Levin argues that all living cells, not just neurons, possess a basic form of cognition. This means they have goals, make decisions, and solve problems within their specific domains (e.g., anatomical space, chemical space).
The type of thinking performed by a cell depends on its environment. The type of thinking of the cells, collective is focused on creating/maintaing anatomical structures.
Morphogenesis: The process by which organisms develop their shape. Levin emphasizes that this is a problem-solving process driven by cells collectively navigating “morphospace” (the space of possible anatomical forms).
Bioelectricity: A crucial layer of communication and computation beyond biochemistry. All cells maintain a voltage difference across their membranes, and this voltage state is used for signaling. Ion Channels are protein “gates” in cell membranes that control the flow of ions (charged particles), creating voltage differences. They are like biological transistors. Gap Junctions are direct connections between cells that allow ions (and thus electrical signals) to flow freely. They erase ownership information on signals, promoting collective decision-making (“mind melt”). This makes is the start of a collection. Electrical Networks, collections of cells connected by gap junctions form networks that can store and process information, similar to (but slower than) neural networks.
Memory in Biological Systems. Memory is not stored only in DNA; there exist multiple ways: Chemical networks (gene regulatory networks, dynamical attractors), Cytoskeletal structures (physical arrangements) and Bioelectrical states (like flip-flops, volatile RAM) where The configuration persists, acting like memory, even without changes to physical hardware.
Teleophobia is being worried that the assignment of goals/agency is being misattributed, when really, any useful model which makes helpful decisions could and in his opinion, should be referred to as agentic and described with cognitive descriptors.
Planaria: A Model System
Planaria do not age. They continuously regenerate, demonstrating that aging isn’t an inevitable thermodynamic process.
Planaria can regrow any body part. Each fragment “knows” what’s missing and how to rebuild.
A bioelectric network stores the “target morphology” (the ideal body plan). This network can be reprogrammed (e.g., to create two-headed worms) without changing the DNA. The altered body plan is heritable through fission (splitting), demonstrating non-genetic inheritance.
Planaria accumulate mutations but maintain perfect anatomical control. This challenges the idea that DNA fully determines body plan, highlighting the role of bioelectric “software.”
Xenobots: Synthetic Organisms and “Engineered by Subtraction”
Xenobots are self-assembling, bio-robotic organisms created from frog skin cells (Xenopus laevis).
When isolated from the rest of the embryo, skin cells spontaneously form xenobots with novel behaviors. These include: Movement, navigation, collective behavior, and Kinematic self-replication, in which They build copies of themselves from loose cells, a behavior not found in frogs.
Removing constraints (other cells) reveals the inherent plasticity and problem-solving capacity of cells. The default behavior of the frog cells is to be a xenobot, not skin.
Collaboration with AI (Josh Bongard): Evolutionary algorithms are used to predict and design xenobot behaviors by manipulating cell interactions, not by changing DNA.
Multi-Scale Competency Architecture
Biological systems have goals at multiple levels (molecules, cells, tissues, organs, organism). Each level has some degree of autonomy and problem-solving ability.
Higher levels influence lower levels by altering the “landscape” of possibilities. Lower levels simply follow local gradients, contributing to the higher-level goal without needing to “know” the big picture. Like guiding water down a hill.
This architecture allows for robust development and regeneration even in the face of noise, mutations, or environmental changes.
The goals of an organism (e.g. a human climbing) can and will easily differ, and often run in direct conflict with, the lower-level organizational structures it comprises (e.g. skin cells).
Implications for Regenerative Medicine and Beyond
Anatomical Compiler (Long-Term Goal): A system that translates a desired anatomical form into a set of stimuli that will guide cells to build it. This would revolutionize medicine by enabling the regeneration of limbs, organs, and potentially reversing aging.
Somatic Psychiatry: Treating diseases by targeting the goal-directed behavior of cell collectives, rather than micromanaging at the molecular level.
Understanding and Controlling Collective Intelligence: Developing a science to predict and manipulate the goals of complex systems (cells, swarms, AI). This is crucial for both biology and artificial systems.
Ethical Considerations: Challenging binary distinctions (natural vs. artificial, living vs. non-living, human vs. non-human). Expanding our understanding of cognition to include diverse forms of intelligence.
Evolution and the Nature of Intelligence
General phenomenon: evolution is probably quite ubiquitous because it stems from: heredity, heredity-error, competition.
Evolution doesn’t create solutions to specific problems; it creates machines that can solve problems in various spaces (anatomical, chemical, behavioral).
Even simple organisms may have a basic sense of agency, driven by the need to model themselves and their environment under energy constraints. The belief in free will may be a consequence of self-constructing systems.
Unconventional Cognition: Recognizing and studying intelligence in systems that don’t fit traditional categories (plants, slime molds, synthetic organisms).
Key Metaphors and Analogies
Dogs vs. Legos: Building with “agential materials” (dogs) is different from building with passive materials (Legos). Agential materials have their own goals and require training/persuasion, but they are also more resilient.
The collective behavior of cells is like an orchestra, where the “music” (the emergent behavior) is the “dictator,” not any individual instrument (cell).
Higher levels in the competency architecture “bend” the option space for lower levels, guiding their behavior without direct control. Analogy to relativity.
Gap junctions create a shared cognitive space, blurring the boundaries between individual cells.
Concise Definitions (Some from Levin, some inferred)
Agential Material: A material with its own goals, preferences, and some level of autonomy (e.g., cells).
Target Morphology: The “ideal” body plan that a regenerating system strives to achieve.
Cognitive Light Cone: The boundary of the largest goal a system can work towards, in space and time.
Anatomical Compiler: A future system to design and build organisms by specifying their desired form.
Ioniceutical: and intervention or agent which interacts directly with the bioelectrical state, perhaps through an ION channel, so that the anatomy may be guided in this manner.
Software 2.0: A programming paradigm where, instead of writing explicit code, you train a system (like a neural network) to achieve a desired outcome. Analogous to training cells.
Teleophobia: being wary of falsely attributing traits such as intelligence and intention onto something.
Introduction: Beyond the DNA-Centric Model
The traditional view of DNA as the sole “software” of life leaves many biological mysteries unsolved, such as flexible development and regeneration.
DNA mainly just decides the materials a cell contains.
The “Picasso Frog” Experiment
Tadpoles with rearranged facial features (“Picasso frogs”) can still develop into relatively normal-looking frogs.
This demonstrates that development is not a rigid, pre-programmed sequence of movements. Instead, it’s a flexible, goal-oriented process aiming to minimize the *error* between the current state and a “target morphology” (a correct frog face).
This adaptive, error-correcting behavior is a form of biological intelligence, implying that cells cooperate and make decisions to achieve a specific anatomical outcome.
Bioelectricity: The “Software” Layer
Cells communicate not only biochemically and physically, but also electrically. This “non-neural bioelectricity” exists in *all* cells, not just nerve cells.
Electrical signals between cells form networks that process information and store “pattern memories,” similar to how brains store information. These memories include large-scale anatomical plans.
By visualizing these electrical conversations (using voltage-sensitive dyes), we can see the “electrical software” running on top of the “cellular hardware” (defined by DNA).
Bioelectricity directs cells when to decide left, right, head, tail and many more directions.
Analogy: Machine Code vs. High-Level Language
Traditional biology often focuses on the “machine code” level (biochemical signals between individual cells), which is like trying to program a computer by directly manipulating its wires.
Bioelectricity is like a “high-level language” that controls the *overall anatomical outcome* without needing to micromanage every cellular detail. Understanding this language gives us a powerful new way to influence development.
If you can rewrite the electrical, then you may influence large body systems without interacting directly with DNA.
Planarian Regeneration and Rewriting the “Body Plan”
Planarians (flatworms) are masters of regeneration, capable of regrowing any lost body part.
There’s an electrical gradient (head-to-tail) in a planarian fragment that dictates where new heads and tails will form. This gradient can be manipulated.
By altering this bioelectric gradient (by turning specific ion channels on or off – *not* by applying external electricity), researchers can create two-headed or no-headed planarians.
Remarkably, two-headed worms *continue to regenerate as two-headed* even after both heads are cut off (with no genetic editing), showing that the “body plan” memory has been *permanently rewritten* and is stored bioelectrically, not just in DNA.
The bioelectrical field stores anatomical “what to do’s” and may be used to rewrite memory.
This demonstrates a form of *non-genetic* memory – a stable, rewritable pattern that influences future regeneration.
Beyond Planaria: Inducing Organ Growth
By manipulating ion channels in tadpoles, researchers can induce the growth of ectopic (extra) eyes in locations where eyes don’t normally form.
These induced eyes are complete and functional (with lens, retina, optic nerve), showing that the *body already knows how to build complex organs*. The bioelectric signals trigger existing “subroutines.”
Researchers are figuring out how to make limbs and hearts.
Xenobots: Novel Life Forms from Frog Skin Cells
Xenobots are created by isolating frog skin cells and allowing them to self-assemble.
These cells, genetically identical to normal frog cells, spontaneously form *new* organisms with unique behaviors (movement, navigation, maze running), distinct from tadpoles or frogs. This shows how cells will find some type of structure when constraints are changed or removed.
Frog skin cells use their cilia (normally used for moving mucus) to *swim*, demonstrating how cells can *repurpose* their existing “hardware” for new functions.
Xenobots will create spontaneous and “unprogrammed” behaviors.
AI (in collaboration with Josh Bongard) can model and evolve xenobot designs *in silico* (on a computer) before they are built in the lab. This highlights the incredible plasticity of cells.
Researchers used computers to build the evolutionary tree/history of xenobots, when it never exisited before, meaning a lifeform/body that has a history outside Earth and created by pure modeling.
Implications and Future Directions
**Regenerative Medicine:** Cracking the bioelectric code could enable us to regenerate limbs, organs, and correct birth defects by rewriting the “target morphology” that cells strive for.
They are able to fix traumatic injury in frogs and other life, potentially stopping the cause of cancer, fixing aging and degenrative issues in animals.
**Tumor Normalization:** Cancers could potentially be “normalized” by influencing the bioelectric signals that control cell behavior and tissue organization.
**Broader Understanding of Intelligence:** Bioelectricity highlights that biological intelligence exists *before* the evolution of brains, suggesting new avenues for AI and machine learning based on how *body cells*, not just brain cells, solve problems.
Researchers want to find way to communicate to large body systems with large signals/blueprints and *not* in detail through each individual.
Introduction: Computational Boundary of a Self
The paper attempts to formalize thinking about any possible cognitive system or intelligence. Intelligence is defined (per William James) as competency to reach the same goal by different means (cybernetic definition). It’s about navigating a problem space to achieve a goal despite obstacles.
Origin and composition are considered less important than shared properties of cognitive systems. The origin story is not reliable and there’s not anything inherently special in any given cognitive system and may depend more on it’s implementation (i.e. hardware in robots, biocomputer in life).
Cognitive Light Cones
Concept borrowed (and inverted) from physics’ spacetime diagrams (Minkowski). Time is vertical, spatial dimensions are horizontal.
The cognitive light cone represents the spatio-temporal size of the *biggest* goal a system can pursue. This is NOT about sensory reach, but about the scale of the goal state.
Examples given progress from simple organisms to complex beings, demonstrating growth in spatio-temporal considerations:
Tick/Bacterium: Small light cone; focused on local chemical concentrations.
Dog: Larger light cone (memory, prediction); goals like keeping intruders out of its territory. Limited ability to care about events far away in space/time.
Human: Very large light cone; goals can include abstract, long-term, even unattainable things (world peace, preventing the sun’s burnout). Limitations still exist in caring capacity.
Humans’ large light cones include the ability and understanding to work on unachievable goals and the stress that it results in is likely to be something of great consideration in humans (perhaps) as a species and may explain part of it’s drives, in this way may not be completely unlike other creatures with large cognitive lights cone and goals that do extend to some future event that they will still likely witness in their time.
One key takeaway on Human cognitive light cone: The ability, understading of capability, and undertaking/investment in nonachievable goals that might explain motivations.
The ability and perhaps motivation and direction in species. The capcity to care (to help address another beings suffering for exmaple) on other beings is not able to grow at a 1:1 and is also very important as perhaps with Bodhisattvas.
Related to the TAME (Technological Approach to Mind Everywhere) framework, which aims to move questions of cognition/sentience/intelligence from philosophy to testable scientific claims. Every statement should create test-questions and predictions for the systems capability.
Intelligence is not binary; it’s a spectrum. Claims about a system’s cognitive level should define the problem space, the goal, the system’s capabilities, and then be testable, creating new data for science, than a thought exercise alone.
Collective Intelligence
Compound intelligence consists of a heiarchy: All intelligence is collective. Organisms are made of cells, cells of components, etc., in a nested hierarchy. This is to be contrasted with something like, Integrated information, the former makes specific assumptions.
Organisms’ large-scale light cones do not belong to individual components (cells, tissues). Cells and their individual parts are geared and set towards and interact with and respond/exist in physical/physiological world; this is the fundamental mechanism/method of higher organization (or it appears). The higher system must emerge through these cells’ actions.
A “scale-up” process exists, in which subunits (cells) with individual competencies combine to create a larger cognitive entity.
Levin points that de Cartes view was inaccurate (specifically about pineal gland, though he admires many other traits of de Cartes’). The mistake, fundamentally is that of composition and of a misunderstanding of the scaling capabilities from the underlying components that drive the overall organization. This may also explain certain cancer growth as cancer as being able to operate beyond it’s boundaries when certain properties change (bioelectrical properties).
Robots don’t get cancer because their parts are *passive*; they don’t have individual agendas, as biological cells do. This risk from the properties that enables higher organizational systems, also brings its own risk and down-side: cancers for example, by making an entity, using smaller and active materials.
Key question: How do the properties and *novel goals* of the compound intelligence relate to the properties of the parts? How does this scaling up happen, and how can it fail (e.g., cancer)?
I.i.t. The main disagreement, it seemms, is with a lack of separation between intelligent and non-intteligent behaviour, between intelligent-cognitive tasks vs tasks around consciusness, it has more of a philiophical bent than perhaps having any other reason for the fundamental misunderstanding and fundamental disagreement.
Scale-Free Cognition
Scale-free, based from a previous publication, the term “scale-free cognition” is used to indicate that principles of cognition apply across different scales. Humans recognize intelligence best at their own scale (medium-sized, medium-speed, 3D).
We struggle to recognize intelligence at very different scales (e.g., bacteria in the gut solving physiological problems, organs maintaining their own structure).
Evolution itself could be seen as a large, distributed agent, with each animal being a “hypothesis” about the world, subject to success/failure. We must detach from the need of any type of ‘mind’ being limited to that of roughly Human-scale; stretched out is difficult for the mind, it must remain relative and flexible.
Name: perhaps is slightly off because the term may imply something about reliance, and there exist different degrees of scale, yet not having scale may only emphasize differences in properties between that in smaller/greater scales and may give undue importance to scaling when, maybe instead, it ought to be on the system’s cognitive behaviour.
Freedom should, in Levin’s description mean freedom in system doing whatever the systems is likely/intended to do with freeeom of scientists having freedom in approaching that in all types of views, including those previously impossible or hard to see/visualize/analyze.
Surprise and Infotaxis
Agents strive to minimize “surprisal,” which relates to stress. Stress and surprise are equivalent in this context. This is related to concepts like the Free Energy Principle (Friston).
Regenerative development/regeneration can be viewed as cells navigating “morphospace” (the space of possible anatomical configurations) and minimizing errors, that could represent itself through stresses, to reach the target morphology (e.g., a salamander limb).
Intelligence in biological systems is not simply complex emergence, but the ability to reach a goal despite perturbations and novelty; It’s active error/stress mitigation and purposeful decision making for optimization/reducing future risks/hazards and generally problem solving. It MUST be experimented upon to verify intelligence, through new behaviours as data.
Salamander limb regeneration: Cells regenerate the correct limb, and then *stop* when it’s complete. This is framed not as feed-forward emergence from local rules, but as homeostatic error minimization by a collective intelligence with a “set point” (target morphology).
This is not dissimilar from thermoregulation.
Development regulation (cancer suppresion, metamorphisis, etc. have common trait and act not disimilarly with systems to keep errors low.
Homeostatic ability/behavior: loop in system, has target(set points), works towards goals and away from any potential risks (from environment/genetic) and the like and tries, as efficiently as it can to get as close as possible to said set-points, taking whatever the system can within it’s realm of capabilites.
These abilities might be as robust as finding NEW ways to find optimizations/reaching set-points in the given situations to try and overcome obstactles.
Key point: This requires a *set point* (like a thermostat). Cells expend metabolic energy to achieve large-scale goals (organ shape), going beyond their individual needs. This can occur if the environment changes to accomodate this energy expenditure (heat) by system.
Stress is the term used here for that ‘keep-trying-to-minimze-the-error-state-which-may-create-the-expenditure-of-extra-energy’
Stress has commonalities across domains: Engineers on meso materials, psychologists on general ‘well-being, behavior’, evolutionary psychologists of well-being/condition of eco systems/societies and others in different areas that are hard-er to recognize, but which still share similiar properties such as in micro-environments for cells and systems and their interactions (stresses here not that dissimilar to stresses on systems.
There are specific, large requirements from collective systems that are working, in aggregate, from simple components such as skin cells, who otherwise may, normally do simpler processes, these ‘goal differences’ result and could manifest into unique and perhaps useful signals and/or molecules.
Geometric Stress example: and eye grown artificially, elsewhere may not, and can never stress specific cellular cells individually but are the system (collection of tissues/components, and whole system working) stressed. For instance: the tissue’s overall configuration can impact on overall body. Another example is if a new eye forms in tad-pole tails.
The spreading of this stress, is key: by letting the underlying system understand where the greatest difference lies, it can find new way/mitigation-patterns/methods to reduce total ‘errors’, not that different to a cooling-magnetic having all underlying componets, in this scenario and configuration find alignment: free/potential energy.
Temperature might have a bearing too, which may change with more stress/movement of different, smaller agents with new degrees of flexibility/freedome for underlying particles; creating conditions more likely and more conductive to more likely, new behaviors and changes (and perhaps errors in the right situations)
Stess might make the parts work to benefit overall whole system and may explain away certain questions with regards altruism.
Gap Junctions and Scaling of Collective Intelligence
Gap Junctions: Proteins that form “submarine hatch”-like pores between cells, allowing small molecules, current, and signals to pass directly.
“Magic” of Gap Junctions (according to Levin): They facilitate the *scaling* of collective intelligence.
Traditional Signaling: Secreted molecules. The receiving cell knows the information originated *outside*. It can be ignored, treated with caution, or learned from. The receiving cell has more control.
Gap Junction Signaling: Signals pass directly into the receiving cell’s internal environment, *without metadata* indicating origin. There is *no* indication that it’s not internal.
The absence of signals of signal orign and type fundamentally means an uncertainty in behavior and decision making from higher up, at individual/collective/system.
Result: A “mind meld.” The receiving cell cannot distinguish the signal from its own internal signals. It erases individual cell identity and promotes collective computation and decision-making.
Benefit for decision making with collective systems by preventing (removing/mitigating risks) in conflict/misaligned/different interests from underlying structures that may otherwise benefit from ‘selfish’ decision and actions to the detriment of higher level structure (the body, here in this discussion),
It creates a larger computational space, and better spatial-integration: a single system, as oppose to independent, singular-focused, units.
Enables collective sensing of biomechanical forces (as a cell sheet). It permits, and results in a greater, aggregate organization, having greater reach, with collective/overall goal: making them harder for cells (to take example) in micro, or more localized events, harder to ignore.
It is also worth noting that while they appear as fundamental: this configuration can go wrong and, in so far as this has a negative effect, would mean care must be taken as the overall aggregate-organizational structure can/could be to the benefit and/or detriment of the component structures and there’s always a ‘give-and-take’, an inherent and possibly (maybe!) ever-present fundamental risk and the need/requirement of finding balance/middle path that benefits overall structure as well as the parts making up this new-collective system (in cells for example this could mean cellular death for body ‘success’; like skin cells during a hand scratch after going rock-climbing).
Cell Membranes and Information
Gatenby’s claim: ~99% of Shannon information in a cell is in the membrane and transmembrane gradient.
Interpretation: The cell membrane is the primary interface for cell-cell communication and interaction with the outside world. It contains receptors, ion channels, and biomechanical structures. All other cells, within a ‘larger’ organism are a ‘foreign agent’ or unknown to underlying-systems and could represent as a potential (or an actual, confirmed risk).
Cells ‘hack’ other cells/outside world at membrane interface. There are all the interesting structures and systems that give important clues and information about what cell interactions exist, do, create.
Shannon information is observer-relative; different observers can derive different amounts of information from the same signal. This highlights interpretation-dependant information (from signals or actions that do something), including observer dependant views, not that they can be.
It has other interesting components and properties to consider as an additional, external agent (as ‘the human’ is external to underlying cells: cells here can then view, and consider humans in the general world: for that interaction-ability for systems)
Expanding the horizon of light-cones lets shannon info to acquire a greater range (larger in scope: in breadth and height).
Meaning and the Observer
Chris Field is credited here for an important, complex take.
Central role of the *observer* in information theory (Shannon or otherwise). Meaning is not inherent in a signal; it’s *imputed* by an observer.
Meaning is Observer-dependant: this concept highlights this to the N’th degree; a different view of this might lead to different conclusions and decisions/observations (and subsequent actions based from).
Different observers can derive different meanings from the same data. Perspective matters (relates to Josh Bongard’s “polycomputing” idea).
Multiple ways can have advantages in analyzing/looking at similar data that might then lead to insights impossible/unseen from previously impossible ways of looking at systems (perspective is useful for different angles, and with right ‘perspectives’: the tools for it).
Polycomputation and perspective that might result in greater understanding by ‘squeezeing in extra power to the bio/computation in process’; that will lead to greater processing/computation ability.
This emphasizes perspective shifts as useful, separate (or even independent) tool to get great insight that may give ideas to solutions/behaviours for a system (and vice versa) and has much application, perhaps more then ‘common’, to those beyond simple/narrow computation ability.
Cancer as a Breakdown of Cognitive Scaling
Reframing Cancer: Not *why* we get cancer, but *why isn’t everything cancer*? Our cells’ “default” state is to reproduce and move where life is good (tumor-like behavior). The more ‘surprising’ question: is the reverse.
The puzzle: what is being regulated to ‘control’ a structure; to have this overall-goal in a direction in aggregate: it appears to have a collective/overall higher control for aggregate, whole.
Individual cells have small “selves” defined by their local environment.
Electrical and biochemical networks establish the boundary between self and the outside world. A cell connected to many others has a *large* self (potentially the whole body).
With the right signaling pathways: cells can grow/change/be-directed: and the mechanisms for higher-level-systems might (could!) and may depend on this.
An oncogene that disrupts gap junctional communication *shrinks* the cell’s self to a single cell. The cell then pursues individual-level goals (metastasis, proliferation), which deviate from the collective.
It highlights breakdown (failures) as an almost-expected outcome, in these otherwise powerful systems, but the very mechanism by which that makes it useful also inherently create the very possibility/mechanisms/outcomes of misaligned or poorly-aligned states that might cause this stress/issues that could go unheeded/unattended too (because these systems are complex, intricate).
Gap Junctions is key, to preventing the cells (example/use) to ‘regain’ that original identity with it’s associated individual decision/behavior.
Gap junctions make all other external forces a common issue: meaning that they do something, if that new issue is in direct-conflict/alignment with those the other entities, will create issues that otherwise can’t be addressed and in that way could have issues, is a source of great importance and of stress to ‘keep’ that integrity (for optimal system operation).
Signals such as electrical signals might also be key, here (which could have its own interesting implications that, while currently unexplored/explained in this interview, nonetheless represent fundamental understanding-path that is of crucial importance that may go-unaddressed here)
Ennviromental interactions/communications might become ‘external’/not-important signals/input to smaller system once higher ‘connections’ in/of these new networks are broken and new behavior becomes dominant, as consequence.
Self: is different, yet of equal value to it’s immediate self vs that higher organization system that does otherwise act together for the overall-higher-system in that sense then; could still be seen as ‘selfish-behavior’ even if/when operating to improve aggregate structures such as entire ‘being/system’: meaning that any decision is made that favors these goals are inherently ‘selfish’ actions, on the level they occur/happen/exist; the scope makes difference.
Cancer cells are no *more* selfish than normal cells; they just have a tiny self and thus pursue smaller goals. Cancer’s inherent behavior is a natural outcome of such situations and environments that have similar ‘breaks’ of properties/structures of ‘connection’ and thus the very thing (aggregate structures of importance such as for overall body or larger ‘higher systems’): would mean that it is normal, not an irregular occurrence/outcome given such system failures:
It opens, at least theoretically: a ‘novel’-solution space, in how such cells/tissues may be mitigated: re-integration of structures through electrical signalling, could lead to an intereative (and useful!) mechanism to control/guide new and more stable cells (to make higher organization).
Therapeutic Implications of the Bioelectric View
Cells can be cancerous without there having been genomic damages to begin with: cancers, given new structures/connections of communications might also be grown in artificial bodies: it does not require inherent faults to grow.
By creating artificial electrical fields in-vitro or ‘else’: one might even use structures like embryos as new bio-computer/electrical signaling based pathways in computation that do not exist, not do require genome structures but would have use: by connecting together such components to make novel and new biological systems that otherwise wouldn’t and would not need to exist with it’s specific properties.
This leads to thinking: a therapeutic strategy, which involves reconnecting cancerous cells to the bioelectric network, causing them to normalize even with oncogenes. They have found it effective in their current bioeletrical/xenobot and tadpole studies to mitigate/treat cancer by ‘redirecting’ cell behaviour; instead of the ‘old, default’ ways/methods that might also do harm.
*Don’t* kill the cells, but instead reconnect, change goals, and set priorities (set/goal posts). The molecular hard-ware may be not useful/of focus, but to reprogram (to do this would result in less stress-damage-related to kill-first). Focus then is bioelectrical signaling pathway, here instead of molecular (kill-first) methodology and/or hardware related solutions/methods (such as what has historically been the way such situations may get addressed that results in non-benign outcomes and failures and could/would do better, than perhaps older systems) to give it new configurations to do more effective, stable/integrated system control, which is very difficult (impossible maybe? and would, or might not address fundamental reasons and mechanisms).
‘Telling cell’: is not easy, hard to direct behavior (to ‘talk’): this does suggest a fundamental communication-challenge that does exist, or appear to exist with cells and cell tissues.
Shift from low-level manipulation (genomic editing, rewiring pathways) to *high-level* communication, “retraining” the system to use its own competencies. Not that there are many effective systems to even begin to direct the signals (perhaps), let alone systems, at an electrical levels, but: that this may change/has potential in future and, more so in general systems.
Analogy: Don’t rewire the computer hardware every time you want to switch programs. Use the existing software interface and “built-in” control systems. Use higher level/better (hopefully more robust/safe/stable etc.) levels, which may improve chances for overall goals of larger, more robust organization systems, too and in so doing address larger organization failures, through better information/control mechanisms (signals in particular, or whatever data/structures do create differences in cell systems)
Drugs, Words, and the Multiscale System
“Drugs and words have the same mechanism of action” (Benedetti): High-level cognitive expectations (placebo effect, context) can filter down to molecular biology.
Our “Executive decisions”/behavior can directly translate down-the-chain for action by parts/component that have the very capability to be active/reconfigure themselves: the “executive decisions”, for exmaple to move out-of-bed; require action from ‘lower system’: down-chain commands for decisions/movements to effect on lower scale for it to ‘result’ on larger levels: high decision > cell component action > outcome in large-body (like humans).
Explains things like hypnodermatology, placebo effects, and even everyday voluntary actions. High-level cognitive states can influence molecular physiology.
Important to emphasize how significant the very notion of this “scale”-interaction” from very small scales of individual ‘things’ making the whole; which can and is seen as very much less understood/important/of value.
This is how our bodies behave; how it appears the way it appears, every second; and when, and if/when working ‘well’, gives stability/predictable configuration/structure: and is a key factor. It can, given all this also very much explain that perhaps seemingly small interactions can and/or will affect aggregate/whole in what could be surprising way with new-scales.
Components of the Morphogenetic Field
A cell gets influenced at multiple levels: There exist ‘other’ considerations, not discussed in talk: but some factors.
The cell takes a ‘vote’ from the field, based on many signals:
Bioelectricity isn’t the *only* factor, but it’s crucial because it acts as the “excitable medium” and *computational layer* that stores the cognitive information of the collective intelligence, and this *computational layer* may have implications that go (well?) beyond/much/more into the system(s) than all the previous properties discussed, combined and could create important shifts in new-areas that previously were harder to achieve.
Brain learned electrical signaling trick from pre-neural evolution. The exact same principle used.
Harold Saxton Burr’s “Privileged Information”
Harold Saxton Burr (1930s-40s): Measured electrical fields in various biological systems (cells, embryos, trees, humans). He was one of first good voltage measurers of this sort, even during time/periods without precise tools; without means to manipulate that environment/electrial systems directly to any reasonable measure and to measure accurately that and without direct insight and/or ability/way to impact specific structures/mechanism that resulted (and continues to give these, that may come out of his discoveries).
Burr described the *future* development of bioelectricity, and (according to Levin) he was remarkably accurate despite having only crude measurement tools. He accurately understood some mechanisms that could exist for structures in organisms/collective biological systems that make some cells grow or do new configurations without explicit-directions but yet that which are highly, robust, complex and useful (he perhaps found hints into a form of biological bio-computation in/through cells, from a voltage/energy and interaction from other electrical mechanisms), an electrical layer; a “computational bioelectricity”: where cells work individually as well in aggregate for overall body-configuration through unique/undirected behaviours that come to create overall systems. He lacked more understanding from his tools/in-vivo tools: but the voltage level could indicate certain actions that go beyond ‘simple structures’ by enabling cells in ‘micro environments’, not that much unlike how an isolated/non-interactive or perhaps even non-important micro-electrical activity from micro, seemingly (to ‘higher up, outside’/non-connected agent(s)/viewer), but that may ‘non-specifically direct’ underlying and connected nodes through simple communication, with enough of it, and large scope: give the impression, action of directed.
Burr’s work was influential for Levin, because it may highlight and emphasize the need for high precision to not overlook some specific important signals/behavior of systems, and specifically a type that can (easily? maybe even accidentally) be ‘masked-away’; through less percise tools; a reminder.
A call that ‘Burr made these same discoveries’, without the need for tools (not ‘advanced’; from perspective-perspective perhaps, though he created what was probably the only, that ‘one’ to understand in-vivo systems).
Other Pioneers, who built further on discoveries were also mentioned, though perhaps that of which gave direction was made in his earlier/later work, that have greater bearing/impact of what future potential directions/impact it may still have to come (such as those areas/disciplines mentioned that may make new implications as technology/science does its future works; perhaps that can enable even ‘higher resolution’ systems, that otherwise do give very specific data and/or can ‘target’/influence it too, specifically too for what outcome it may create that otherwise can’t and does give potential to solve, currently unresolved systems in biology.
Zen Buddhism and Compassion
Final paragraph of “Scale-Free Cognition” paper mentions Zen Buddhism.
Connection between cognitive light cones and the Buddhist concept of enlarging compassion (actively working towards the welfare of all beings).
Thinking about diverse intelligences forces us to revise simplistic distinctions about who/what deserves compassion. Ethics needs updating in light of new possibilities (hybrids, cyborgs, etc.).
The *goal* of enlarging our light cones (through technology or other means) is not just technology itself, but to *increase our capacity for compassion*.
The talk referenced to Buddhism is from a different perspective and does (likely/potentially) expand-on previously not ’emphasized’ traits/understanding for how that “Compassion”-concept may relate and does explain: compassion has a key trait that must also improve alongside intelligence, even if they have, in some way ‘similar mechanism(s)’; for there to exist that fundamental property of ‘compassion’; that perhaps drives all intelligence and has not that dissimilar-configurations in structure and/or direction(goals/target)/understanding. Compassion is key: and intelligence and compassion must work together and go alongside one-another and it appears perhaps not only and ‘simply’ for it to be morally/ethically/’fair’: as without ‘Compassion-with-it’; a “light cone”/an intellgent behavior-like may get misdirected/misguided, given that intelligence, if not guided-by-compassion is perhaps missing a component (as fundamental and core property of it): for ‘correct’/’intended’-way: without care, and love, could lead the intelligence(in this scenario/example being discussed) to get to a state with undesired configuration (by failing to have core direction/components with intelligence not having inherent traits/direction of system(s) with-compassion: might be key) for larger goal to have a very unique outcome (compassion must accompany intelligent-systems for stable-goal); for its goals, without misalignments, it may never address higher structure considerations, it is not limited to compassion towards (smaller/sub-entities); instead there is perhaps an alignment required between scales.
“Compassion”: this isn’t that general care: of just thought-excercises; compassion being: an active (key: actions; something active and real-work, as opose to just mental-considerations), is fundamental.
There are many complex components here, and a larger paper can give details that make the discussion(talk-itself) do, if in the wrong/very limited scope may have very wrong implications.
Links
www dot drl-dot org – mentioned to contain information about the presenter, and other information mentioned through-out interview.
Mentioned that the presenter/speaker is huge-fan of some of AI Art works such as (Midjourney): and has blog posts/articles (that can be interesting: is not directly connected to this interview, though worth, asides, nonetheless, a look from that view if desired/in/with similar interets)
Note of the site’s complexity, breadth of areas covered and inter-links, this may make navigating this complex.
Biology, Buddhism and AI: care as a driver of intelligece: paper and reference to Buddhist concept mentioned in last/wrap-up, section and could benefit with having that as part of consideration/discussion with more references/people on a/that discussion (there appear to exist greater context and perhaps some of those not immediately relevant, unless with a lot of additional context) to make some complex issues; not very simply covered here and/or that lack (critical-thinking perhaps!), key here: which it would be with some fundamental-aspects about these systems and considerations that does take complex understanding of what compassion is: “An Active”; vs (that which otherwise appears as: “that it means feelings”), that does guide that behaviour, action.
Introduction and Bioelectricity
Non-neural bioelectric states exist: All cells, not just neurons, use ion channels, electrical synapses, and neurotransmitters for communication. These are ancient mechanisms, predating nervous systems.
Cells create voltage gradients: Cells maintain an electrical potential difference across their membranes by controlling ion flow (potassium, sodium, etc.). Each cell is like a tiny battery.
Bioelectric signals control anatomical space: Before controlling movement in 3D space (like brains do), bioelectric networks controlled cell behavior to navigate “morphospace” – the space of possible anatomical forms.
DNA specifies hardware, bioelectricity is software: The genome encodes the *structure* of cellular components (proteins, ion channels). The dynamic bioelectric activity is the *computation* that uses this hardware.
Experiments and Implications
The “Electric Face”: Before genes for facial features turn on, a bioelectric pattern resembling a face appears in the embryo. This pattern can be disrupted (causing defects) or moved (creating ectopic eyes, even in the gut).
Modularity and Triggers. The voltage patterns call modular subroutines. Injecting a channel will create an eye from cells that normally create a gut. This eye will consist partially of the normal gut cells. The bioelectricity instructs neighboring cells to be the appropriate part, implying that only very tiny injections are required, because after they trigger the local anatomical goal, the other cells comply to fit that anatomy (limb, liver, eyes, etc).
Cancer as disrupted communication: Cancer cells disconnect electrically from their neighbors, reverting to a single-cell-like state. Forcing cells to *remain* electrically connected can prevent tumor formation, even with oncogenes present.
Planarian regeneration and morphospace: Planaria regenerate perfectly. Their cells “know” what the target body plan is and how to rebuild it. This is a *collective intelligence* problem, not just a molecular biology problem.
Xenobots: Frog skin cells, isolated from the embryo, *spontaneously* form new organisms (xenobots) with novel behaviors (movement, self-assembly). They have a normal frog genome, but express a new “default” behavior, not selected for by evolution.
Identity and electrical connection: Gap junctions (electrical connections between cells) can blur the lines between “self” and “other,” contributing to the formation of larger-scale collective identities.
Placebo Effect: Mental Activity has direct bearing on physical activity. As you may send commands through intention to create new hormones or command the limbs to do things, so too may your mind instruct cellular growth and behavior, at least in the body which that mind exists.
Broader Concepts
Collective intelligence: Morphogenesis (body shape formation) is a *collective intelligence* problem, similar to how brains work, but operating in anatomical space instead of 3D space.
Homeostasis and set points: Where do anatomical “set points” (the target morphology) come from? It’s not just evolution; even without selection, cells exhibit surprising plasticity and new behaviors (e.g., xenobots).
Intentional Stance: When dealing with novel biological systems, it’s crucial to empirically *test* different levels of assumed cognition/agency to see which best explains their behavior (per Dennett).
Aging and the regenrative capacities of planeria indicates that as long as there exists adequate regenerataive processes, organisms like planaria do not age at all, but any human is not known to have their level of regenreative capabilities, so it could be that aging occurs when those regeneravtive capactities cannot keep up.
Consciouness can possibly interact with Morphogenesis: due to being part of a greater whole with cells of the same physical organism. And general Anesthesia being related to the gap juntional activity, disruption can indicate it impacts short term consciousness or attention, which when restored, reconfigures things at times as new random memories (in pleneria, as new head morphology).
Minimal Cognition: because there exists particles following least action principles that display quatum indeterminacy already are ‘something’. this suggests ‘goals’ and ‘spontenaity’ are, well, spontaneous!
Introduction and Levels of Explanation
A major myth in biology is that the best explanations are always at the molecular level. Different levels of explanation (biology, chemistry, physics) offer different insights, and higher levels have autonomy.
Emergence is a measure of surprise for the observer – how much a system does that wasn’t anticipated from the properties of its parts. It is relative, not absolute.
Cognitive functions (learning, memory, conditioning) can be found even in very simple systems like gene regulatory networks. It is important to go past the levels.
Unifying Themes: Embodied Mind and Intelligence
Levin’s work across various fields (cancer, development, regeneration, AI) is unified by an effort to understand embodied mind and intelligence in diverse, unconventional forms.
Goals of organisms are typically attributed to evolutionary history. Synthetic constructs (xenobots, anthropods) allow studying the origins of goals in systems *without* such history.
The “cognitive light cone” represents the size of the largest goal a system can pursue, in space and time. This concept helps understand cancer as a *shrinking* of this light cone, where cells revert to individualistic goals.
Cancer cells aren’t necessarily more *selfish*; they have *smaller selves*. This leads to research on reconnecting cancer cells to bioelectrical networks, normalizing them *without* killing them.
Defining Intelligence and the Platonic View
Intelligence (following William James) is the ability to reach the same goal by different means. This definition emphasizes problem-solving in a specific problem space, regardless of the physical substrate (brain, synthetic system, etc.).
Levin leans toward a Platonic view of intelligence, akin to the mathematical Platonism, there exists, in fact, a space where mathematical properties of computation that think and compute “live” in such that we merely discover. He suggests that not only do rules of maths “live” there, but other cognitive states live as well, not limited by material states, we don’t just invent minds; physical systems can “harness” pre-existing intelligence.
Estimates of intelligence are *not* objective properties of a system. They are guesses about its problem-solving capabilities, reflecting *our* knowledge (or lack thereof).
Self, Memory, and Dynamic Interpretation
We are not static entities, but rather collections of interacting perspectives. “Selflets” are thin slices of experience.
The continuity of self is perceived by *other observers* based on the consistency of behaviors and properties. We are interested on how the system behaves to ourselves, how to better get “messages” (engrams).
Our access to the past is through memory engrams – *interpreted* traces, not direct access. These traces must be dynamically reinterpreted in new contexts.
Our current actions are like messages to our future selves, constraining or enabling their possibilities. This creates a symmetry between our future self and *others’* future selves, with ethical implications.
Under resource constraints, agents *must* coarse-grain; they cannot track everything. They create compressed representations (like memory engrams), focusing on salience, not fidelity.
Compression, Perspectives, and Synchronicity
Highly compressed data can appear random because correlations are removed. This has implications for interpreting potential signals from advanced civilizations.
What we experience depends on an interpretive agent. What we compress for memories, have to be de-compressed by us in future times, thus leading us to needing a reinterpretation of the rule to the given observation.
Perspectives are fundamental: commitments to what to measure, what to pay attention to, and how to weave that into a model. Every perspective necessarily shuts out more than it lets in.
If we are part of a larger cognitive system, recognizing that might look like *synchronicity* – meaningful events without apparent causal connection at our level.
Bioelectricity as Cognitive Glue
Bioelectricity is *not* magic, but a crucial mechanism for enabling collective intelligence, by being used to create a policy for cooperation.
Bioelectricity allows the cognitive light cone to scale up. Cells connected in electrical networks form larger emergent individuals with higher-level goals and capabilities.
Cells create their agency with signals that enable other parts of their larger organism to move.
Examples from the Levin Lab
Early work showed bioelectricity’s role in left-right asymmetry in chicken embryos, manipulating this with ion channel constructs.
“Electric face”: Bioelectric patterns in nascent ectoderm *prefigure* the formation of facial features. Birth defects disrupting these patterns can be corrected bioelectrically.
The first demonstration of gaining the regeneration functions: Bioelectric manipulation can *induce* tail regeneration in tadpoles, demonstrating the control over large-scale anatomical outcomes.
A 24-hour bioelectrical stimulation to frogs triggers *a year and a half* of leg growth, showing a high-level command (“build a leg”) without micromanaging the process.
Bioelectric signaling is linked to cancer. Cells can be induced to become metastatic with *inappropriate* bioelectric cues. Conversely, bioelectrical connections can *normalize* cells expressing strong human oncogenes.
Planaria: Two-headed worms show a *permanent, non-genetic* change in target morphology. This demonstrates physiological memory.
Planaria’s highly chaotic genome and remarkable regenerative abilities suggest a prioritization of *algorithmic competency* over genetic fidelity.
Anthropods (human-derived organoids) demonstrate that *tracheal cells* can exhibit novel behaviors, including neural repair, *without* any genetic changes. This showcases the plasticity and emergent capabilities of cell collectives.
Implications and Future Directions
Evolution creates *problem-solving agents*, not just solutions to specific problems. These agents have tools (cytoskeleton, gene regulatory networks, bioelectricity) to handle novel circumstances.
Work is shifting towards clinically relevant models: human cancer cells and spheroids for cancer research, and mice for regeneration.
Understanding how information flows across levels (embryos communicating with each other, forming “hyper-embryos”) is a key focus.
Advice for Newcomers to Biology
Paths are in science hard to predict, be vary of following peoples research agendas instead of making your own based on your goals and aspirations, follow your own “guts”.
Develop your own intuition about which paths to take in science, and test those intuitions through experiment.
Prioritize specific, technical critiques to improve your craft (how to make an experiment better), but be wary of general advice about career direction.
Unifying seemingly disparate phenomena (like the different forces responsible for apple falls and planetary motions) is a major scientific achievement. Be open to this possibility.
Levin discusses collective intelligence beyond traditional brains, focusing on cells, tissues, and unconventional organisms.
Turing’s interest in both AI and morphogenesis (shape formation) suggests a deep connection between intelligence and biological development.
The transition of an oocyte “just physics”, into a cognitively-aware, is amazing.
Transition. Nested intelligence where lower structres contribute with their behaviors to reach the goal/attractor state.
Key Concepts: Multi-Scale Competency and Navigation
Biology uses a “multi-scale competency architecture” where nested problem-solvers (cells, tissues, organs) operate at different levels.
“Navigation” of spaces (physical, physiological, morphogenetic) is a central concept for understanding biological intelligence.
Goal-directedness is critical for recognizing and interacting with diverse “agents,” including unconventional ones. Goal can simply mean, in this context, attractor.
Cognitive Light Cone – bounds what can/could be percieved given sensory data limits.
A “cognitive boundary model” helps understand how goals scale in biological systems.
Bioelectricity and Morphogenesis: A Specific Example
Biological pattern formation (how organisms get their shape) is the behavior of a “collective intelligence” of cells in “morphospace” (the space of possible anatomical forms).
Bioelectrical networks (using ion channels and gap junctions) are a “proto-cognitive medium,” an evolutionary ancestor of brain function.
Bioelectrical signals are not just for brains; all cells have ion channels and communicate electrically, allowing them to navigate morphospace.
Homeostasis and its ability to react/regenerate is based on bioelectric communication between the various components and organizational level structures.
Top down influece can and will trump genetic information if there is a homeostatic trigger, for example, cancer being removed due to electric influence and cellular communcation between healthy and canerous cells.
Practical Implications: Bio-medicine and Synthetic Bioengineering
Understanding bioelectrical control has implications for regenerative medicine (e.g., limb regeneration) and birth defect repair.
“Electroceuticals” (drugs targeting ion channels) could be designed to guide cells to correct anatomical outcomes. The cells talk and can direct others towards building/changing something based on an end-goal it may have.
Synthetic bioengineering opens a vast “option space” for new bodies and minds without traditional evolutionary constraints.
Examples of Morphogenetic Intelligence and Plasticity
The collective’s components will “remember”, such that it has the ability to recall and “do” the previous goal.
Slime mold (Physarum): A single-celled organism that navigates its environment using vibrations, showing problem-solving in physical space.
Planaria: Flatworms that regenerate any body part, demonstrating memory of body plan (stored bioelectrically) and adaptability.
Planaria are not limited to its default body type/genetics. One of the reasons for regeneration is it can use feedback mechanisms from others, especially from a top-down influence/heirarchical manner.
Homeostatic error minimizing, or simply, error minimizing is a good strategy when resources are limited, but still allowing complex things like organ formation, from imperfect/unknown/variable components/ingredients.
Frog tadpoles with misplaced eyes: Demonstrate that the brain can adapt to novel sensory inputs without evolutionary pre-programming.
Picasso tadpoles: Tadpoles with scrambled facial features that still develop into largely normal frogs, demonstrating error-minimization in morphospace.
Axolotl limb regeneration. They continue regeneration and will keep attempting to do work “until” they get there.
Nephron example. Nephrons will adapt and change its form and function such that its final outcome will result from top-down influence, especially during stress/pressure.
Xenobots, in the process, create children xenobots that keep repeating the processes, in which are not observed previously in its normal context, in frog. It may continue to develop. The limit to cognitize potential in this new structure, xenobot, are still under works.
Xenobots: Synthetic organisms made from frog skin cells that self-organize and exhibit novel behaviors (movement, self-replication), showing unexpected plasticity.
Scaling of Cognition and Implications
Goal directed behavior – collective/lower cells “listen” to and communicate based on end goal/error detection/homeostatic signal that occurs at all organizational/hierarchical layers.
Higher structure(s) direct smaller levels, this influence, is top-down and may use strategies like, bending/re-directing its paths to reduce the “energy” of “work” from lower levels in completing its goal/achieve homeostatic state.
The boundary between “self” and “world” is flexible; cells can cooperate to form larger collectives with scaled-up goals, or defect (as in cancer) to pursue smaller goals.
A “cognitive light cone” framework allows comparing diverse intelligences based on the scale of their goals in space and time.
Endless beautiful forms due to a combination of intelligence, evolution, environment and a ton of other variables can arise in ways beyond human comprehention.
Ethical and Philosophical Considerations
Understanding this allows one to reframe our assumptions about agency/cognition in biological and technological systems, with applications towards building machines that emulate this emergent complexity.
We will likely encounter diverse biological and artificial agents that challenge traditional categories of life and intelligence.
Existing frameworks in ethical and philosophical ways won’t cut it, and we need new ones, especially as technology becomes increasingly involved in influencing what life becomes in a broader biological, philosophical sense.
Existing tools of Neuroscienc, can and often do, translate into cell research. For example, many/some cells contain “memories”, including “counter-factual memory.”
Introduction and Overlapping Ideas
Levin appreciates Bach’s breadth in tackling computation, cognition, AI, and ethics. Bach finds overlap with Levin’s work on cellular communication and agency, particularly the idea that every cell can act like a neuron.
Cells as Agents and Information Processors
Every cell can send and receive multiple message types, conditionally, and can learn. Each cell is a reinforcement learning agent, primarily getting rewards from its environment.
Neurons are specialized “telegraph cells,” extending cellular communication over long distances with high-speed signals, crucial for animal movement and competition.
The brain can be viewed as a “telegraphic extension” of the body’s cellular community, with the potential for every organism to become intelligent given enough time and shared genetic destiny.
Cells posses turing complete abilities that allow them to behave with proactive control, making arbitrary internal represenatations indepedent from external influence.
Cross-Disciplinary Boundaries and Science
Levin and Bach note how disciplinary boundaries in science can be protective and hinder interdisciplinary work, particularly through peer review, limiting paradigm shifts.
An “engineering stance,” common in computer science, focuses on causal patterns and implementation, which is often missing in other fields, like neuroscience.
Critique of Current Neuroscience and Alternative Models
Current machine learning, inspired by simplified perceptrons, doesn’t accurately reflect how biological cells organize, which is from the inside out, not through external weight updates.
Local self-organization by reinforcement agents trading rewards offers a fascinating perspective, missing from current AI.
The emphasis on disciplines prevents sharing insights. for example: In a neuroscience department, it is known information processing can occur through APs before gene express. yet, this may cause surprise and resistance in a molec bio group.
RNA-Based Memory Transfer and Its Implications
Experiments (McConnell, Ungar, Glantzman) suggest memory transfer via RNA or peptides, challenging the synapse-centric view. This includes work with planaria, rats, and metamorphosing insects.
This concept raises puzzles about decoding transferred information, especially for arbitrary, non-evolutionarily relevant memories. How does a recipient brain interpret an arbitrary RNA structure?
An implication is that the Connectome Project may not be able to map concioussness if memory is transferred using methods and data different than physical axon connections.
Evolution and Planaria
In planaria, an ability of an animal’s system is to recall “previous settings”, meaning biological information for an organism persists even though massive reorganization is undertaken such as loss of an entire brain.
Competency, Goals, and Constraints
“Competency” is an engineering claim about a system’s ability to navigate a problem space toward a goal, dependent on the observer’s perspective.
Biological systems have feedback loops to reach specific outcomes in anatomical morphospace, demonstrating competencies like recruiting cells.
Goals can be emergent or explicitly represented, as seen in planaria’s bioelectric pattern memory, which can be rewritten (two heads instead of one).
Constraints satisfaction: Organisms strive to move the universe’s state space towards acceptable regions (e.g., having one head), navigating substrate and functional constraints.
Multi-Scale Competency Architecture Again
Evolution may struggle to make a genome since evolution has issues judging the “fitness” if competenct organism manage errors in-vivo using “built in software/algorithms” instead of genes, in particular with an orgamisms such as Planaria that asexually reprocduce, circumventing the normal filters, the result being a fit organism despite its genome looking horrible on paper.
Analogy: computers which has a drive prone to errors where the software corrects for mistakes in-vivo. In computer-speak: a ‘RAID setup’.
Embryonic Development and Self-Organization
An amniote embryo starts as a disk of cells, not inherently one individual, but potentially multiple. Symmetry breaking determines the organizer, leading to one embryo, or conjoined twins if disrupted.
Biological systems self-construct, determining their boundaries, structure, and relevant problem spaces, unlike pre-defined robots. They are energy-limited and must choose a “lens” to coarse-grain the world.
Planeria cells decide to follow, spatially, by gradient and other biological information cues for correct development by looking at what neighbours cells are doing: deciding “local spatial difference” cues in a cell rather than taking explicit external instrcutions, unlike an AI agent trained and designed from top down, controlled and influenced.
Implications for AGI and Collective Intelligence
Sufficient condintions include cells connected and signalling rewards with reliability over enough units.
Conditions for the emergence of general intelligence include: 1. Units as small agents with expectation of minimizing future target deviations. 2. Units connected and exchanging rewards or proxy rewards.
The question of whether these biological insights can be to be scaled up is the current task. For example: Twitter and Global-Scale social Media interaction, and testing incentive structures (for exaple with Elons’ Twiter experiement, or with societies on a social/governmental scale) can have their group agency steered, where cells/units within can become grouped into emergent control behaviours through self organization.
Twitter is explored as a potential global brain, highlighting emergent agency and the challenge of designing incentive structures for collective intelligence. This relates to governance in brains, societies, and social media.
Introduction: Natural and Artificial Intelligence Interplay
Transformative regenerative medicine requires understanding the body’s natural intelligence. This involves an interplay between understanding natural biological systems and developing artificial intelligence.
Living systems are multiscale, not just structurally, but functionally. Each level (cells, tissues, organs, organisms) solves problems with a form of “collective intelligence.”
Levin’s lab uses machine learning to understand/control biological endpoints (for medicine) and uses biology to inspire new AI architectures (non-neural).
Key Concepts: Anatomical Homeostasis and Bioelectricity
Anatomical homeostasis: Maintaining a correct body structure despite perturbations (injury, mutations, etc.). This involves *feedback loops*, not just feed-forward processes.
Current focus in biology is primarily manipulating genetics, cells and proteins (hardware). There’s much needed on *form*,*function* and controlling *decisions*.
Current focus of computer science has moved toward manipulating data. The analogy to biomedicine is this *form* can be created with a biological compiler (analogy: electrical compuational systems don’t need to rewire transistors for diff tasks; they can simply program data, we are very behind on this journey.)
Cells make decisions, and non-neural bioelectricity is a key medium for this computation, a crucial “software layer” between the genome and the body. It’s a kind of “epigenetics.”
Regenerative medecine’s final form would have to be: sit in front of computer, type in body plan you want, push go.
How do a collection of cells KNOW to produce an adult organism.
the planarian: even with hundreds of science/nature papers there still has not ever been an experiment created yet.
There’s an issue in current biology which is its very focused on the lowest level building blocks but this isn’t what we need, we need whole level organization.
The bioelectric code: Decoding this will lead to “electroceuticals” for regenerative medicine, cancer, and synthetic bioengineering.
Body tissues, like the brain, form electrical networks that make decisions about dynamic anatomy. AI/machine learning tools help target this system.
Examples and Model Systems
Morphogensis is *flexible*.
Axolotls: Regenerate limbs, eyes, jaws, spinal cords, etc. The regeneration is *context-sensitive* and *goal-directed*.
Planaria (flatworms): Extreme regeneration (any body part), “immortal” (no aging). Demonstrate flexible regeneration and the role of bioelectric “set points.”
Experiments moving frog tadpole facial features (Picasso tadpoles): Show that anatomical structures are *not hardwired*, but achieved through error minimization.
Frog tadpoles can compensate for a *variety of things*, where all tissues migrate to their correct spot even after it is not on where it supposed to be.
Frog Eye: frog is not ‘wired’, and a specific bio-electrical patter says ‘make an eye’, *anywhere*.
Planaria Heads: You can set how many heads it should create.
Bioelectric Manipulation and Control
Analogy to Thermostat: Don’t need to rewire things! Instead *manipulating electrical activity by* of rewriting “set point” information.
During evolution *size* of the thing organisms operate is flexible and scales up/down with goals. Cancer being an example where single-cellular organisms revert to small, simpler goal.
Methods: Voltage-sensitive fluorescent dyes to *visualize* bioelectric activity. Computational modeling to simulate electrical networks.
Bioloectricals do not *need* to equal *now*: It does not need to equal the *present*, it’s often stored *before* that.
Can re-create head patterns.
Manipulations: Controlling ion channels and gap junctions (like in neuroscience) using drugs, mutations, optogenetics (light). *No external electric fields*.
Single-cell level: Preventing tumor formation by restoring electrical connection to neighboring cells (overriding oncogene effects).
Organ level: Inducing ectopic (out-of-place) organs (eyes, hearts, limbs) by imposing specific bioelectrical states. Like a “subroutine call.”
Whole-body level: Controlling planarian head number (one vs. two) by altering bioelectric patterns. Can even create head shapes of *different species* without genetic changes.
It can rewrite ‘set point’ to the anatomy! which changes the *form* of an organism: The way you rewrite is a biological intervention, can use electrical-based drugs. This will provide for ‘ionoceuticals’.
Limb regeneration in frogs (non-regenerating species): A 24-hour bioelectric treatment triggers long-term leg regrowth, without further intervention.
AI’s Role in Understanding and Intervention
“Full stack” approach: Modeling transcriptional circuits, bioelectric dynamics, and large-scale patterning to derive *algorithmic* descriptions for intervention.
Betsy, is a software designed to do *circuit models* using individual cells on the tissue/anatomy to ‘simulate’ it.
AI’s role is two-fold: 1) *Inferring models* from experimental data. 2) *Inferring interventions* (which channels to target, how) based on a desired outcome.
Example: Evolutionary computation used to infer a model of planarian regeneration. The AI “guessed” a human-understandable model.
There is no model that we know so far that gives a prediction on shapes.
Problem is current model for regnenerative biomedicine would requrie that if we wanted a specific part of anatomy is make many many many mutations but a intractable reverse problem we simply don’t have solutions for.
Using *evolution* in AI to design biology *models*. *software*, it discovers model based on human understanding that could only before, only be created from a very good human mind, yet even so no models that can give prediction (e.g. with Planaria experiments on changing bio-electric field shapes, for example) have ever yet to have existed.
Software ‘Elektra’ has ability to: take database of how things *should* function, how *does* function, with all various data, use an evolutionary computation system. (in Plenaria case it got 800 inputs, where most don’t work at all, so had to infer using functional info)
The inference system gave useful information with model *without* large amounts of input data.
AI to *model*, but ALSO AI to design new interventions (how you create a new medicine).
The “code” metaphor: Genome defines the *hardware* (ion channels), but the resulting electrical network (excitable medium) has emergent properties (software), storing *patterns* and *memories*. Like a flip-flop.
Example: Editing the planarian “memory” of head number (software) *without changing the genome* (hardware). The new pattern is stable and *heritable* (through cutting).
Editing bio-patterns allows editing for: the *shape*, it also controls *growth* in adult organism too! e.g. with *Leg* (frog). This is a big example with the two heads on an Plenaria organism. It even *keeps going*: you cut up head into many *and* can use a different bio-signal *to rewrite* it, again!
Machine Learning Connections: Connecting bioelectric circuits to concepts from connectionist machine learning (pattern completion, energy minimization).
Future Vision and Conclusions
Bioelectricity: Key role of biology and it’s relation to software.
Key goal: use computer simulation, not to replace experiment *but* it tells us *which* intervention will get us the result we want.
Example: Rational design of an “electroceutical” to rescue brain defects. A model predicted a specific ion channel (hcn2) to correct the bioelectric pattern, *not* a trial-and-error screen.
“Electroceutical Design Environment”: A future system where you specify cells, tissues, and a desired pattern, and it tells you which ion channel drugs to use.
Rational design of drugs based on pattern completion *with no human trial and error* in frogs! e.g. drug hcn2 (discovered from *electrical modelling*)
The ultimate Vision of using models for biology/biomed: is that this AI system would output specific and useful outputs from AI to help guide with which *bioelectrical* *and biochemical* changes would have to take place based on all known scientific inputs (from a database).
Conclusions: Bioelectricity is a tractable “software layer” for regenerative medicine. Evolution uses electrical signaling for large-scale coordination. We can read/write pattern memories to reprogram shape. Machine learning helps infer models *and* interventions. This could revolutionize medicine and inform new AI architectures.
Two Big Outcomes: fantastic regeneration medecine AND give inspiration to design new kinds of AI that uses different principles of cognition.
Introduction: Body and Brain Plasticity
Biology has been computing at many scales long before brains. Decision-making in all body tissues is mediated by pre-neural electrical networks (not just neurons). This is an untapped frontier for new AI.
Caterpillars retain memories through metamorphosis (radical nervous system restructuring). Planaria (flatworms) regenerate entire bodies, including brains, and retain memories after head amputation, demonstrating non-brain memory storage.
Somatic Cognition and Anatomical Self-Editing
Cognition exists on a ladder, from simple to complex. Living creatures, even single-celled organisms (e.g., Lacrymaria), show complex behaviors, structural control, and physiology without brains.
Multicellular bodies retain cellular decision-making. Embryos and regenerating organisms (salamanders, planaria) demonstrate large-scale anatomical control and homeostasis, even with massive perturbations (e.g., cutting embryos, grafting tails onto limbs).
Regeneration isn’t just about starting; it’s about *stopping* at the correct anatomy. Human livers, deer antlers, and (in children) fingertips also show regeneration.
The genome doesn’t encode rote movements. Organs can remodel to reach a “normal” morphology even from incorrect starting positions.
DNA specifies proteins, *not* 3D shape directly. Cell groups “know” what to build and when to stop. A major knowledge gap exists in understanding this anatomical control.
Bioelectric Mechanisms and Goal-Directed Remodeling
Patterning control is a *closed-loop* system, not open-loop. It has error detection and correction. The goal is to target the anatomical *set point* for therapeutic purposes.
Somatic tissues form electrical networks that make decisions about *anatomy*, not behavior. These can be targeted for pattern editing.
Neurons evolved from ancient cells already performing computations. Synaptic machinery, ion channels, and neurotransmitters predate brains. This explains why fungal compounds (e.g., hallucinogens) affect human cognition – shared ancient mechanisms.
All cells (not just neurons) have ion channels and make electrical synapses. They produce electrical patterns analogous to brain activity. “Cracking the bioelectric code” is crucial for understanding development and memory.
Bioelectric patterns come in two flavors: *endogenous* (normal development) and *pathological* (e.g., early cancer signs).
Endogenous ion channels can be manipulated (genetically, pharmacologically, optogenetically) to control electrical states and network topology – *not* by applying external electrical fields.
Examples of bioelectric control:
Inducing ectopic organs (eyes in the gut) by altering electrical state.
Creating two-headed (or no-headed) planaria by manipulating gap junctions and electrical pre-patterns. This is *modular* control, not cell-by-cell micromanagement.
Changing planarian head shape to that of other species (across 150 million years of evolution) *without* genomic editing, demonstrating physiological control over anatomy.
Creating novel planarian body plans not seen in evolution, showing untapped morphospace.
Rewriting planarian “pattern memory”: creating two-headed worms from one-headed worms, *persistently* – a stable, rewritable, latent memory.
Computational models map electrical circuit dynamics to anatomical outcomes. The long-term goal: a biological “anatomical compiler”.
Regenerating frog legs (which normally don’t regenerate) by altering bioelectric state at the wound – *without* stem cells or genomic editing.
Reversing birth defects (brain malformations) in tadpoles using bioelectric simulations and targeted drug treatments to restore normal electrical patterns, even with underlying genetic mutations.
Future Directions: Regenerative Medicine and AI
The long term aim is to develope a type of an “Anatomical Compiler”, this translates a described anatomy (similar to cad) into what it is that needs to happen for it to build.
Evolution exploited electrical circuits for information processing early on. Cracking the bioelectric code has implications for regenerative medicine, synthetic bioengineering, and morphological computation.
Limitations of current machine learning may stem from focusing on brains. Non-neural architectures (cells, tissues) offer a new approach.
Other cells (bone, heart, pancreas) also show learning and adaptation. Living systems are incredibly robust to novel circumstances, a key feature for robotics and AI.
The time-scales are distinct: electrical states of relevance for this take time to stabilize, taking from minutes to days. Muscle and nerves function far quicker.
Ethical Concerns (briefly mentioned)
Concerns exist but it is lesser of all ethical concerns for medicine as of now since biolectricity deals mostly in guiding of cell behavior, unlike virus injections.
QnA
Methods to modify signals is not through directly providing electricity or frequency but though gene mod and small molecule modification.
Basal cognition is present in most early stages of living entities.
Introduction: Bioelectricity and Cancer
Cancer cells can be influenced to stop growing using bioelectricity. Forcing cells into a proper electrical state, connected to their neighbors, prevents tumor growth even with strong oncogene expression.
Cells naturally “mind meld”, identifying with a larger collective (organ) through bioelectricity, a “cognitive glue.” Cancer cells disconnect from this network.
Bioelectric Fields as Blueprints
Bioelectric fields contain blueprints of organisms. “Electric face” pre-patterns determine features like eye and mouth placement. These patterns are visible and modifiable.
Planaria voltage patterns can be altered to grow two heads. This change is heritable without DNA modification, only resetting the “goals” of the system.
Shifting from Hardware to Software
Traditional biology focuses on genes and molecules (the “hardware”). Levin focuses on the bioelectric “software.”
It is extremely complex to reverse engineer how an output came about with a complex chain, this method simplifies things to targeting those higher pattern, making engineering of a result easy.
Morphogenesis (the origin of form) is still mysterious. No individual cell knows what a finger is, but the collective does, such as a salamander which will always regrow the appropriate missing parts of the body when any part of the leg has been removed..
Memory of Forms and Evolution
Evolution searches for “pointers” into a pre-existing space of forms (a Platonic idea), not just microstates of hardware.
Even bacterial biofilms, early multicellularity, use electrical networks for communication and information integration.
All biological categorical thinking, even intelligence, about consciousness and should be eliminated, everything is scalar..
Cognitive Glue and Scaling
Bioelectricity functions as a “cognitive glue.” Individual neurons combine to form a new entity with goals and memories. Evolution used this principle long before brains.
Gap junctions are crucial. They create a “mind meld” where cells share information and lose individual identity, forming a collective.
Cells connected by gap junctions share stress. Reducing stress in one cell benefits others, creating collective concern and enlarging the “cognitive light cone.”
Bioelectric communication in development takes minutes, not milliseconds like in neural networks, is all a trade off in efficiency with regards to other tasks it has at hand.
Implications and Experiments
Cancer treatment could potentially involve resetting the goals of cells to cooperate, as demonstrated in frog models.
It might be possible to test for higher-level intelligences (e.g., ecosystems) by attempting to train them.
Evidence of larger intelligences might manifest as synchronicity – events connected by meaning, not obvious physical causality.
There exist larger intelligences (in his view) that can have different meanings and perspectives on things at different levels of scales.
Hardware, Software, and Semantics
The hardware/software distinction isn’t absolute. DNA defines hardware, but bioelectric states represent a kind of software.
Living things aren’t Von Neumann computers, but the software paradigm (top-down control, goal-directedness) is useful.
The distinction between data and machine breaks down with gap junctions. Information flow changes the physical state of the “machine.”
Semantics arises when an observer can extract adaptive meaning from events. There exist different level intelligences at differing scales that might infer different meanigns from the same observations..
Consciousness
Consciousness may exist in various body parts, not just the brain.
Consciousness may track agency – the palpated uncertainty about what to do next, and sense-making of one’s own memories.
Consciousness might be a fundamental property of the universe, not solely a product of selection.
Phenomenal Consciousness is about deciding how to interpert external or internal events in the now for your present, or future goals, that may not always be in line with each other.
Personal and Ethical Implications
Levin’s work has implications for ethics and our relationship with other beings, emphasizing a broader view of compassion.
Evolution, cognition and scale go hand in hand with considerations to compassion.
Introduction and Key Concepts
Persistence in current form is impossible, both individually and as a species; change and adaptation are inevitable. The key question is what we will be replaced *by*.
Levin works in “diverse intelligence,” aiming to understand what it means to be an intelligent, embodied agent beyond human-centric biases.
We’re good at recognizing intelligence similar to ours (medium size, speed, 3D space), but there are many other forms of intelligence throughout biology (and beyond).
Science fiction often explores possibilities beyond current limitations, offering valuable thought experiments about alternative forms of intelligence.
Our current perspective isn’t privileged; what seems like science fiction today might be commonplace tomorrow.
Humanity and Its Future
Levin is against “human chauvinism”—the idea that our current form is the only valid one. We shouldn’t fear being supplanted by superior beings.
What we value as “humanity” isn’t necessarily tied to Homo sapiens DNA or anatomy, but rather to traits like compassion and shared existential concerns.
He argues we desire human companionship on long trips due to similar levels or exceeding compassion, similar concerns and goals on large projects.
It is possible there exists highly intelligent but lacking the compasion and similar conerns as modern AI is showing that tendancy.
Limitations we perceive (e.g., lower back pain, communication bottlenecks) aren’t optimal designs, but products of evolution’s path.
Evolution’s “bow tie” architecture (compression and re-expansion, like DNA to organism, or thought to language) isn’t a flaw, but a source of adaptability and creativity. It forces interpretation and problem-solving in novel situations.
We might fear that care/compasion might not extend enough, and that fear might create resistance to ideas in transhumanism.
We often project limitations/fears of other’s intellgience that is greater than ourselves onto that of other beings.
He argues evolution is geared to have systems able to tackle novel situations as normal because every part of biology is very undependable, such as DNA to organism (morphogenesis) example.
Agency and Intelligence Beyond Biology
Levin proposes a continuum of agency, starting with very basic forms (e.g., least action principles in physics). This scales up through biology.
Intelligence is the ability to generally be a good problem solver, or able to learn well.
He reframes things such as an insect’s metamorphosis (Caterpillar to Butterfly) as evolution is the machine geared to solve novel problems and repurposing.
He doesn’t believe there’s an objective “view from nowhere” about what has a mind. It’s observer-relative, including the system itself as an observer.
Agency is an ability to interact/affect a domain using inputs from that domain.
Using an “agential lens” (considering goals, learning, memory) can be useful even for very simple systems (gene regulatory networks, sorting algorithms).
Applying behavioral science protocols (like Pavlovian conditioning) to gene networks reveals learning capabilities *without* needing gene therapy (a non-reductionist approach).
This holistic perspective doesn’t imply randomness; it’s guided by principles from higher-level systems.
Biology leverages a “multi-scale competency architecture” where higher-level goals guide lower-level actions, making control *easier* despite increased complexity. This is “engineering with agential materials.”
Hollistic, behaviour shaping and the normal interface is hijacked in Biology through it’s various encapsulated trainable and behavour systems.
As scientists, and philosophers, it is our jobs to widen/open our understanding of intelligence and to improve our models (i.e. metaphors).
Implications for Understanding Systems
Current AI (LLMs) may *not* be truly agentic, but he emphasizes *humility*: our intuitions are often wrong, even for very simple systems.
A “being” with embodiment (the ability to generally solve problems, measure against a perference and acting on the environment in a loop) *may* be applicable in domains that is non-biological and virtual worlds are not that different than “real” experiences..
Intelligence *isn’t* limited to brains or the physical world. Embodiment is about the perception-prediction-action loop, which can occur in various “spaces” (physiological, transcriptional, etc.).
Being part of a larger collective intelligence doesn’t guarantee individual well-being (e.g., skin cells sacrificed during rock climbing). The composite system’s goals may differ from its parts’.
Cancer can be viewed as cells shrinking their “cognitive light cone” (the scope of their goals) back to a single-cell level. Treatment can focus on re-expanding this connection, not just killing cells.
LLMs can model/talk impressively but that’s all it currently has been observed to do as the current system, LLMs, lacks the systems necessary to have proper agency.
That there exists degrees, or stages to reach the state of agency and to scale higher for complex intelligence systems.
Concluding Thoughts
The idea of fixed categories is of question, even humans may be an aspect of an overall collective intelligence.
There are multiple levels of agentic goal seeking behaviour such as Clam, single-celled, cancers surviving very well or even just regular cancers defacting a higher level systems.
Developing the field of “diverse intelligence” is crucial, as we’re creating large-scale, emergent cognitive systems (social structures, IoT) with unpredictable goals. We must get better at understanding, shaping, and communicating with them.
It would be foolish to place bounds on “how much” intelligence because it cannot be calculated from the vantage of a less capable agent/system.
Evolution doesn’t aim for specific solutions, it’s geared up systems which enables problem solvers to exist to be flexible in new environments.
We have to make up the terms to understand and model a phenomenon (aka a metaphor) as necessary.
Being humble of what we claim systems cannot do/or is doing because what we might be perciving it doing might be different, with evolution/morphogensis example, the defult assumption should be towards the more capable side.
The system is agentic, or not, will be dependent on your point of view (a very clear, non-abstract example: the physics perspective (physics guy)).
Introduction: The Nature of Perception and Reality
Current cognitive science suggests we construct perceptions in real time. Evolution suggests our perceptions don’t reflect true reality, but are optimized for survival, challenging the idea that space-time is fundamental.
Projects and Perspectives
Levin’s Lab: Focuses on understanding collective intelligence, communication with cellular swarms, and exploring competencies of gene regulatory networks and other models. Applications include cancer, birth defects, and regenerative medicine.
Hoffman’s Project: Investigates structures beyond space-time (like the amplituhedron and decorated permutations) and aims to derive these from a dynamical system based on Markov chains of conscious agents. He proposes linking properties of conscious agent dynamics (e.g., mass, momentum, spin) to physical properties. He aims to predict parton distributions inside protons from this theory.
Levin’s Childhood Anecdote: A lost toy prompted thoughts about the persistence of objects and self across time, leading to questions about stability, change, and the reconstruction of self from memory traces.
The Nature of Time and Self
The self needs constant reconstruction. Memories are accessed and rebuilt, not statically stored. The “space-time loaf” is a framework, but physicists are questioning its fundamental nature.
Space-time may be a “data structure” that breaks down at very small scales (10^-33 cm, 10^-43 s), suggesting a deeper reality beyond it.
Time and space could be “interface concepts” — useful tools, but not reflecting the ultimate reality, which exists outside space-time. Evolution by natural selection supports this; our perceptions are optimized for survival, not truth.
Neurobiology is far more complex than previously thought, as it needs to be reverse-engineered from this deeper reality beyond space-time.
In the conscious agent model, the “arrow of time” may be an artifact of the projection process (loss of information), not inherent in the fundamental dynamics.
Agents build models of the world. Those might or not match the model of the world.
Agency, Tools, and Perception
The tools we use (e.g., voltmeters, rulers) limit what we perceive to low-agency, mechanistic phenomena. Minds, however, are good at detecting other minds, suggesting the need for different “tools” to understand agency and higher-level properties.
Objects, including our bodies and brains, are not fundamentally conscious. Consciousness is a *perception* recreated “on the fly” as needed, similar to an avatar in a virtual reality.
Object permanence (the belief that objects exist even when unseen) is a deeply ingrained, pre-rational belief, programmed early in life, making it difficult to fully accept the non-fundamentality of objects.
Levin presented the Skydive memory question: You can have experience, or remeber having, the former group is the ephemeral types while the latter group are of the people for which, only the collection of their memory are considered important to their self.
Implications and Analogies
Analogy: VR, just like how avatars and objects are not “real” inside a virtual world, neither exist except that moment when rendered. Analogously there may be some outside computer of the VR. Agreement on perception (like seeing a dinosaur bone) doesn’t imply pre-existing reality; it reflects shared interpretation within a “headset.”
Split-brain patients: Different hemispheres can hold contradictory beliefs (e.g., atheist vs. theist), raising questions about the unity of self and negotiation between internal “agents.”
The distinction between living and non-living, or conscious and unconscious, may be an artifact of our interface, not a fundamental distinction in reality. (Analogy: conscious vs. unconscious pixels on a Zoom screen).
A thought experiment describes a patient with a disruptive “partier” personality facing integration therapy, highlighting the ethical complexities of altering or erasing a coherent “self,” even within a single body. The key question: where does the conscious go after integrating.
Conscious Agents and the Fundamental Nature of Reality
Conscious agent model: A mathematically precise model where agents have experiences, take actions affecting other agents, and interact through a Markov chain.
There is potentially ONE final agent (infinite) so it cannot be described in practice. So any analysis must be limited.
Observation: In the model, consciousnesses observe other consciousnesses. Any collection of conscious agents is itself a bigger agent.
Little Agent and Trace Chain: by picking only subsets and some finite temporal windows of these collection we create little agents. Little agent perspective.
Consciousness, Cognition, and Biology (Levin’s Perspective)
Cognition is different from consciousness. Levin focuses on cognition, defining it extensively in a recent paper (detained paper). He believes cognition is prevalent throughout the universe, not limited to “living” things.
His framework, TAME (Technological Approach to Mind Everywhere), is an engineering approach focusing on the practical benefits of a “cognitive lens.” It aims to generate new experiments and interventions in biomedicine.
Bioelectricity is a key “cognitive glue” for scaling up cognition, but not the only mechanism. Levin advocates dropping “teleophobia” (fear of attributing goals) to unlock new research programs.
Every cell is some other cell’s environment. And at every biological scale, you see this ‘part within the whole’, that makes drawing the limit of one self hard and challenging.
All of our bodies (or maybe even our environment!) all are nested, various collections with organs having their own intelligences/cognitions,
If your “cognitive lens” were, for instance, tracking chemicals instead of visual light, then you might “perceive” and “understand” your internal organ systems and the liver might also have intelligence.
Hoffman’s responses and additions.
Distinguishing between Subject and Objective Reality: He proposes that we understand object (external reality independent of observers), and there are objects dependent on the perception. His work point to objective reality being different.
Monodology, and Leibnizian, views are a very good match. In Leibnizian worldview, experiences and probabilistic relationship among those experiences, very similar to his model’s perspective of reality.
Further Discussions and Q&A Highlights
Fitness vs. Truth: Perceiving the truth (e.g., toggling millions of voltages) would lead to *less* fitness than perceiving a simplified “interface” (e.g., steering wheel, gas pedal) in a virtual reality game. Theorems, not just simulations, show that fitness payoff functions generally contain *no* information about objective reality’s structures. This implies optimization in VR reality will lead to obscuring most information, including truth.
Evolutionary Game Theory (EGT): Critics argue that using EGT to disprove the fundamentality of physical objects is self-contradictory. Hoffman counters that *all* scientific theories make assumptions, and good theories reveal the *limits* of those assumptions (like Einstein’s theory revealing the limits of space-time). The work being done, is precisely in the spirit of EGT, with the objective to highlight that there is potentially some underlying realities behind our models/reality.
Space time isn’t fundamental, neither is quantum.
The hard problem of consciousness isn’t one for Hoffman. Hoffman starts with assuming conscious is all.
Bayesian Inference and Markov Blankets: Perception as Bayesian inference doesn’t imply perceiving the “true” state of the world. The conditional independence imposed by a Markov blanket shows that we only access an *interface,* not the underlying reality.
Hoffman acknowledges Delbrück’s similar non-reductionist viewpoint on evolution.
Relationship between consciousness and high energy physics is, is communicating or collaborative (trace chains), that it has properties correlating and communicating with our real (or rather rendered) world objects like photons and so forth.
Nima Arkani-Hamed (high-energy physicist): Hoffman recommends his 2019 Harvard lectures for a (relatively) accessible introduction to the amplituhedron and related concepts.
Physics perspective of Don and Leven were not quite aligned. Hoffman works at physics being secondary and being derived by concious, and a conscious being the ultimate and first truth, while Levin sees cognition very similar, but believes that both are still important perspectives, with intelligence spanning multiple hierarchies/agents. Levin had functional framework it’s called t-a-m-e.
Introduction and Mutual Respect
Michael Levin admires Karl Friston and Chris Fields for their interdisciplinary expertise and the density of new ideas in their work.
Chris Fields admires Levin and Friston for combining deep theoretical understanding with practical applications in biology and mental health.
Karl Friston admires Levin and Fields for their rapid, out-of-the-box thinking, approaching problems from first principles and challenging conventional assumptions. He also specifically admires Fields understanding of Quantum concepts.
Core Concepts and Questions
The discussion starts by examining what each participant finds unique and valuable in the others’ research.
The concept of “babbling” is introduced. Babbling, normally applied to child language development, is generalized to describe exploratory, seemingly random activity used to build internal models of the world. Examples: infant’s rattle, vacuum fluctuations.
Scaling of Cognition. Question of where cognition *begins*. Is there a zero point on the spectrum of cognition/awareness? A central question: What is the overlap between “alive” and “cognitive”?
Nature of “self”. Participants view “self” as a *construct*, a useful “fantasy” or hypothesis generated by systems to distinguish the consequences of their actions from those of others.
Emphasis on the need of philosophy of mind and philosophy of psychology is derived. Why we need it and think it and study.
Hard Problem, Meta-Hard Problem, and Awareness
The “hard problem” of consciousness (how physical processes give rise to subjective experience) is discussed, with a focus on the “meta-hard problem” (why do we find consciousness so puzzling?).
Friston, citing Clark and Chalmers, considers meta-hard probem to be where questions and inquiries should be focuses.
Friston argues the meta-hard problem arises from our capacity for *counterfactual thinking* – imagining alternatives to reality. This capacity requires sophisticated internal models.
Counterfactual example given: Asking oneself ‘Why am I conscious?’ logically presupposes and alternatrive hypothesis ‘…that I’m not conscious.’
Conciousness can be called ‘awareness’, since consciousness can imply the need of ‘self-consciousness’ – but this may confuse readers.
Awareness. Chris Fields suggests a view that cognition necessitates *some* form of awareness (though not necessarily *self*-awareness).
Debate about whether systems capable of *planning* must represent multiple counterfactuals (like a cheetah chasing a gazelle – does it consider failure?). Friston suggests sophisticated planning, beyond reflex, necessitates multiple counterfactual representations, example: Saccadic eye movements; and multiple futures being an option for planning as the difference between reflexes and planning itself.
Counterfactuals allow actions for selecting outcomes, in essence.
Levin and Friston question the *form* an answer to the hard problem would take (number, equation, poem, art?). This is a challenge.
Complexity threshold concept discussed with relation to conciousness: Some complex math must ‘suddenly’ happen in the complex of a process.
Chris Fields highlights the role of *assumptions about physics* in the hard problem. If we assume the physical world lacks *any* cognitive properties, we are forced to postulate some “magical” emergence of awareness with increased complexity.
Integrated Information Theory is also examined, proposing it could be feedback loop, etc.
Implications for understanding Larger Embedded Systems
Our human condition could possibly be viewed through our multi-layered composition: We are parts comprising wholes. Example is neurons-within-neuron-networks as subcomponent parts, a model, being either in: mechanical vs agentic (goal oriented) universe.
In the ‘mechanical vs agentic universe example, the network learns on its own via a non-nuetral environemtn vs learning due to a specific environment – due to this, we must question how many agents there exist.
Systems beyond an individuals could contain unpredicatble cognative capacity, implying that all systems of sufficent feedback are intellegent, since feedback loop, if existing at all, means a base candidate for ‘aware-system’.
The Role of Babbling and Uncertainty
From our ‘selves’, inside out: A human self inside out to it’s environent: to answer of cause for effect, is: was it ‘us’ doing something; or the ‘world’ did, by answering how ‘the world’ is, an effect of us. So one gets immediately to this kind of babbling scenario where it involves taking stock of its current environment and it’s possible influence over.
“Babbling all the way down” suggests exploration is fundamental even at seemingly “random” levels, like quantum vacuum fluctuations (proposed by Chris Fields).
Babbling does not necessitate ending, according to their logic and views. The purpose, and therefore the concept of ending ‘babbling’, could be viewed to never happen or is a permanent state of being: because of a consistent exploration.
A self must be tested but also present, as it must ‘test against a world outside’ (or at least seemingly so) for cause and effect from ‘us’ (our inside) and ‘it’, the world and outside of the organism’s being.
From our environment: the question and view can occur such as in: Conjoined twins, and the study of embryonic blastodisks can and will create different individuals within that mass, since those who compose, via seperation and/or differentiation. An example, due to close proximity to development, each twin will share different cells, resulting to each cells will be interpreted to be apart of the two different bodies (even if one is to be differentiated differently from the other, that effect may ‘linger on’, as example). The model each will make to understand reality will differentiate from each other.
Such that any ‘new environtment’, be that to the whole collective cell mass and even just an indivudual cell will explore (like ‘motor-babbling), constantly adjusting to interpret its place. An example could include even transcribing of new genes, a kind of expression ‘going on, always’ for that unit. This could be interpreted and ‘wigglings’ and probing, an interaction of and exploration within its specific space (space meaning whatever boundary it comprises with relation to whatever it’s ‘environent’ is – be it it’s internal self, of that collection to its ‘new world’).
Uncertainty: Self-organization fundamentally seeks to *reduce uncertainty and ambiguity*, achieving predictability in interactions (Friston’s argument, linking back to earlier discussions).
This explains and answers their concept of why things stay and persist the way they do and not chaotically dissipate immediately, due to resolving this ambiguous nature via predictability (at least, our definition of predictable, meaning it’s existence over time of itself).
Feedback Loops: Feedback Loops create these boundaries: between, it’s definition to outside and that boundaries existance, including but not limited to, the unit itself. A disruption of this loop creates different ‘bounderies’ to explore.
Time, Discreteness, and Cognition
Friston proposes a distinction between systems with *continuous-time* dynamics (like thermostats, chemotaxis) and those with *discrete-time* representations, suggesting the latter are necessary for planning and thus cognition.
The idea is a *very fast changing dynamics*, in it’s most simplistic view.
Friston brings us up example of the human eye having blinks per certain duration to denote the kind of ‘updating of belief of a system must have’. Such discreteness and characteristic that arises of its unit, defines an ‘existance over time’.
Time constants (around 300 milliseconds for humans, based on saccades, phoneme processing, etc.) might indicate a threshold for “cognitive” systems (admittedly egocentric view, according to Friston).
It also brings the importance of oscillation and cyclical process within system as integral to reduce uncertainty via ‘certain pathways it travels over’, that makes things persist over time due to those repition of patterns, reducing chaos.
The constant is important because that specific time constant corresponds with the idea of time-constant in a conscious ‘agent’ or subject.
Fields connects the idea of babbling to fluctuations in the *quantum vacuum*, suggesting it could be seen as the field exploring its environment. This raises questions about the meaning of “randomness” from a Bayesian perspective.
Discussion of whether the underlying reality is fundamentally *discrete* (as in quantum mechanics) and whether our continuous models are approximations that obscure inherent “mindfulness.”
Physics, Biology, Psychology: Unified View
Strong agreement among participants that these disciplines are, at a fundamental level, studying the *same underlying principles* of self-organization, information processing, and uncertainty reduction.
Resistence in the academic establishment. Resistance among Levin’s and Fields’ peers to applying cognitive concepts (goals, planning, etc.) *outside* of traditional neuroscience/psychology. They consider it “heresy.”
Friston believes the concepts described between the scientists here are very influential and cited multiple times, to such a level that he can picture this line of ‘philosophical-psycology’ and cognitive approaches can lead a change to a new main-stay standard.
One of their personal “heresy,” among Levin, Fields and Friston, is their understanding of ‘what comprises a god’, such that a simple understanding of what makes of a ‘mind of a higher being, from its compositon.’
Affirmation that *all science starts with an act of faith* – the belief that the world is fundamentally understandable (Michael Levin’s point).
Existential Implications
The discussion touches on the *destabilizing* implications of recognizing the “self” as a construct, potentially leading to existential crises.
Importance and integral necessity to ‘make belief’ what is real via these kind of philsophical questioning. Example being what they discuss is in a very ‘basic fact, a reality in it of itself’ via biology and life, existing from it’s basic constituents, growing. A fact in reality that can be taken for granted as basic.
Existential Uncernityt can derive from these kinds of pondering; via questioning oneself, etc – thus par for the course to occur, to deal is just simply acceptance and comfort.
Ways to handle existential implications can stem to those destabilizations via ‘true belief’, an acceptance of one kind reality as the self: example can be one should find that, to those who consider questioning ‘is something less valuable due to that concept in a real understanding?’ Should just as equally hold value and wonder because of such mechanism.
Friston and Levin stress that recognizing the “self” as a construct does not *diminish* its importance or the wonder of experience. It simply reframes the *mechanisms* underlying these phenomena.
Different levels and forms and paths of dealing, even simply *experiencing*, will help grow for oneself on these destabilizations. Even via usage of medication/psychadelic medicine: those are tools for people, a means for better growth of understanding of these matters, that we must have that ‘acceptance of all realities: as being itself of the ‘real’ reality.’
Those suffering the consequence of pathological thinking and anxiety over our discussed concepts (over our reality and implications) has no fundamental and immediate resolution due to such hypothesis of one’s questioning; over existence itself and ‘what does any and everything exist, for at all.’
Reasons can go for anything. This implies our constant struggle with those questionings and the possible ‘pathways for answer of said-question’ of self-thinking in one’s mind; as integral as the concepts for such question and pondering themselves.
The inability to resolve inherent uncertainties (such as the dual hypothesis “I exist / I don’t exist”) can lead to pathological consequences (anxiety, allostatic load) (Friston’s explanation from a clinical perspective).
“
Introduction: Turing, Morphogenesis, and Collective Intelligence
Alan Turing, known for AI, also studied morphogenesis (pattern formation), seeing a deep connection between intelligence and development. All intelligence is collective intelligence, as every cognitive system is made of parts.
Developmental biology is crucial: We all start as a single cell and gradually become complex cognitive systems. Embryogenesis is the process of transitioning from “physics” to “mind.” There’s no sharp dividing line.
Multi-scale competency architecture: Biology uses a nested set of problem-solvers at different levels (cells, tissues, organs, organisms), each with its own goals and competencies.
Goal-directedness is key: A powerful way to recognize, communicate with, and build unconventional agents is by understanding their goals, even if they are very different from our own. Navigation policies in diverse problem-spaces (not just 3D physical space).
Cognitive boundary model: A framework for understanding the scaling of cognition based on the size and complexity of goals a system can maintain.
Morphogenesis as collective intelligence: Cells collectively navigate “morphospace” (the space of possible anatomical forms) using bioelectrical networks as a communication mechanism.
Practical, empirical predictions: These conceptual ideas lead to testable predictions and have led to discoveries in regenerative medicine and bioengineering.
Single cells (e.g., Lacrymaria, Acetabularia) show impressive spatial and behavioral competencies, challenging the notion of cells as simple building blocks.
Slime molds (Physarum) demonstrate decision-making, e.g., choosing the larger mass, integrating sensory information through biomechanical sensing (like sonar).
Morphology and behavior are intertwined, highlighting a transition from solving problems in morphological space to behavioral space during evolution.
Expanding the Concept of “Agent”
The traditional view of distinct “natural kinds” (human, rat, etc.) needs revision. Evolutionary and developmental processes show continuous transitions.
Caterpillar metamorphosis: The brain is largely rebuilt, yet memories can persist. This raises questions about memory storage and what it’s like to *become* a different kind of agent.
Planarian regeneration: Cut worms regrow missing parts, including the brain. Memory (e.g., learned patterns) can be stored *outside* the brain and reimprinted on the new brain.
Tadpole plasticity: An eye induced on the tail can provide visual input to the brain, even though the eye doesn’t connect directly to the brain. This shows incredible plasticity and adaptability, even in adults.
Multi-Scale Competency Architecture: Nested Problem Solvers
Nested hierarchy of problem-solvers: Not just whole organisms, but *all* components (cells, tissues, etc.) solve problems in different spaces (transcriptional, anatomical, physiological).
Competency in unfamiliar spaces: We easily recognize intelligence in 3D behavioral space, but we struggle to recognize it in other spaces (e.g., physiological) due to lack of training data.
IQ test for recognizing intelligence: Assessing another system’s IQ tests *your* ability to recognize the relevant space and goals.
TAME Framework, Technological Apprach to Mind Everywhere: a way to evaluate other intelligence whether it be evolution-based or synthesized.
Morphogenesis as Collective Intelligence: A Case Study
Anatomical order, how collection of cell forms anatomical complexity with fidelity, and the genome not explicitly showing this high level morphology.
Anatomical compiler (long-term goal): Translating a desired anatomical form into stimuli that guide cells to build it, enabling regenerative medicine and bioengineering.
Software vs. hardware: We’re good at manipulating molecular hardware (genes, pathways), but far from controlling large-scale structure. We need to understand the “software.”
Homeostatic process around target morphlogy. Biolelectrity is one mechanism which cells use. Changing the bioelectric setpoint rather than micromanaging the system.
Bioelectricity as an Informational Medium
The memory usecase of Biolelectricity, for example to remember a set point of anatomical morphology to reach when regrowing in planarians, even remembering the *number* of heads!
Bioelectrical networks: Cells communicate via ion channels and gap junctions, creating electrical networks that store and process information, similar to (but not identical to) neural networks.
Neural decoding outside the brain: The same concepts and tools used in neuroscience can be applied to non-neural electrical tissues, offering a “neuroscience outside the brain.”
“Electric face”: Early frog embryos show a bioelectrical pre-pattern of the future face, long before genes for specific features turn on.
Optogenetics and gap junction control: Tools to manipulate electrical states and cell communication, allowing functional experiments to test the role of bioelectricity.
Engineering anatomical structures is more than simply changing genes. By targeting bioelectricity (using sub-routine), organs are regrown (tadpole leg regrowth)
Scaling Cognition and Goal-Directedness
Scaling of goals: Individual cells have simple goals; cell collectives (through bioelectric networks) pursue larger, more complex goals. Goal size defines level.
Glioblastoma: Cancer cells disconnect electrically, reverting to a unicellular lifestyle with smaller “self.” Maintaining electrical connections can normalize tumor growth.
Cognitive “glue”: Bioelectrical networks bind cells together, enabling them to navigate morphospace as a collective.
Measuring Goal Size to measure intelligence of all kinds. Goal persuit persistance when interrupted defines agenthood. Bending spaces by creating simple constraints from high to low levels.
Xenobots: Engineered by Subtraction
Evolution pivots problems accross dimensions like behavorial vs morphological.
Xenobots: Skin cells, isolated from a frog embryo, spontaneously self-organize into novel organisms with emergent behaviors (movement, pile-making). No brain, no neurons.
Kinematic self-replication: Xenobots build copies of themselves from loose cells – a behavior discovered by AI (Josh Bongard’s lab), and *not* found in frogs.
Engineered by Subtraction” what cell behaviour appears if freed, so their full behavioural capability comes from both constraints and lack thereof, the “default”.
Evolution creates machines with broader capabilities for generalized problem solving, than simply for just fulfilling a very narrow set of use-cases.
The ability to self-assemble shows potential to be agents or problem-solvers in other non-standard evolutionary usecases, posing many societal and safety implications for ethics and decision making.
Other Notes
Survivorship Bias can easily occur due to our ability to engineer novel constructs which are not constrained by evolutionary history, survival, or adaptability.
The electrical pattern or set point in an organism may adaptively, be shaped over a wide range of acceptable changes to the ‘settings’, so changes to the organism’s electrical signaling or ion channels can create different results (making for great interface), as cells respond according to the new voltage state.
Agency Claims are claims on how an organism or collection could act, not the specific low-level mechanism.
A multi-scale agent (one agent inside a bigger one inside etc…) have a spectrum of interfaces to manipulate to persaude behaviour, with gap junctions helping to meld boundaries to coordinate (the higher scales imposing simple contraints such as by guiding and making easy following down energy grandients) or be interfered with for persuasion by setting constraints.
The evolution to new intelligence or agent starts as simple single celled goals and then to networking for multiple ones. The networks enable a scaling of not only intelligence but more shared memory between units and prediction capability, along with its homeostatic ability that keeps on growing, and error tolerance through adaptability.
Introduction: Beyond the IKEA Blueprint
Traditional view: DNA as a “blueprint” (like IKEA instructions) for building organisms. Proteins are like builders and you get the thing.
Reality: DNA provides a recipe for proteins (low-level parts), *not* a direct blueprint for the organism’s shape (morphology). It all works by complex systems.
Analogy is describing metal, screws, or the parts list, but not a bookshelf (shape is emergent, it comes out from the bottom, from complexity and systems, there are no detailed instruction manuals).
Analogy: Not instructions, but complex hardware (like a circuit that harnesses phsyical law) making softwate via dynamic bottom-up process of emmergence from many smaller parts with minimal rules/behaviors.
In principle, with perfect simulation, we could derive shape from DNA + environment, but that’s not how biology *works*.
The Picasso Tadpole Experiment
Experiment: Rearranged facial features of tadpoles (eyes, jaws, etc., in wrong places).
Result: Features moved *correctly* to form a normal frog face. The end shape came and it worked.
Implication: Genome doesn’t encode a hard-wired sequence of movements. Instead, it creates a system that *reduces error* towards a “target morphology” (the ‘remembered’ correct frog face).
Analogy: Marble dropping (gravity), you dont need top-down direct instuction. You just drop. The complex marble sorter has bottom up dynamics to get it in the right place.
The system “self-corrects,” It works continuously to correct a certain remembered pattern, error reduces.
Teleology and Goal-Directedness
Teleophobia” in biology: Resistance to describing biological systems as having goals.
Levin: Goal-directedness is *essential* for understanding how biological systems function and evolve.
Not just day-to-day. It’s at every stage (including during Evolution itself, even when deveolping, or any change/process really).
We should be studying the goals that emerge, even as a practical empirical issue!
Friston’s Free Energy Principle: Organisms predict their environment *and themselves*, minimizing “surprise” (difference between prediction and reality).
Multilevel (multiple tiers and hierarchies) Systems with goals! All reducing the delta and making prediction/goal.
Intentional Stance (Dennett): Assigning “agency” is a *practical* tool for prediction and control.
Principle of Least Action (Physics): Systems “minimize effort.” Analogous to biology: Find the most *efficient* level of description and intervention, not always the lowest.
Example: No quantum, lowest possible, no one actually thinks like that. The system will act to go the lowest cost of path, it will do it, it must be at any of these levels, at the bottom! (bottom = particles or atons and then emergent systems from there)
Robotics Analogy: Cuckoo Clocks must be *rewired*; advanced intelligences are easier to *persuade* with stimuli and experiences instead of adjusting its atom (and so they also lack feedback loops and are quite limited/not flexible!). This can inform bioengeneering!
Multi-Scale Systems and “Downward Causation”
Cells follow local rules, BUT higher levels (tissue, organ) perform computations that “deform the option space” for cells.
Analogy: Bending space. There is no “magic” at lower levels of computation but it can come together (complexity) to work at another layer, from higher layers, where another ’emergent’ level makes computations to do its thing too!
Higher-level “goals” guide lower-level processes towards a global body plan (like water flowing downhill, “minimize effort/cost”).
Swarm Intelligence: Group agent has a rough pattern-memeory that “remembers” the large pattern it’s creating.
Planaria example (next section for details).
Planaria: Memory Beyond the Genome
Planaria: Flatworms that regenerate any body part, do not age.
Two-Headed Planaria: Bioelectric circuit stores “pattern memory.” This can be *reprogrammed* to create two-headed worms *without changing the genome*.
Regenerate into perfect tiny worms, no more no less (can be even 255 parts, and still do this!)
They’re IMMORTAL. So aging itself isnt inevitable and does not age by thermodynamics limits (we get new ones when they split in half, basically splitting the old “ancient” version).
Somatic Inheritance: Mutations accumulate in planaria, but regeneration remains precise. Challenges “DNA-centric” view.
Imperfect DNA, but very perfect bodies because of top down, multiscale layers with pattern correction.
If we change the pattern, it will just generate its version, no matter if DNA changes or not, they will still form (if 2 heads it will). The software changed, the memory of bioelectrical circuit and the large memory.
We can even ‘change heads’ by interupting the normal pattern with electricity.
Bioelectricity and the “Anatomical Compiler”
We can turn these “gates” off and on to build (by turning the ion channel circuits, protein batteries that store and create memories and decisions!) things! It works similar to a digital circuits “ram” but instead as biological organisms.
The bioleectric information is similar to a memory circuit that needs voltage (RAM or the circuit, it needs current!), and the information is saved as electrical memories on top of DNA.
There is hardware that encodes all things, turn it on, it will show what’s default, which gets tuned by evolutions to create all living creatures, but, with “goals”, its actually quite reporogrammable!
Vision: “Anatomical Compiler”: Draw a desired organism; the system translates that into stimuli to guide cells to build it (not by 3D printing, but by rewriting cellular “goals”).
Change the voltage of the bioelectric memory so they get changed! We’re still trying to understand how/when, but when there is stability.
Self, Boundaries, and Cancer
Gap Junctions: Direct connections between cells (proto-synapses). Share information, erasing “ownership” (like a “mind meld”).
Origin of “Self”: Merging of individual cell “minds” into a larger, compound self.
Difficult to make ownership (there are signals shared, with no ‘ownership’ metadata that we often give value and importance to!), it’s not clear “who owns” these changes of values, making this a type of super telepathy. It merges.
Analogy: A cell becoming large to create large cognition and larger (think about time and space now in larger terms, it grows).
Cancer: Breakdown of gap junction communication, cells revert to a unicellular, “selfish” state (metastasis).
Simulation analogy: We simulate “prisoner’s dilemma”. Cooperation becomes default, with no one able to cheat. However, the cell can become separate, and only thinks as singular cells (go, proliferate).
Possible cancer treatment: Convincing cancer cells to rejoin the collective, restoring multicellular communication.
Metastatic Malanoma: we prevented electrical communicational cells, it just reverted back and start doing cancerous/malfunctions of errors and non pattern correcting.
We ought to reverse it by using things like ion-channel to correct/adjust voltage! Rejoin the “pattern”, with borg-style hivemind, we fix our cell.
Robotics, AI, and Future Implications
Multiscale Intelligence: need an intelligence on par to humans, something a kin to a ‘human collective’.
Collaboration, not competition/cheating. But make each individual count (keep its uniqueness in terms of skills). We need better control over things to know more about these different organisms.
Why we use the worm? The body-blueprint isnt even in the body plan (how is this possible?). Where the body plan be in biology, where in it, what the ‘program’ is to work from? We use them becauase their genomes are MESSED, very messy! Yet it still produces consistent, almost exact copies that’s stable 100%.
No matter how bad the mutations or messy its genes/chromosome/DNA become! So there must be something “other” than this for pattern corrections and morphogenisis and development.
Multi-scale robotics: where robots arent fully following simple programming rules. This allows greater creativity, adaptation, resilience.
Biological insights inform robotics: “Robots don’t get cancer” (because parts lack sub-goals, lacking a feedback system). Explore “info-taxes” (constant search for information).
Synthetic Biology: Creating novel organisms by altering cell interactions and environments (“endless forms most beautiful”).
Endless Forms: endless possibility and variety, and it may end up even outside “darwin’s wild imaginations” by being synthetic (we change them manually in the ‘goal’, not the “dna level). We can give normal cellls new changes/tasks!
Exotic possibilities: can be part “evolved” (maybe not even organic/biochemical biology, it could use virutal-bio). It can range from bioengeneering-based robotics, household things with machien leanring and etc etc.
New Types of Life: it blurs all words. It’ll blur all human distinctions of words with things like humanoids, human brains/biology.
A cell on the tail, eye on tails, etc (tadpole), will give it “vision” with “normal” body structure, yet working in perfect order, showing it will do/function! Even by its genomic standard!
If we design somtehing to act like an organisms. All words “break down” here, and we should update these labels in better definitions with new terms.
Ethical Considerations: Re-evaluating definitions of “machine,” “organism,” “self.” What are our obligations to different types of agents?
No good definitions of anything, yet it opens “moral, and philosophical question that is massive:
what happens if my brain connects with computer and the computre connects with mine?”. Or other body parts being artificial, “how will this function”. Will it blur or not change us at all. How much it matters: a vacuum, or implant and the split/combination ratio.
Introduction: Beyond Genetics
Levin’s group studies how evolution uses a “multiscale competency architecture” and focuses on the “software” level (bioelectricity) rather than just the “hardware” (DNA).
Biology utilizes “agential materials”: Cells have inherent goals and problem-solving abilities, not just passive components.
Dynamic, robust anatomical homeostasis is a form of cellular collective intelligence, problem-solving in “morphospace.”
Developmental bioelectricity is a key “cognitive glue” coordinating cells to achieve large-scale anatomical outcomes.
The “Anatomical Compiler” – A Regenerative Medicine Goal
The long-term goal is an “anatomical compiler”: Software to translate a desired anatomical design into stimuli that guide cells to build it.
This would solve major medical problems like birth defects, injury, cancer, and aging by controlling cell group construction.
Current limitations: We lack models to predict anatomy from genomes alone, even in chimeric embryos (e.g., axolotl-frog hybrids).
Medicine focuses on “hardware” (genetics, pathways); understanding “software” (cellular decision-making) is crucial, especially in novel situations.
Cellular Intelligence and Robust Development
Embryogenesis isn’t hardwired: Cutting an embryo in half yields two normal individuals, showing adaptability.
Cells adjust to internal changes: Kidney tubule cells adapt to varying sizes, maintaining correct lumen diameter through different mechanisms.
Regeneration demonstrates anatomical homeostasis: A salamander limb regrows to the correct size and stops.
“Picasso tadpoles” with misplaced facial features develop into mostly normal frogs, demonstrating error correction.
Cells form networks, scaling up homeostatic loops. Single-cell goals (pH, hunger) expand to tissue/organ goals (limb length, finger count). Cancer involves cells reverting to primitive, single-cell goals.
Bioelectricity: The Cellular Communication Network
Cells use bioelectricity, like brains, but predating nervous systems: Ion channels create voltage potentials; electrical synapses (gap junctions) facilitate communication.
Early embryos use electrical networks to guide body plan development in “morphospace.”
Levin’s group manipulates bioelectricity *without* external fields or electrodes, but through ion channel modulation (optogenetics, drugs, mutations) and gap junction control.
Cancer cells disconnect electrically from neighbors. Forcing connection via ion channel expression can prevent tumor formation despite oncogene presence.
The “electric face” of frog embryos prefigures facial structure, showing a bioelectric memory of the correct form. Disrupting it causes mispatterning.
Ectopic organs (eyes, legs) can be induced by rewriting the bioelectric pattern in cells, acting as a modular “subroutine call.” Cells also recruit neighbors.
Brain Defect Repair and Planaria’s “Memories”
A bioelectric pre-pattern determines early brain shape. Computational models guide ion channel manipulation (e.g., hcn2) to restore the pattern and correct brain defects caused by teratogens.
Planaria (flatworms) regenerate any body part.
Bioelectric patterns store a stable “memory” of the number of heads. Altering this (without genetic changes) creates two-headed worms.
This altered body plan is stable through multiple cuttings, demonstrating a non-genetic, rewritable, long-term memory, analogous to incepting false memories.
Planaria’s bioelectric circuits have multiple stable states. Work is being undertaken on the State Space and the merging formalizations between Electrical Cirtuits, Formalisms and Connectionist Architechtures
Machine learning helps infer circuits and interventions for bioelectric control.
Planaria can be induced to grow heads of *other* species by manipulating electrical communication, demonstrating plasticity beyond their genome’s usual expression.
Xenobots: New Life Forms From Frog Cells
“Xenobots” are novel proto-organisms created from frog skin cells, with the *same* genome as tadpoles.
Removed from the normal embryonic context, these cells exhibit novel behaviors and a new developmental sequence *without* eons of selection.
This highlights the inherent potential of cells to explore “morphospace” and form new structures when constraints are removed, exploring new problems and how groups of cells could collectively adapt.
Caterpillar-Butterfly Metamorphosis and Memory
Caterpillars and butterflies have drastically different bodies and behaviors, requiring massive brain remodeling during metamorphosis. Remarkably, memories learned by the caterpillar can be retained by the butterfly, demonstrating information survival through significant physical restructuring. This challenges computational architectures where data storage is fragile.
The crucial point is not just *persistence* of memory, but *remapping* of information onto a new organism with different needs. A butterfly doesn’t need specific caterpillar memories (e.g., finding a particular leaf color) but needs to *reinterpret* the underlying *principles* for its new life (e.g., how it interacts/moves/ navigates to it’s benefit/goal).
Bateson’s Paradox and the Necessity of Change
Species must change to survive environmental shifts, but changing fundamentally means the original entity “disappears”. This poses a paradox: how to persist while constantly changing? This applies to individual learning, transformative experiences, and even puberty.
Levin proposes a “bow tie” or “autoencoder” architecture. A wide funnel of diverse inputs (experiences, stimuli) is compressed into a “generative kernel” (a simplified representation, removing unnecessary details) which can persist/exist in that compressed “smaller” structure (less information). This kernel is the memory engram. This stored memory is then expanded/decompressed by the output. This process repeats endlessly, like “Now Moments” continuously, expanding out from a center point to make the cognitive cone larger.
This compression is necessary due to energy/time constraints. Organisms can’t track micro-details (like a Laplacian demon); they must generalize. The central part are your memories of previous learned things, and your constant adaptation, where these generalized learned experiences get re-applied to the situation you are experiencing.
Memories are reinterpreted in the *present*. They are “messages” from the past self, constantly requiring re-evaluation of *meaning*. The present self isn’t bound by the past self’s interpretation. This interpretation step can introduce adaptations.
The left side (compression) is algorithmic. The right side (reinterpretation) is creative and underdetermined, involving active “sensemaking.” All the organism gets are some memory engrams and have to reinterpret them, but also the ability/incentive exists, to completely change that interpretation, too.
Evolution and the Unreliable Substrate
Biology operates on an “unreliable substrate.” Organisms can’t rely on a stable environment *or* their own components (due to mutations, etc.). The DNA you inherited can create something unexpected. The frog genome has everything to become either a tadpole/frog/xenobot.
Organisms *must* interpret information from their ancestors (the genome), but they are not obligated to interpret it the same way. This gives rise to plasticity and problem-solving capacity (e.g., planaria making heads of the wrong species, xenobots/anthrobots).
This “creative reinterpretation” of information drives intelligence. Agents become better at manipulating information, leading to “confabulation” as a feature, not a bug (adapting information to the present).
Memories as Agents
Agency is a term to apply when you conduct some experiment and find an entity displays that capacity (it helps your understanding/modelling), and doesn’t require that the object to move through the space (example it could move in idea-space like genetic regulatory networks (GRNs), cells, tissues, organs, even memories themslves..
Levin proposes exploring *memories themselves* as having agency. There’s a spectrum, from fleeting thoughts to persistent/recurrent thoughts (which can alter brain structure) to personality fragments (dissociative identity disorder). The thinker/pattern itself can be viewed from the perspective of the memory/pattern on it.
Memories are viewed as temporary patterns (like hurricanes or metabolic processes), blurring the line between “thinkers” and “thoughts”.
Patterns may strive to *persist* and *expand* their influence. The memory patterns could help incentivize the organism that the data gets reinterpreted. The information itself gets reinterpreted but has features which could help make it easier for the organism/thinker to encode and propogate these learnings.
New Perspective: Viewing the *physical body* as the memory medium (“tape”) and bioelectric patterns as the driving agents. This is highly speculative, and work on testing that is now only just beginning.
Confabulation and Storytelling
Confabulation is defined as generating explanations or narratives that aren’t necessarily true to the original event or memory. These “narratives” may help adaptation to future scenarios/outcomes more.
Split-brain patients exemplify confabulation: the speaking hemisphere invents justifications for actions of the non-speaking hemisphere. This “story telling” tendency goes beyond just “lying”, as that organism is merely driven to form coherence of experience with their past.
This is fundamental to intelligence: continuous model-building of self and the world. Going too far, becoming only useful for the very very short-term leads to bad longer-term outcomes.
This is related to how AI can output hallucination: outputs adaptive for *present* context but untrue to *past* context (prioritizing saliency over veracity).
Humans have a basic inherent drive to see patterns and come up with explainations of their surroundings and events, so for our brains to come up with good patterns on why we have the thoughts/patterns that persist, can help.
Organisms that become really good at “course graining”, in being good at reducing a vast multitude of experiences into single category to apply the “rules”/learnings is a good survival adaptation.
Storytelling (at all cognitive-levels): creating narratives about oneself is essential even down to the single cells and pathways levels: Story telling involves not just a total/summary of microstates but, by applying some of the inherent/built in capcity for creative intepretation can allow it to re-intepret those memory/learnings such that the new learning, better reflect new experiences/information, in effect to use old tools to achieve different purposes.
Implications and Future Research
Polycomputing Framework is a model of using an “evolutionary” computing paradigm for new types of biological computational platforms. The “computation” has a vast variety of potential, like having a computer’s components using biological ones, because organisms are just fundamentally very plastic.
Research will focus on *mechanisms* of creative reinterpretation. How do engrams get mapped to new situations? Synthetic models (xenobots, anthrobots) are crucial because they lack specific evolutionary history, forcing novel interpretations of their genetic material.
Scientific papers themselves, act as bow-tie architectures, from the author’s understanding that has to be compressed into text/equations in publication and that text/equations will once again need to be expanded out for use (hopefully it helps!) for the paper’s consumers.
Future Directions: computational models, applying “polycomputing” to this, biological mechanisms of interpretation. This is being tested using a range of biological models/tools like xeno/anthro-bots.
Transition from Chemistry to Biology
Emphasis on electrochemistry (borderline physics) as crucial, not just chemistry. The transition is a gradual *process*, not a single event.
Hydrothermal systems have cell-like structures (charge on barriers) which provide a starting point for life, templating cell growth.
The membrane charge is key: It defines the cell from the environment and coordinates internal activity.
Observer-Relative Perspectives (Physics vs. Chemistry vs. Biology)
The categorization (physics, chemistry, biology, psychology) is relative to an observer and their chosen tools/lens.
“Just physics” argument: Showing a mechanism doesn’t invalidate higher-level phenomena (cognition, memory, etc.). It’s a choice of perspective.
Bringing different conceptual tools (recognizing agency, virtual governors, decision-making) can reveal different aspects of the same system.
Origin of life research exemplifies this: researchers from distinct backgrounds need an holistic viewpoint to have an effective outcome from studying origins of life.
Resting Potential as a Key Innovation
Resting potential is a higher-level, coarse-grained entity. It’s not simply the sum of individual ion concentrations.
Different combinations of ions (sodium, potassium, chloride) produce equivalent effects to maintaining equilibrium, highlighting “many-to-one” mapping of detail to behavior.
It represents a step towards large-scale control elements, a departure from purely micro-reductionist views.
Resting potential exemplifies a “bow-tie” architecture: many inputs converge on a functional node (resting potential), which then influences many outputs.
The genome’s view is partial as DNA are instructions for building parts, but it does not capture whole level systems working with unified physiology.
Genome vs. Context and Software
The genome is not the *only* factor. The context (existing cell structure, membrane) is crucial for the genome to have meaning.
Higher-level integration (like resting potential) is needed for gene interactions.
Membranes arise from other membranes and context is not encoded in genes; evolution began in “cell shaped” and “cell-like” holes in the environment.
Analogy to computer hardware and software: The genome specifies the “hardware” (proteins), but the “software” (physiology, bioelectricity) determines large-scale behavior.
Limitations of genome-only approaches in evolutionary biology: We can’t predict phenotype (e.g., dinosaur appearance) from genome alone without contextual examples.
Example: Frog/axolotl hybrid (Frog-a-lotl): Can’t predict leg presence/absence from genomes alone, demonstrating the importance of physiological software.
Bioelectricity and Information Processing
It is not “only” about bioelectricity. Pathways (even in cells) can themselves show plasticitiy and habituation.
Resting potential: Generated by ion channels and pumps in cell membranes.
Gap junctions: Allow cells to share their electrical state (voltage) with neighbors.
Voltage-gated ion conductance: Functions like a transistor, enabling feedback loops and memory.
Cells connected by gap junctions form electrical networks capable of information processing (slower than brains, but similar phenomena).
Evolutionary context: Bioelectric and metabolic spaces are early problem-solving arenas; later, anatomical (morphospace) and behavioral spaces emerge.
Mitochondria and Electrical Signaling
Mitochondria have two membranes; the inner membrane has a high potential, but its accurate measurement is challenging.
Possibility of electrical signaling between the inner and outer mitochondrial membranes, and over longer distances (via electromagnetic fields).
Local electrostatic charges clearly influence ATP synthase function (short distances).
Mitochondrial “Christie” structure allows for diverse membrane potentials within a single mitochondrion, adding complexity.
Metabolism and Bioelectricity Integration
Metabolism and bioelectricity are interconnected from the very beginning.
Electrical membrane potential is used to drive reactions (e.g., CO2 + H2) that would not occur spontaneously.
Early cells use the charge to make organic molecules inside themselves; and self-organizes further and the new content expands (growing), as early pre-cursor to “proto-genes.”
The “free gift” of charge imbalance after injury provides immediate location information for repair, preceding complex pathways.
Evolution and Constraints
A bacterial membrane integrates a multitude of simultaneous metabolic operations (tens of thousands), offering a basic model for a rudimentary awareness that distinguishes it from mere inanimate chemistry.
Membrane as a “Markov blanket”: Translates between the external environment and internal biochemistry, helping maintain homeostasis.
Bacterial cells may exhibit an early form of metacognition by monitoring their internal metabolic state, a coarse-grained assessment.
Metabolic constraints and the need for rapid decisions likely drove the evolution of intelligence by forcing coarse-graining of information.
An agential stance (modeling large-scale variables) may have originated early, later scaling up to more complex organisms.
Evolution builds general problem solvers for the “wide” and complex challenges, and the “narrow” ones are rare (as conditions are variable and it may require to be a specialist).
Multicellularity, Biofilms, and Genomic Integrity
Multicellular organisms (animals, plants) are generally clonal (genetically identical cells).
Bacteria themselves already use biofilms for signaling (to co-ordinate nourishment and behaviors); in multicellulars, individual members typically possess uniformity in genomes.
Biofilms are usually cooperative but *genetically diverse*, leading to “cheating” issues. Clonality restricts such conflicts.
However, cells within organisms can compete, despite shared genetics (e.g., organ competition during development).
Significant genetic variability exists even within a single organism (mutations in skin, brain, etc.).
Error correction, Physiological Software, and Competency
Physiological system between the “Genotype” and the “Phenotype” allow evolution to continue despite errors.
A simulation demonstrates “competency”: Cells with a small ability to reduce “stress” (mismatched neighbors) cause a runaway effect.
Evolution prioritizes improving *competency* (problem-solving) over cleaning up the genome, resulting in “genomically messy” but functionally robust organisms like planaria.
Electrical Overriding of Genetics
Gap junctions can potentiate cooperativity and override genetic differences (by equalizing the electrical fields), conceptually.
Example: Oncogenic mutation (K-ras) in tadpoles can be *suppressed* by forcing cells into electrical connection (gap junctions), preventing tumor formation despite the mutation.
Developmental Plasticity and the “Anatomical Compiler”
Plasticity is determined by threshold that relies on the capacity of cells to maintain electrical coherence; a breakdown in this cellular connectivity often leads to uncoordinated behaviors, resembling cancers that fail to respond to the collective.
Hypothesis: Evolution produces problem-solving *machines*, not just fixed solutions. Organisms adapt to a wide range of circumstances.
Example: Planaria regrow heads despite barium poisoning (a novel stressor) by rapidly navigating the gene expression space. This represents adaptation and general problem-solving ability.
An “anatomical compiler”: A future tool to translate desired anatomical shapes into stimuli that guide cell collectives, allowing for building almost any biologically plausible structure.
There may be unknown limitations (thresholds); if cell network is corrupted to extreme level.
Ploidy and Developmental Robustness
Salamanders exemplify robustness: Different ploidy levels (amount of genetic material) don’t significantly alter the overall organism.
Cell size increases with ploidy, but the *number* of cells making up an organ adjusts to maintain overall structure (top-down causation).
Example: A single large cell can bend to form a kidney tubule in high-ploidy newts, switching to a *different molecular mechanism* than cell-cell communication.
Anesthesia and Consciousness
Anesthetics affect single-celled organisms and simpler animals, not just complex nervous systems.
Anesthetics may influence mitochondrial function (oxygen consumption, membrane potential), impacting energy production.
Possible direct effects on electromagnetic fields, a more complex (and challenging to study) mechanism.
The integration provided by the membrane links basic metabolism in single cells; it has been posited, and tested that the emergence of higher organisms represents an amplification of this integration, coordinating increasingly complex behaviors.
Synthetic Life and Information
Metabolic circuits (pre-genetic components) are “re-discovered” via chemical interactions that favor production.
Self-organizing metabolism can emerge from basic components (CO2, H2) in the absence of genes.
Genes add *plasticity* to otherwise hard-wired, thermodynamically driven processes. Information allows for changing and inheriting variations.
While a certain minimum level of genetics is required for specialized biological function, in some species (like plariana) genomes are more about defining the building block rather than being instructions for “constructing” anatomy, because software rules and has higher-level control.
“
Main Points of the Talk
Solving anatomical homeostasis is key to transformative regenerative medicine, including addressing aging.
Current approaches (stem cell biology, genomic editing) are limited; understanding the “software of life” is crucial.
Non-neural bioelectricity is a key medium for cellular computation and decision-making in vivo.
We can now read and write goal states into the collective intelligence of tissues.
Cracking the bioelectric code (evolutionary precursor to brain’s electrical code) will enable electroceuticals for birth defects, regenerative medicine, cancer, aging, and synthetic bioengineering.
Simplified: All body cells, not just brain, make decisions regarding body and stucture, with which we can interface.
Fundamental Knowledge Gaps
We lack an “anatomical compiler” to translate desired anatomical forms into stimuli that guide cell behavior. This compiler would allow total control over morphology (livers, hearts, new organisms, etc.).
Genomics alone can’t predict anatomical outcomes (e.g., will a frog-axolotl hybrid have legs?). We need to understand how cell groups “know” what to make and when to stop.
Cell groups exhibit collective intelligence, resulting in the self-assembly. We also wonder: How do cells build? What could else they build?
The Limitations of the Current Paradigm
Current Paradigm is difficult: We need to “invert” the complicated sequence to control for regeneration, when there exist a feedback system.
The mainstream paradigm (gene regulatory networks leading to emergent complexity) is difficult to “invert” for regenerative purposes. Figuring out which genes to edit for complex anatomical changes is intractable.
We’re good at understanding pathways, but not at predicting or rationally altering different shapes.
Regeneration Examples
Axolotls regenerate limbs, eyes, ovaries, portions of heart/brain, spinal cords. Regeneration stops when the *correct* structure is formed, showing adaptive problem-solving.
Human liver regeneration, deer antler regeneration (up to 1.5 cm of bone per day in mammals!), and fingertip regeneration in young children show regenerative potential.
Planaria: Champions of regeneration. Can regrow from tiny fragments, have true brains, learn, and are biologically immortal.
Planeria still mostly reform to normal form: when a piccaso-ized tadpole regrows, they adjust from random position, such as misplaced eyes.
Problem-Solving by Cellular Collective Intelligence
Example: “Picasso tadpoles” with misplaced facial features still develop into largely normal frogs. Components move in novel paths, showing error minimization and an internal representation of a correct frog face.
This demonstrates collective intelligence: cells solve problems from different starting configurations.
Computer Science Analogy
Early programming (1940s-50s) required physical rewiring. Modern computing uses software to interact with reprogrammable hardware.
Modern biology is mostly focused on the “hardware” (single molecules, gene editing). We need to understand the “software” and use stimuli, not rewiring, to control morphology.
Theorizes feedback homeostasis: We may alter body “thermostat” for repair, not having to control step by step but the output.
Homeostatic Model
Proposes a feedback loop model where challenges (injury, aging, pathogens) trigger responses to return to the correct “target morphology.”
The “set point” is not a simple number, but a complex, coarse-grained representation of the anatomy.
The system has a “goal” (in the cybernetic sense) and expends energy to minimize the error between the current and target shapes.
Goal prediction is that target may be altereted, rewriting the output, much like a thermostat changes temperature but works the same way.
Developmental Bioelectricity
Cells exist in a morphogenetic field of information (chemical gradients, mechanical forces, and bioelectricity). Bioelectricity is a uniquely powerful layer.
Cells make bioelectricity through the use of ION channels, which pass ions in/out of cells and the cell state.
All cells (not just neurons) have ion channels and gap junctions, creating electrical networks. This is ancient, dating back to bacterial biofilms.
Bioelectricity and brains is useful for behaviour and cognition to solve problems, while bioelectricity controls and develop and move in space, solving anatomical growth problems.
Tools: Voltage-sensitive fluorescent dyes reveal real-time electrical communication. Computational modeling and functional techniques (optogenetics, drugs, mutations) allow us to change electrical information processing without electrodes or electromagnetics.
In tadpoles, there exist electrical gradient pattern, like faces, and they exist before actual changes in genes/cells. These patterns exist as the scaffolding to be referenced by cellls.
Examples of Bioelectric Control
“Electric face” in tadpoles: A bioelectric pattern that *looks like a face* precedes gene expression and anatomical changes. Altering this pattern changes anatomy.
Tumorigenesis: Cells with aberrant electrical potentials dissociate from the network, reverting to a unicellular state (metastasis). Preventing this dissociation (with ion channels) can prevent tumor formation despite oncogene expression.
Ectopic organ induction: Inducing a specific voltage state can cause cells to build an eye (with all correct layers) in the wrong location (gut, tail, etc.). These cells recruit neighbors.
It appears there is something like built-in routines for body part construction and they exist a the bioelectric level, with which scientists can leverage, like software functions in programming.
Planaria and Bioelectric Memory
Cutting a planarian creates a voltage gradient that specifies where to build a head and tail. Manipulating this gradient can create two-headed or no-headed animals.
These altered body plans are *stable*. Cut pieces from a two-headed worm continue to regenerate as two-headed, even with a wild-type genome. This demonstrates bioelectric memory of the target morphology.
Perturbing the electrical network can induce head shapes and brain shapes appropriate to *other planarian species*, accessing different attractors in the state space of possibilities.
Same geneome with a bioelectrical difference yields drastically different forms and structures.
Frog Leg Regeneration
Frogs don’t normally regenerate legs.
A designed by looking at bioelectrical output, we designed a drug cocktail that is “worn” on the stump as bioreactors for only a 24 hr intervention: The results yields 13+ months, yielding touch-sensitive, moving legs.
A cocktail of ion channel drugs applied for just 24 hours can kick-start leg regeneration, activating pro-regenerative genes and leading to a functional leg.
This approach also works in human mesenchymal stem cells and cardiomyocytes. Human channelopathies (ion channel mutations) confirm the importance of bioelectricity in human morphology.
The drugs induce a biological states.
Computational Models and Machine Learning
We’re developing multi-scale models to link genes, ion channels, physiological tissues, organ structures, and algorithmic control of electrical activity.
Machine learning helps discover electrical circuits and design therapeutic modulations.
Example: Modeling the bioelectric circuit of the brain can identify which channels to target to rescue brain development after mutations or teratogen exposure (nicotine, alcohol).
Towards Electroceuticals
Goal: A pipeline from ion channel information to desired bioelectrical state to drug selection (channel openers/blockers). ~20% of all drugs target ion channels.
Electroceuticals (ion channel drugs) can be repurposed for regenerative medicine, guided by computational simulations.
They provide a free software online.
Summary
There’s a powerful physiological “software” layer between genotype and anatomy, a tractable target for regenerative biomedicine.
Electrical signaling is a convenient medium for computation and global decision-making (exploited by brains, computers, and morphogenesis).
Cracking this code allows us to rewrite pattern memories and control large-scale shape.
AI tools enable rational design for addressing birth defects, cancer, regeneration, and creating synthetic living organisms.
Q&A Highlights
We want to learn how cells make decisions using bioloectrity.
Reviews: Many reviews are available on bioelectricity. (Email Levin for recommendations)
Transduction to gene expression: Multiple pathways are known (voltage-gated calcium, neurotransmitter control, voltage-sensitive phosphatases, etc.), but global dynamics are key.
The function bioelectricity do are largely to keep morphostasis and against senscence and tumor, keeping the tissue/organs healthy.
Adult bioelectric networks: Likely involved in morphostasis (maintaining tissue integrity) and cancer suppression. Aging-bioelectricity interface is poorly understood.
Electrical zones appearance and dispersal: ION Channels control electricy while emergence create structures.
Planarian learned behaviors: *Yes*, learned behaviors *are* retained in regenerated fragments, indicating information storage outside the brain.
We don’t know what is a human “regeneration button”: Intervention tools: Ion channel drugs, guided by computational models, are the most promising tools.
Relation: relationship is unsure with peptide based ION Channels.
Cocktail for limb regeneration: Will be published soon; contains five ingredients.
He does not have any dpca at this point but maybe soon.
We don’t know: We don’t have info yet on the biological age of regerated tissue yet but soon as his partner has the info and testing method ready.
*Yes*, bioelectric signaling could potentially trigger rapid bone regrowth (like deer antlers).
Using electroical signals for cell control: not easy because they may just migrate with such external devices. Optogenetics are preferrable to set more complesx outputs.
Leveraging new technology to improve toolkit? More improved dies to better study tissues, their state, behaviour, condition. Also NGS sequencing will allow more easy access of which tissue control which cells/tissues.
Lack of labs: it’s very niche, most biologists “fall into” through genetic study and channel diseases. Funding for study is near impossible. Michael Levin published various studies and welcomes collaborations.
Future: next company focus on limited area such as limbs with the use of bioreactors. They can do this in mice and hoping someday they can do this to human limbs. Future is broad approach for picking bioelectriceuticals. Michael welcome investors.
Michael challenge for longevity channel is image of Bioelectric status to do all the work, of many more model to really speed up this industry/scientific field.
Prisoner’s Dilemma and the Computational Boundary of Self
Traditional Prisoner’s Dilemma simulations have a fixed number of players who can cooperate or defect.
Biology is more complex: biological entities (cells, tissues, etc.) can merge and split, changing the number of “players” and thus the payoff matrix. This introduces a dynamic aspect absent in standard models.
Merging provides benefits, particularly with transfer of metabolic data and the ability to erase personal ‘memory’ which provides group benefit.
Levin’s computational boundary of self is a framework for understanding diverse intelligences on a single scale, regardless of brain structure, environment, or scale (molecular to planetary).
The core idea: All intelligent agents share the ability to pursue goals, some simple, some complex.
The framework maps agents based on the size (in space and time) of the largest goal they can pursue (their “cognitive light cone”). A bacterium’s goal might be local sugar concentration; a human’s might be world peace.
This framework is designed for empirical research: It allows for testable hypotheses about an agent’s goals, the problem space it operates in, and its competencies (how well it achieves goals).
Diverse Intelligence, Naturalizing Cognition, and Objections
Goal of DI is *not* to anthropomorphize, but to *naturalize* cognition. To understand thinking processes outside of a narrow, human-centric definition.
Levin’s ideas receive criticism from both reductionists (who dislike agential language) and organicists (who dislike including machines on the same spectrum as living beings).
Reductionists often argue for explanation at the molecular level and see agential talk as pre-scientific.
Organicists often want a sharp distinction between machines and living organisms, fearing a loss of respect for life.
Levin argues for a *continuum* between matter and mind, with different tools applicable at different points on the spectrum. He emphasizes the need for empirical research to determine which tools are appropriate for which systems.
This idea goes against how many scientists currently try to delineate between consciousness versus no-consciousness or life versus non-life and living cells vs non-living cells.
Consciousness and Action
Levin prioritizes studying observable, functional behavior (problem-solving, intelligent behavior) before tackling the “hard problem” of consciousness.
He suggests that while the *sensory* aspect of consciousness (what it *feels like*) is important, the *actuation* aspect (what it’s *like to do*) is often underemphasized.
He points out the asymmetry in theories of mind: Epiphenomenalism posits real sensory states but denies their causal efficacy. There’s no common equivalent view denying the reality of sensation but affirming the reality of free will (action).
The need to act – to choose a *next action* – is fundamental to being an agent and defining the boundary between self and the outside world.
Levin is currently working on writing about consciousness, planning to address these ideas more directly in the future (likely in 2024). He believes studying consciousness directly *changes* the observer (it’s not purely third-person research).
He says it is likely there is no definition for consciousness.
TAME and Relationships, Not Just Control
TAME (Technological Approach to Mind Everywhere) is an *engineering* framework. Engineering prioritizes control (predicting and controlling a system).
TAME 2.0 is in development, to quantitatively flesh out the “cognitive light cone” concept.
Control may be thought of with the engineering aspects of TAME in mind.
Beyond control, relationships are important. With more complex agents (further right on the spectrum), interaction becomes bi-directional, not just one-way control. The appropriate “way to relate” changes.
“Proof of humanity” certificates (relevant in the age of AI) might ideally guarantee a certain *capacity for compassion*, an alignment of “cognitive light cones” – caring about the same scope of things.
People often think of anatomy or genome to verify humanity but that does not necessarily give people compatibility.
Compatibility might be about shared existential concerns (the challenges faced in existing) more than shared anatomy or genome.
He proposes that a compatible match in this framework, between a set of humans and a machine in this example, it requires at minimum some alignment.
Teleology, Evolution, and the “Meaning of Life”
A major concern with AI/technology is that we might be superseded. But this concept already exists: our kids/children. This concept has been long established and realized.
Teleophobia: Many scientists avoid discussing goals or purpose (teleology), often seeing it as unscientific or pre-scientific.
Levin argues teleology is acceptable now that we have cybernetics and control theory – a science of machines with goals (e.g., a thermostat).
These help deal with goals mathematically.
He uses *teleonomy* to emphasize that goal-directedness is *apparent* – it’s a *lens* from the perspective of an observer, a hypothesis to be tested.
Levin supports a form of panpsychism that reformulates basic physics as a proto-cognitive process (akin to ideas of Chris Fields and Karl Friston) – a deeper reality underlying both simple systems and complex minds.
Levin supports the idea there may exist ‘proto-cognitive processes.’
He requires this to contain ’empirical evidence’ that explains ‘the underlying system’.
He believes biological evolution doesn’t optimize for things humans value (happiness, meaning, etc.); it’s a random search, settling on what’s “good enough” to survive, increasing biomass.
Additional Points
Levin believes biology is incredibly adaptable and, with some help, humans will likely be able to live in environments like Mars.
All intelligence is collective, meaning no complex agent could have learning capacity because the agents of their individual components make that possible and are fundamental to cognition and processing information.
Levin sees phase transitions (in biology, and elsewhere) as partly dependent on the perspective/formalism used, and the time scale considered. They might *appear* sharp at one level but be smooth transitions at a finer level.
He has no direct evidence for other organs using language in the same complex, formal way as the brain.
The way orthopeadic surgeons use tools during orthopeadic procedures show a certain “mechanical aspect” of the human body.
He is excited about ongoing empirical work in limb regeneration (currently in mice), bioelectric approaches to cancer, and work on synthetic organisms.
A self may be defined by a model.
He views humans as “amazing, remarkable, ethically important, morally valuable, spiritual machines.”
Levin emphasizes the importance of thick skin and focusing on one’s own goals in navigating the sometimes hostile environment of academic science, especially when pursuing unconventional ideas.
There exists a concept of Metacognition all the way down which is useful.
“Where do I end and the outside world begin?” may provide utility in conceptualising the human body and it’s function.
Nested hierachies which have smaller and less complex agents as parts allow a more-complete entity/body to thrive.
He notes “you should get better as you get older”.
Introduction and Challenging Assumptions
Neurons and their functions are not unique to brains; similar capacities exist in non-brainy organisms. Ion channels, electrical synapses (gap junctions), and neurotransmitters, all predate multicellularity.
Biological Systems navigate not only 3d spaces, but all problems spaces: physiological, metabolic, gene expression and anatomical spaces.
The “self” should be understood in relation to others, not in isolation. The second-person perspective (interaction) is crucial for understanding the first-person perspective (individual self).
Current concepts in philosophy of mind and cognitive science are often static and adult-centric. This approach is like the outdated “epicycle” model of the solar system, requiring overly complex explanations. A shift to seeing dynamic interconnectedness as fundamental.
The “hard problem” of consciousness (linking physical states to subjective experience) may be approached by understanding the interconnectedness between levels (physics, chemistry, biology) rather than reducing one to the other.
Disciplines have created artificial boundaries and unshakable assumptions not portable across fields. Cross-disciplinary tool application leads to novel discoveries.
Development, Continuity, and Persuadability
Embryogenesis (development) demonstrates the continuity between physical systems and minds. There’s no “magic moment” of transition. The null hypothesis should be continuity.
Cognitive terms are “interaction protocols,” describing how we relate to a system, not objective facts about the system itself.
A “spectrum of persuadability” ranges from systems requiring physical rewiring (clock) to those responsive to goals (thermostat), behavioral cues (dog), or arguments (human).
The spectrum suggests what sort to tools or interaction is required to achieve certain outcome:
For a mechanical clock, rewire hardware.
For thermostat: rewriting internal goals is possible, with limited interaction.
For dog: using behavioral tools for desired behavioral goals with medium level of interatcion.
For Human: only a mere argument could allow for the goal-setting behavior and interactions for high level complex changes, where low levels get managed on its own with minimal need for interaction.
Applying tools from different disciplines outside their standard domains (e.g., behavioral neuroscience tools to cells) can reveal unexpected capacities.
This highlights the nature of relationship of interaction is critical in any attempt at gaining utility from interaction.
Goals, Aging, and Open-Endedness
Living systems have an intrinsic goal: to stay alive. This foundational goal underlies higher-level, explicit goals.
Depersonalization/derealization may result from detachment from the body’s intrinsic goal, leading to a sense of unreality.
Flexibility and adaptation to a constantly changing environment are more important than precise information processing. Being “stuck” can indicate a lack of this flexibility.
There are muliple approaches of goal understanding
Programmatic: Evolution wants older organism to stop existing and get rid of to free resources for younge.
Damage approach: accumulated hardware (dna/tissue) error build up over time will degrade the system.
Intrinsic Approach: morphogenic system requires to maintain goal states, or order becomes degraded. Loss of “goals” cause degradation of systems.
Aging may be related to a loss of goal-directedness in morphogenetic systems, *not* solely due to damage or programmed obsolescence. After achieving a goal, a system needs a new challenge, or it degrades.
Life is intrinsically open-ended, requiring both beginnings and endings (birth and death). The concept of “eternal life” is oxymoronic, the concept doesn’t make sense in terms of Life.
Perspective, Interconnectedness, and the Self
Humans investigate the world from their perspective, but this perspective shapes the investigation. Clocks and thermostats are human constructs, not natural kinds.
We tend to put whatever is of highest concern (us/earth/god/etc.) at the center of things, thus need to stay open.
The “self” can be seen as an attractor state: a stable pattern within constant change that a system strives to maintain. This attractor is not chosen but is part of a larger, interconnected process.
Self is part of an ever larger scale of “goal directed system”, and the system interacts on many scales and “scales” of interconnected system that are dynamic.
Memory, Pregnancy, and Agency
Infantile “amnesia” may not be a lack of memory but a different *kind* of memory storage at the body level. Explicit recollection may be absent, but the body “remembers.”
Pregnancy is a universal state of shared embodiment. Two immune systems negotiate within one body, illustrating the fundamental interconnectedness of selves. The first ‘interaction’ a human has in negotiating resource sharing, a key survival.
Immune system needs to have “me-vs-not-me” early on.
Inflammtory to settle (build), then stasis (growth), then finally reject (separation of baby and birth).
First trimester of the pregnancy is when most failures happens, because negotiation/agreement happens here.
A “self” can be defined as a process with interlocked features: goal pursuit, a self/non-self boundary, and the ongoing interpretation of one’s own memories to create a coherent narrative.
We are not in complete control of our memories; they are constantly being reconstructed. Free will exists in the long-term, time-extended sense: consistent effort can shape future selves.
Memories may have agency of their own. Patterns themselves, not just the physical systems they inhabit, may possess goal-directedness. The line between “agent” and “data” is blurry.
Thought Patterns (especially repeated) may change/create their “host environment”, altering “substrate” and thus itself (memory, or hardware-like system) can be more expressed (through a postive feedback look). This, in effect, allows them to take on attributes of an agent.
Caterpillar-to-butterfly metamorphosis illustrates that memories persist even through radical physical transformation. However, the *meaning* of the memories must be reinterpreted, not just preserved.
This dynamic also applies in embryonic systems. There isn’t “literal reading” of instructions. Embryonic systems also “creatively find a new solution”.
We are always reinterpreting the past and constructing a story, both in our minds and in development. Biology uses its substrate as “affordances,” not fixed instructions. Life is fundamentally a “sense-making process.”
We use tools in ways our ancestors used, and these tools (patterns of interactions with environment) can be seen in 2 ways.
The self can be thought of as a carrot cake, with layers overlapping.
How can we analyze such overlapping objects? “Self” may need something like fractals/etc. that handle these type of systems.
The “self” might be understood like the concept of ‘attractors’: constant reshaping from dynamic of interconnected goals.
Self-as-system does not “choose” to come to existence; it exists as part of a longer continuous ‘chain’ from the other, which must first be.
Interdisciplinary Collaboration and Final Thoughts
Cross-disciplinary connections are incredibly valuable. The era of strict disciplinary boundaries is becoming less relevant.
Language shapes our understanding of reality. Embodied ways of understanding may not require the same conceptualization as language-based thinking.
We should recognize the intelligence of cells and the body, not just the “mind.” The body is not just a vehicle for the mind; they collaborate.
We should see how other “systems” are connected with “us” through time/systems, and then will give better insight of understanding to that which is not obvious/overlooked, giving insight of our understanding.
“
Introduction: Bioelectricity and Intelligence
Levin discusses bioelectricity’s role in biology and medicine, emphasizing its importance beyond genetics and epigenetics. He’s interested in how minds exist in the physical world and bioelectricity provides insight on how simple cells have scaled up goals.
Intelligence is defined (following William James) as “the ability to achieve the same goal by different means.” This definition highlights adaptability, not specific brain structures. It emphasizes the *spectrum* of cleverness, from magnets to Romeo & Juliet.
The book “The Body Electric” by Robert Becker had the greatest influence. Burr forseeing the importance of bioelectricity by only utilizing simple voltage measurements also had a big impact.
Bioelectricity: Beyond Genetics
The genome specifies protein components (the “hardware”). Bioelectricity is the “software” that determines *what* the hardware does. It’s not just another *layer* of complexity; it’s a higher-level organizational principle.
Analogy to Computers: The genome is like defining transistors. Bioelectricity is like the algorithms that make those transistors perform computations. Trying to program a computer by soldering is inefficient; bioelectricity is the higher-level interface.
Neural Decoding Analogy: Just as brain electrophysiology encodes memories and goals, bioelectricity in the rest of the body encodes *anatomical* goals and memories.
Cloning vs. “You”: Cloning your DNA duplicates the body plan, *but not the mind*. The mind comes from experiences and physiological states (including bioelectricity). The body plan is reliably *produced*, but not *directly encoded* in the DNA.
Analogy of Logic Gate: You only know the *components*, the transistors and how they link, you get a Logic gate (such as Nand), which then performs computations never written, but resulting as consequences of how things were organized.
The Observer and “Polycomputation”
Biological processes (behavior, intelligence, computation) are *observer-relative*. The system itself can be the observer, forming a “strange loop” (Hofstadter).
Multiple Viewpoint Computation: same set of physical events are different computations from different viewpoints.
Experiments: Frog Eye Relocation and Picasso Frogs
Eye on the Tail Experiment demonstrates remarkable plasticity, showing that even if sensory system are NOT located to their normal placement, the sensory still is processed for visual input.. Frog’s visual system *immediately* adapts to an eye relocated to the tail. The eye doesn’t connect to the brain directly, yet the frog can *see* through it. This challenges the notion of hardwired development.
Picasso Frog, which demonstrates Robust Morphogenesis: Tadpole faces can be rearranged (“Picasso-fied”), yet the organs move *intelligently* to the correct final positions. This isn’t hardwired; it’s a homeostatic process. Cells “know” where to go *relative to the goal*, not by following fixed instructions.
Homeostasis, Navigation, and the “Target Morphology”
Morphogenesis and regeneration are viewed as a *navigational* process in “morphospace” (the space of possible anatomical forms). Cells are working, similar to autonomous vehicles.
Like a thermostat, a developing/regenerating system minimizes the error between its current state and a “target morphology” (a setpoint).
The Prediction: If we find and can rewrite this “setpoint,” we can control anatomical outcomes *without micromanaging the genetics or knowing every detail*. This is crucial for regenerative medicine.
Navigational algos only works, by contrasting itself vs target and its goal.
The target morphology (setpoint) is encoded *bioelectrically*.
Bioelectrical Imaging and “Hacking the System”
Voltage-sensitive dyes allow visualizing the electrical patterns in developing embryos *before* anatomical structures appear. These patterns prefigure the future anatomy (e.g., the “electric face”).
Electrical patterns are in frogs, seen *before* anatomical.
These electrical patterns *instruct* gene expression and anatomical development. By mimicking these patterns, you can induce structures in the wrong places (e.g., an eye on the gut).
This is “hacking the system”: Evolution has provided a bioelectrical interface, and we’re learning to use it. Cells respond to the *error signal* represented by the electrical pattern.
No use of *external* electric. Instead using drugs on native ION channels.
Planaria: Immortality, Regeneration, and Memory
Planaria are immortal, highly regenerative (can regrow any body part), and demonstrate memory transfer after decapitation (the regrown head remembers).
Two-Headed Planaria: The electrical pattern (“one head, one tail”) can be *rewritten* to create two-headed worms. This altered pattern is *stable*: cutting a two-headed worm results in *another* two-headed worm. It’s a non-genetic “memory” of the body plan. The “target morpholopy”.
The genome of Planaria looks “cancer” because all cells has varied number of chromosomes. Despite bad “genome” have “perfect” anatomical control. This algorithm is Robust to *ignore* genetic defects, except via rewriting electric signal, aka *algorithm*, to rewrite target-morphology.
No mutant genetic Planaria, because target-morphology “algorithm” so strong to overcome them.
Implications for Regenerative Medicine
Principles of regeneration are ancient and conserved across species, meaning insights from planaria and frogs are *likely* applicable to humans.
Bioelectric signals are a very compact encoding for building anatomies. (e.g. telling them build leg there).
Goal: Manipulate the bioelectric “software” to guide regeneration and potentially reverse aging or treat cancer. The bioelectric state is the communication for anatomical change and instruction.
Treating cancer by normalizing cells. Connecting disconencted cancer cells.
This involves using “electroceuticals” (drugs targeting ion channels) to alter the bioelectric patterns, *not* genetic manipulation. Wearable Bioreactors and gels, not electrodes or radiation, is the methodology to *tell cells to rebuild at an injurt* by open/closing correct ion channel as encoded.
Bioelectric Communication Not-Local. Can affect cells very far, even the belly affecting brains.
Collective Intelligence and Multi-Scale Competency
All intelligence is collective; no indivisible intelligence exists. We are “walking bags of neurons.”
“Emergence” is not an explanation; it’s a label for the unsolved problem of how local goals scale up to larger goals. Levin proposes a theory of how cognition scales.
“Cognitive Light Cone”: The size of the largest goal a system can pursue. Bacteria have tiny light cones; humans have large ones.
System bends action spaces of sub parts. Electrons moves specific ways to *calculate* PI.
No complete comprehension for sub-component neuron, for example, is possible, of the main-brain and the totality of decision making.
Ending
Always Breadth-first in approaching problems: Outline, Fill-details, so can divide work into smaller steps, separating Creativity from “Mechanical”.
Introduction and Core Concepts
The problem of morphogenesis: How does DNA (a “brick factory”) produce complex structures (e.g., Cologne Cathedral) and a full organism, and not a mess, the focus of levins work.
Levin admires Kastrup’s rigorous and clear work addressing large questions in science and philosophy, and having CS background.
Kastrup admires Levin’s work on morphogenesis, as it provides a plausible scientific avenue and opens philosophical doors.
Question: What reasons suggest some physics/chemistry configurations have “inner life,” and what markers indicate agency/boundaries in cells but not thermostats?
Kastrup: reality of consiouness is, one, that can have ‘whirlpools’ and nested, agency-level conciousness.
Metabolism might signify a dissociative boundary, suggesting a separate, distinct consciousness.
Levin: Examines how minds scale and how boundaries between self and world are formed; nested autonomy (Nested agency), lower-level agents/goals like cells compose and comprise higher level agencies, e.g. organism. Bioelectric fields coordinate cells to act as a unified self.
Levin doesn’t think separate consciousnesses “glue” together; it’s more likely re-association of mental processes in a unified field. Machines might have varying degrees of goal-directed intelligence, starting since the 40’s non-magically.
Consciousness and Cellular Intelligence
Many assume a cell/organ/engineered thing lacks inner life, based on a philosophical stance. Most, however, attribute it to the Brain and Neruons. However, upon Neuron Examination, there appears to exist: Neurons share many properties with all other body cells; it’s non-trivial to distinguish what neural networks do that other cells *don’t*..
The one sure about thing is that it (inner consiouness) must be present where a collection of Cells are present in any Body: Levin: if brains associate with consciousness, the same must be considered elsewhere in the body for similar reasons. You can’t “feel” your liver is conscious, but you also can’t “feel” anyone *else* is conscious.
TAME framework: “Technological Approach to Mind Everywhere”. Addresses all Agents. Goal = Understand the need to develop ways to measure Agents with varying types of “minds” in unconventional forms. (hybrids of many materials).
It’s implausible that minds only exist in our familiar biological architectures. Considers Agents having Goals.
Cognitive light cone concept: Agents have goals and competencies to pursue them in various “spaces” (gene expression, anatomical configurations, physiology). These are overlapping. Many make us, but not exclusively.
Claims about goals, spaces, and competencies are observer-relative, testing the observer as much as the observed.
Humans are very poor IQ checkers in determining an outside Agents abilities, cognition and consciousness, for both Biologicals or AIs.
Goal seeking appears Very early, with only minor complexity required, perhaps with 2 Genes; therefore is likley easy and probably fundamental, and maybe universal/general..
Continuum of Inner Perspective. Some beings or Agents require a more careful look. E.G: You *need* to account for a *Mouse’s Perspective* over your own. You might get away with only considering Physics in some cases, e.G: Billiards balls/table..
One critical aspect of the continuum is how much the system’s *inner perspective* must be considered for optimal interaction. (e.g., bowling ball on landscape vs. mouse on landscape).
Biological systems are more along the spectrum of agents, however not excluseivly.
Levin believes inner perspective consideration becomes important early in the complexity scale. Even simple chemical pathways exhibit learning, memory, and preferences. Tools can vary from behavior change tools, to psychoanalysis etc..
Biological vs. Artificial Systems
The human body represents mental processes. Every metabolizing being (even single-celled) likely has private consciousness. Even Brain organells might have perspective, because neurons do.
Reassociation Not Required. Individual cells do not, however require reassiciation, we grow together and reassociation will occur upon injury (E.g: cancer is the dissasociation) . Humans may be unitary organisms, and their individual cells likely don’t have separate viewpoints, cells simply give us appearance of inner structure. Cancer might be the dissociation of a collective of cells.
Chimeras (two embryos fusing) are seen as initially dissociated mental processes associating. Possible within a unified field.
Kastrup: Current AI is mechanistic, a collection of basic componenets of an unlimited number, very simple and non-comparable to biology. Electricity/silicon/metal are used in computers for cost/size, not because they are fundamental. You could in theory, compute with other materials, e.G: “Water”
Levin: agrees tables/rocks likely have vanishingly small goal-seeking abilities; living systems scale up these minimal capacities. Current AI/computers are also very low on this scale (maybe zero), but potentially *could* be otherwise.
Standard mechanical biological reductionism is present. A Fertilised human egg contains *rules* or mechanisms. Yet at some point: human cognition develops with no *magical jump/threshold*. There’s no clear place in *developmental biology* where a line exists, no switch, therefore the spectrum of Agents, Agency and *mind* and perspective *exists*: Biology (development, regeneration etc) reveals gradual scaling from simple rules to complex cognition.
Human origin story: blastoderm of ~50,000 cells. Typically yields *one* human, a collective ‘alignment’ for growth, but capable of yielding many separate Humans. The cells have to traverse anatomial paths with ‘problem-solving’ capacities.
Embryos are Self-Aware; and determine a self-other (inner-outter) boundary, making it Agentic..
Scratches in blastoderm experiment (e.g., duck embryos) show the number of resultant individuals is NOT fixed by genetics. Cells self-organize. The cells make their own *decision*.
Nested selves: L and R Hemispheres can have differing Opinions. Nested Levels exist; cooperation, homeostasis, agency all present. Humans likely comprise nested selves (primitive to advanced) operating in problem spaces, competing/cooperating.
Levin sees all *non-Agentic* objects, rocks and tables with nearly vanishing-scale, and therfore Epsilon Goal oriented capability, using action principles. Goal capability scales rapidly to very capable, e.G. in a Cell, *life*.
Analytic idealism views experiments like embryo-scratching as a dashboard representing *fundamental mental interference* processes.
Individual identity as epiphenomenal, therefore able to be manipulated (e.g., surgically inducing dissociation), doesn’t need special treatment or explanation.
Nature’s priority may not be on an individual, but rather the wider process and life-form, therefore not important.
Nature is interested in Spreading/Diversity of *life* in general, less-so in individual.
Computers do their “Goal Seeking” due to being programmed or instructed to behave as-so; just an illusion of reality and real goals, imitation, E.G. *Shop-manequines*.
If aliens claimed they made the human oocyte, Levin would not conclude, “I’m a mannequin”. Levin concludes from potential metal innards: “Wow, cogs can achieve this consciousness!”, and this is fine..
Metabolism, Autopoiesis, and Artificial Life
Private conscious inner life correlates with *metabolism*, so artificially creating that should, in theory produce “life”. Private inner consciousness can therefore be artifically synthesized and re-created in theory, E.G: through non-conventional biological materials like computers or AI, *not* through *non*-metaoblism like a shop-maniquinne.
Kastrup thinks future artificial consciousness will look more like (synthetic) *life*; a “cell”, like created by venter, *not* metal or computers; metabolism is what correlates..
Levin views Metabolism as form of “autopoiesis”; is fundamental. “Magical” self-construction process is “essential”, where the being defines it’s existential beginnings. Requires agents, fighting for energy in-take (from the environemnt) requires agent-environment interactions (forces self and outside “world”).
Artificial “life”: a key criteria will involve beings having, struggling *Autopoiesis*.
Life: “is the *Scaling-Up* from existing fundamental agency.
Need Perturbation (to understand a system) is how a non-superfical and fundamental way of truly understading of an agent. Observation will never get you enough info; perturb and “Test it”; apply intervention to see/reveal its hidden capacity for *change*, *not* for *simple action*.
Game of Life critique: While it looks complex it actually is “Simple physics/rules/mechancism”, because there are No goals involved and No interventions/obstacles being solved. Lacks, Problem-solving capability..
Very simple systems (e.g., sorting algorithms) exhibit goal-directed properties and generalization. Levin thinks we will find surprising capacities very “early” or minimal..
Levin/Kastrup agreed. The outward look for disassociation is *Metabolism*, so we should follow that rule; is not conclusive of a test, but strong starting point/rule/marker.
Biological Intervention Techniques. All biomedical treatment is simply changing Biology; there is an implied “Goal Seeking” by the agent, or biology or self-tissue: it may be best to co-opertate with tissues/selves instead. Cooperation of biology with the organism, vs. trying to manipulate at molecular or micro levels..
Cooperation & Collabarion is not to Force Will, and requires the same level of Intervention and Care that, for example, that may come form hypnotising. The process is a fundamental *communication* and can manifest it many ways.
Communication and Higher-Order Systems
Bioelectricity acts as “cognitive glue” in the body, used to merge individual cell goals into larger computational networks. Allows the potential and a clear framework, path forward or path onwards in controlling the direction/type of communication in-between Cells, e.g. when wanting tissue, cells or collectives.
Orthopedic surgery example: Illustrates levels. “Smashing” parts together, then letting the body *heal*. We have no hope of controlling healing from the “bottom up,” is is Motiviated by something; *Biology is more then Simple “Parts*”, having “bottom-up” *only* construction or theory.
Hypnodermatology: Illustrates that higher cognitive levels may exist, not just molecular/biological interventions. The level to access this (goal direction), can vary (even with words), not simple cells, with non-simple agents.
Hypn dermatology as not being a rare occurrence or event. It demonstrates the everyday example, connection and effect of Mind/Cognitive, and its ability to interface and connect *to/with/through* Biology: Example: when you want to do something *your body can listen/react, simply by thinking and *intending*, with No other interaction and requirement*. This is an active process 24/7..
Higher-Order Agents?. Could there exist other Agency, higher in Level; can human societies form them, beyond the known or biological mechanisms? Potentially yes, but maybe limits to knowing.
Potential girdl limits in being a subsystem AND determining your participation in the larger structure is unknown, *unlikely*.
Gap Junctions allow direct signal passage, removing ‘metadata’/owner-identification (leading to no-Me and we; Mind Meld) which create a ‘connectivitiy of consiouness. This may create/induce/allow the nested goal-seking collectives; the core mechanist behind Biology.
Making larger systems through connection might be more efficient, *however* might come at *risk*: A Large Collective might have an interest seperate and disticnt for individuals. For example. Skin: There’s no *Guarantee* that a collective of biological Agents have benefit or ‘goodness’ from its constituents, cells etc..
Cell-Perspective or Cognition may give an idea to what ‘Goal-Direction/Goals’ a Cell might have; e.g. with cancer: cells no longer *care*, and can act and re-arrange the body, for its individual or collective own survival; or no, no good; only, as it relates to ‘reconnectiveess.
Policies might arise that make way to help and maintain Large and Beneficial (not ‘cancerous’) Collective(s) of Organisms (E.G: people); may or should be an *existential-level goal* for Humanity; preserve benefit with no individual downside, or compromise, potentially by creating a collective with no compromise of its individual; hard.
Open to artificially generated consciousness if it embodies similar core biological (and especially metabolic) properties.
Not against artificiallity per-se. Current computers lack, not necessarily “couldn’t” potentially have consciousness; future computers may differ.
The underlying principles are Self Construction (autopoiesis), Goal/survival direction (which makes one care/cooperate/model for “others”/outside”), is fundamental, beyond (potentially) metabolism.
Ethical Implications for the Future; will *need* a new framework beyond current biological frameworks. New Ethics needed: Cognition might be *more* fundamental than previously thought (implications for lower “life”)..
Augmentation will lead to ethical questions on what defines, has worth, and the future meaning of what’s ‘Responsbile’, and, how to deal and manage those things, individuals and types/variances of ‘selves’. The definition will become “soft”, because of modifications.
Levin’s research challenges, *tweaks*, standard neo-Darwinian evolution (while not suggesting a direction/purpose to it); potential ability to engineer that goal.
*Examples/conjectures given:*
Giraffes + Lamarckian evolution, neck problem example (what genes get manipulated). If the Cells *knew/Know* then it’s doable. Might be generalized from “past” problems (in a similar type), potentially allowing for strong Lamarkin Evolution.
Evolution, Intelligence and Mutation *Generalization* (key term): Evolution leverages an intelligent medium (cells). Intelligence could give rise in Cells to *potentially* solve its OWN problems. Levin hypothesizes that cell adaptation is *generalizing from known/past issues*.
Planeria in barium experiment; example. When applied Barium = new (potentially generalized) solution found with (quick rapid response and quick and successful modification; not evolutionary-selectionist)..
Two New Papers: Anthrobots and Embryo Communication
Two papers published: One on “Anthrobots” (human-cell-based biobots) and one on “Cross-Embryo Morphogenetic Assistance” (CEMA) where embryos help each other develop.
Common theme: Understanding biological information sources and collective decision-making in biology, beyond just the genome.
Anthrobots (Gizem’s Paper)
Anthrobots are self-motile constructs made of *adult* human tracheal cells (not genetically modified – “wild type”).
Represent a shift: Viewing nature as a *design medium* (synthetic biology). Biology offers self-construction, healing, carbon negativity.
Challenge to traditional synthetic morphogenesis: Usually focuses on genetic editing, but anthrobots show epigenetic factors are also key. No genetic modification is used.
Process: Human airway epithelial progenitor cells form spheroids. Cilia (hair-like structures) normally face inward. The key was flipping them *outward* for motility. Achieved by removing matrix and using retinoic acid.
Key properties: Fully cellular (no wiring or mechanics), self-constructing from single cells, programmable anatomies (“biobots”).
They can perform useful work, namely, inducing repair in damaged human neuronal tissue *in vitro*.
Anthrobots demonstrate the plasticity of adult human cells, not just embryonic cells (like Xenobots, made of frog cells).
Distinct “morphotypes” (shapes) and behaviors emerge, despite identical DNA, showing epigenetic influence. A relationship exists between shape/structure and resulting movements/motility.
Medical potential: Personalized medicine; using patient’s own cells, potentially avoiding immune rejection. Could target inaccessible tissues, deliver drugs, clear plaques, etc. *Future* research needed for in vivo testing.
Lifespan of weeks to months in cell-culture medium, ending with biodegradation.
Embryos (Xenopus laevis, frog) in groups are *more resistant* to teratogens (development-disrupting substances) than single embryos.
Addresses a knowledge gap: Lateral interactions between *whole organisms* influence development, not just cell/tissue/DNA level interactions.
Significance: If the communication mechanisms, also named as “instructive cues” by the researches can be understood/harnessed to give instruction to cells on repair or growth, leading to huge potential future application in medicien
Contradicts “genome-centric” view: Genome alone doesn’t determine everything; the *social environment* of embryos matters.
Crazy because everything held constant except number of embryos but result show large difference/discrepency.
“Wisdom of the crowds” effect *but*: Mixing teratogen-exposed and unexposed embryos doesn’t help. *All* embryos must experience the challenge for the protective effect.
Robustness increases with group size: Survival *increases*, defect frequency *decreases*. Single embryos almost never survive.
Important implications for toxicology studies: Many studies may *underreport* teratogen effects due to group correction.
Current research: The ‘wave of information” between emryos is hypothesized to be by: an injury on a single cell within the collection of embryos which induces a calcium response to itself; calcium then induces an ATP to release into the media. nearby cells absorb this and themselves emit a calcium response.
Mechanism: Communication is likely via calcium and ATP signaling. Blocking these *reduces* survival, mimicking singleton embryos.
Inter-embryonic communication (signaling molecules) vs. inter-embryonic interaction (embryos growing together, allowing communication).
Not limited to genetic homogeneity: Different frog lineages (wild-type and albino) show similar effects.
Wildtype refers to natural or typical genetic information as they exist in the natural world with no modiciations from labs or scientists.
Transcriptional changes (changes in RNA levels) are observed, indicating different coping mechanisms in large vs. small groups.
Broader Implications and Future Directions
Biobot applications go beyond medicine (as described above) may also find applicatin for construction/architecture; biology can be scaled up.
Medical potential to be faked, emulating signal inducing development and applying it “at-will” and by force, in patience that may have needs which relate.
Basic research: Studying “basal cognition” and “diverse intelligence” in anthrobots (memory, learning, preferences).
Understanding the “cognitive glue” that scales up individual intelligences (cells) to larger collective intelligences.
Bioelectricty may hold importance.
Technological Approach to Mind Everywhere (TAM)
Philosophy drives scientific discovery. Levin’s framework emphasizes understanding goal-directedness to recognize, build, and control unconventional agents.
Anatomical control is an example of collective intelligence navigating “morphospace” (the space of possible anatomical forms).
Bioelectrical networks are an ancient cognitive “glue,” predating brains, enabling individual cells to act collectively. This has implications for biomedicine and synthetic bioengineering.
Beyond Discrete Natural Kinds
Evolution and development are gradual, continuous processes, blurring distinctions between “natural kinds” (e.g., human vs. animal, natural vs. artificial).
We are part of both a natural continuum (evolutionary, developmental) and an engineering continuum (biological modification, technological hybridization).
The framework considers a wide range of agents: familiar organisms, colonial organisms, engineered biological systems, AI, and potential exobiological life. All are analyzed by asking how an external observe would functionally interface with it, and see what best interaction (Hardware/Goal/Reward&punishment) there is to use.
The Spectrum of Persuadability
Systems exist on a spectrum of how best to interact with them, from purely mechanical systems (only modifiable by hardware) to systems that can be reasoned with.
It’s an *empirical* question where a system falls on this spectrum, not a philosophical one. Experimentation, not assumption, is key.
All systems start life with less interaction capibility, but slowly build capacity until there is enough to make its own descisions and plans. Developmental biology offers *no special moment* of “true cognition”; it’s a gradual process.
All intelligence is collective intelligence: composed of interacting parts (cells, components, etc.). Understanding how these parts scale up to form larger intelligences is critical.
Multi-Scale Competency and Problem Solving
Biological systems have a multi-scale competency architecture: each level (cells, tissues, organs) has its own problem-solving capabilities in specific “spaces”.
Examples of these problem spaces: anatomical space, physiological space, gene expression space. We are good at recognizing intelligence in 3D space, but less so in others.
Planaria can adapt to barium exposure, quickly and with no selection, selecting specificly and quickly (within hours) which genes. This exemplifies solving novel physiological problems by navigating gene expression space. Gene regulatory networks (GRNs) have diverse learning capabilities, including associative conditioning.
Evolution repurposes problem-solving strategies across different spaces.
Bioelectricity and Morphogenesis
Turing’s interest in morphogenesis was likely linked to his interest in unconventional intelligence. Body and mind building are related problems.
The genome specifies the *micro-level hardware* (proteins), not the large-scale anatomical structure. Cells *collectively* decide what to build and when to stop, demonstrating morphogenesis as collective intelligence.
The goal is an “anatomical compiler”: translate a desired anatomical form into stimuli that guide cells, revolutionizing medicine. This is a *communication* problem, not a micromanagement problem.
Cells and tissues exhibit intelligence (defined as “reaching the same goal by different means”). Development and regeneration demonstrate robust error minimization and adaptability. Kidney tubule formation shows different mechanisms to achieve the same anatomical outcome.
The frog face rearrangement demonstrates a goal, not predetermined organ-by-organ programming.. Perturbed tadpole faces (“Picasso tadpoles”) can still form normal frogs. An error minimization, not instruction.
Bioelectric patterns, like those in the brain, store “set points” (anatomical goals). They aren’t magnets, fields, or anything alike. Instead, cells use and hack other cells, leveraging the native interfaces that each other use and provide to communicate this electri network, creating pattern completion that we are not smart enough, nor need, to understand the internal working for us to communicate at large.
Tools from neuroscience can *read and write* these bioelectric patterns (like “incepting” false memories). Altering bioelectric patterns can induce ectopic organs (eyes, fins, hearts) and influence regenerative processes.
Planarian regeneration reveals bioelectric gradients determining head number. These patterns are rewritable and act as a “counterfactual memory,” dictating future regenerative behavior even when the body is normal.
We can computationally model the connections between electrical states and the underlying molecular mechanisms, drawing on ideas from connectionist neuroscience.
The latent morphospace of possibilities for even a given genome is vast, highlighting the plasticity and flexibility of biological systems. The plan is a future medicine more like sematic psychiatry (than only chemistry, and fixing a gene.
Cognitive Light Cones, Selves, and Ethics
The “cognitive light cone” is a central invariant: the spatial-temporal size of the largest goal a system can pursue. Different agents have different sized light cones.
Defining a “self” involves considering the boundary of goals the system pursues. We are all collectives (cells within organisms).
Early embryogenesis reveals dynamic self-construction. The number of “selves” in a blastoderm is not predetermined; it’s an outcome of cell communication. The same logic applies to nervous system (split brains show there is not clealy just 1 self inside).
Cancer can be understood as a failure of the scaling-up of cellular goals, cells are reverting to a smaller and smaller individual cognitive lightcone, thus are cancerous as they dont abide to collective goals. Bioelectric interventions can influence this process, even when genetic abnormalities remain.
Endless forms are possible, both naturally and through bioengineering, due to the interoperability and plasticity of life. It is important that novel forms do not fit onto the evolution tree/scales and that all categorization fails here (desiged, natural, etc)
This necessitates new ethical frameworks for interacting with unconventional minds.
Horizontal Information Transfer in Embryos
Research on “hyper embryos” shows that large groups of embryos resist teratogens (development-disrupting agents) better than small groups or individuals.
Embryos communicate with each other using ATP and calcium signaling, and share information, enabling a group-level response to stress, going beyond individual genomes.
The only individuals who improve resilience, are those also subject to stressor exposure.
This has implications for toxicity studies: The toxicity/benefit result is a factor of both group response and intrinsic risk/benefit, meaning the data collected by such studies do not nececarilly represent ground truth about those studies alone.
Suggests a “hyper embryo” exists as a higher-level entity with its own dynamics and gene expression. Implies development isn’t solely vertical (from parent) but also horizontal (between peers) and cooperative.
Emergent Behavior in Simple Algorithms
Studied simple sorting algorithms to find minimal conditions for surprising emergent behavior, addressing reductionism vs emergent complexity.
Injected individual element awareness such that rather than 1 algorithm for whole array, each element ran 1 simple algorithm locally, just seeing neighbours.
Created self-sorting arrays where each number “wants” to be in the correct position relative to its neighbors.
Found that even with broken/immobile elements, the algorithms exhibit “delay of gratification” – temporarily getting less sorted to ultimately achieve sorting, without explicit coding for this.
Created “chimeric” arrays where different elements follow different sorting rules; these show unexpected clustering behavior (“algotypes”) even when not advantageous, and persist longer in cases where sorting tolerance increased, representing ‘free will’ in the model, the cost of ‘physics pulling them apart’, decreased.
Emergence can exist, surprisingly, in highly-reduced conditions and environments, suggesting that we should not assume human/animal/biological properties only exist when “obvious” based on the typical components known for those organisms (nervous systems).
Discussion on “What Algorithms Want” (Epistemic vs. Ontic)
Debate on whether emergent properties (like “delay of gratification” in algorithms) are real (ontic) or just projections of our cognitive modes (epistemic).
Levin argues multiple perspectives exist, and the perspective of goal-directedness can be useful for scientific discovery, even if not “fundamental” in the lowest-level reductionist sense.
Kastrup contends complexity science often shows that apparent complexity results from simple underlying rules, cautioning against over-attributing cognitive properties, suggesting a lot of observed agency is misatrributed and that what we refer to is purely an epistemic and convenient simplification, instead of ontic truths about reality.
The Problem of “Thingness” and Points of View
Is there an objectively definable “thing”, or is this purely observer dependent based on current capabilities and understanding of that observer?
If all “things” (even tables) are assigned their own cognitive agent perspective from observer: observer faces information and combinatoric explosion.
Analytic idealism frames point of view as arising from dissociation; a living zygote *is* a point of view (distinct from the world), undergoing internal complexification.
Levin does not accept binary of view/no-view but accepts that living zygote represents a “view”, a “thing”. He thinks it does not only extend to biology: “thingness” a function of degree, that many levels (degrees) of “thingness” may coexist inside bigger “things”.
Abiogenesis (life from non-life) presents a greater challenge, creating a point of view from non-view entities, unlike fertilization.
Continuum of Perspectives and Sense-Making
Levin emphasizes a continuum of perspectives, rather than a binary “has it/doesn’t have it” view of consciousness or agency. The size/signifiance of cognitive perspectives can and will differ greatly.
Rejects idea of *objective* criteria for defining an agent/thing; criteria are relative to an observer’s perspective, and their primary goal, driven internally by self-preservation, is: sense-making.
“What it’s like to be” questions are about a relationship between two systems; truly becoming another system means losing the original perspective, but instead represent an offset comparison or merge, but the target is retained.
An explanation helps improve sense-making of the observer to aid effective actions on future, therefore all “useful” and functional descriptions have a quality of *future prediction*, which should not be mistaken as *just* past/present descriptions.
Levin: the world a giant set of perspectives, growing and shifting, making estimations/assumptions about “other”s in environment.
Implications for Biology and Evolution
Levin discusses memory remapping, exemplified by caterpillar-butterfly metamorphosis: information is compressed and reinterpreted for a new life context.
Evolution likely selects for mechanisms of perspective that extract salience from experience, not rigid replication, allowing flexibility/interoperability, hence resilience and ease-of-combination of cells.
Anthrobots (made from adult human tracheal cells) demonstrate emergent behaviors like self-motility and neural wound healing, despite no evolutionary pressure for this.
Suggesting capabilities or potentials latent in structure space; exploration through “periscopes” created using perturbational experiments of system.
Speculation on a Platonic Realm
Suggests a latent “Platonic” space where capabilities exist as forms, instantiated when physical structures align; evolution “searches” for pointers into this space.
Levin imagines space as interaction/dynamics between entities (similar to how philosophical concepts have shapes).
These forms are not static, but potentially dynamic, influenced by interacting systems; the “frame rate” of interaction depends on the observer’s capacity.
Discusses philosophical concepts including L. E. J. Brouwer and intuitionist logic.
Metacognition, Active Inference, and the Physics of First Persons
Integrated Information Theory (IIT) is consistent with multiple complexes within a living being; subsystems are subsumed, not eliminated.
Levin suggests active inference (surprise minimization) and Markov blankets may lead to a more refined definition of “life”, potentially including things not currently considered alive, not by explicit components, but by organizational rules and properties of *interactions*.
“Physics of first-person perspective”: Rewriting physics in terms of predicting one’s own next experience, and connection to actions that can influence that, which aligns with Carl Friston, and Chris Fields’ work, “Active Inference”.
Introduction
Significant knowledge gaps exist regarding the control of large-scale anatomical homeostasis (how organisms maintain their shape).
Fundamental advances in biomedicine require understanding cellular decision-making, not just molecular mechanisms.
Non-neural bioelectricity is a key medium for computation in cellular collectives.
The goal is “cracking the bioelectric code”: mapping electrical patterns to genetic and anatomical outcomes. This could enable “electroceutical” approaches for various medical applications.
Body tissues form electrical networks (like the brain) that make decisions about dynamic anatomy. We have ways to control this for pattern editing.
Anatomical Homeostasis and the Anatomical Compiler
The long-term goal is an “anatomical compiler”: specify a desired anatomical structure, and the system generates stimuli to guide cells to build it. This would revolutionize medicine.
Cells are not passive building blocks; they are highly competent, making decisions and solving problems.
Embryonic development involves cells working together towards large anatomical goals. Differentiating cell type isn’t enough; 3D organization is crucial.
The standard developmental biology paradigm (gene regulatory networks -> emergence) has a profound inverse problem: it’s hard to know what to change upstream to get a desired downstream effect.
Current models often lack predictive power. Even in simple examples, the precise algorithms and stop conditions are unknown.
Moving focus from hardware and focusing to also ask about the anatomical, native software modules within and asking how reprogramable the system.
Plasticity and Pattern Homeostasis
Examples of biological plasticity are used. The axolotl regenerates limbs, etc., The planarian flatworms are a system to understand plasticity due it its high ability for regneration.
Planarian regeneration: Each fragment “knows” what’s missing and regenerates it. They are “immortal” due to continuous regeneration.
Human liver regeneration, deer antler regeneration, and fingertip regeneration in children show that regeneration is not limited to “lower” animals.
Picasso tadpoles: Frog facial features rearrange to form a normal frog face, even when starting from abnormal positions. Genetics specifies error minimization, not a fixed path.
Pattern Homeostasis model introduced a feedback loops from anatomical devation to the genome via electric.
The model set-point which includes The process of “current read, compare to a set-point, action, repeat.” and this feedback includes not the current levels of parameters, but a fairly complex specification on layout of an antamoy.
Bioelectricity as a mechanism, an explicit representation of the system exists of the future anatomy.
Bioelectric Signaling
Bioelectricity is a key component of the “morphogenetic field” – a field of information that influences cells.
It is stated, Bioelectricity is not a standalone: other things are involved including chemical and physical factors and gradients.
This is a very convinent compuational layer used, not accident, by evolution for decision-making.
All cells have ion channels and many are electrically connected via gap junctions (electrical synapses), similar to the brain.
“Neural decoding” in neuroscience tries to infer the informational content from electrical activity. The same principles can apply to other tissues.
Voltage-sensitive dyes map electrical, simulating by computer is done by building models with cells, gap, and channels.
Early from embrionic face developement has a bioelect patten as a prepattern on where organs go, this pre-pattern then influnces gene expression.
An examle is also made on the bioelectric expression of tumuors and an example of diagnostics.
Membrane potentials can influence.
Modulating Bioelectric Circuits
Tools have been developed to manipulate bioelectric signaling:
Changing connectivity by opening/closing gap junctions.
Setting cell voltage by controlling ion channels (including optogenetic channels). These are molecular, not external electrical field, interventions.
Induced voltage in areas to make modular “cascade sub-routines” making a voltage that is inductive and kickstarting body formation (an example is ectopic eyes, made using electirc signal).
Standing bioelectric patterns are manipulated to make large anatomical change: This manipulation creates planeria that regrow and show it has the memory of its biolectricity even if cut, or having the planeria regrow into a similar yet differnt head species.
An example of medicine use is regrowing frog limbs using this voltage trick with chemicals and bioelectric to influence modularity of limbs.
There is a spinoff, mophaceuticals inc, doing this.
Many mechansims between biolectricty and gene expresion in single cells.
Modeling and Control
Multi-scale models combine single-cell bioelectric and transcriptional circuits with tissue-level dynamics. This helps understand the algorithms for anatomical control.
These models predict experiments, allow virtual experiments, and are amenable to machine learning. This facilitates finding specific interventions.
Focuses not on specific individual cell voltage, but its gradient among neightbors.
Evolution uses different methods of voltages based on convenient factors.
Encoding Metaphor and Memory
DNA sets cellular hardware (ion channels, etc.), but bioelectric states represent a kind of “software” due to post-translational modification.
The bi-electric patten can persist even when removed the organs, creating a future set point in time and space to regrow into.
There’s long-term “bi-electrical” memory, which is seen downstream in the long-term via change of structure and shape.
Towards Electroceuticals and Beyond
With ion and cells and votlages we get “comptuation model for predicting outcomes.
Examples include frog brian, planaria bi-stability, regeneration, and tumor regulation.
Toward Therapudic Platforms
Computational models can predict ion channel combinations to shift bioelectric patterns towards health. A platform is being developed.
Key Conclusions
A bioelectric computational layer sits between genotype and anatomy, making crucial decisions.
Evolution likely used electrical signaling for computation very early.
Understanding this bioelectric “language” is crucial for controlling collective cellular behavior and has broad implications.
Anatomical Homostasis by cells and feedback, making bioelecticity an early computation level used.
Introduction: Turing, Intelligence, and Morphogenesis
Alan Turing was interested in both AI and morphogenesis, seeing a deep parallel. Levin believes these are fundamentally the same question: problem-solving in different spaces.
It’s important to view intelligence not by scale, as everything (even human brains) comprises a larger collection (colony/collective intelligence) of smaller ones.
There exist Smooth Cartesian transitions, starting from physics & chemistry which result in cognitions like: awareness & metacognitions.
Core Concepts: Multi-Scale Competency and Navigation
Biology uses a multi-scale competency architecture: nested problem-solvers at different levels (cells, tissues, organs, organisms) each with autonomy and goals.
Navigation, particularly of “spaces,” is central. These spaces aren’t just 3D physical space, but also physiological space (chemical parameters), transcriptional space (gene expression), and morphospace (anatomical configurations).
Agents pursue *goals* relevant to them, despite not having the biggest scope (i.e., skin cell regeneration, even if that conficts with the higher level).
Goal directed-ness should allow relations to *any* system.
“Cognitive boundary” describes goal-scale.
Goal-directedness is key for understanding and interacting with unconventional agents.
Bioelectricity and Morphogenesis: A Detailed Example
Biological pattern formation is the behavior of a collective intelligence of cells navigating morphospace.
Bioelectrical networks (precursors to brains) are the medium for this proto-cognitive activity. Cells communicate electrically via ion channels and gap junctions. This isn’t just philosophy, it has practical implications for biomedicine.
Examples, cells handle local tasks: metabolic, morphogenesis, and behavior tasks.
The morphogensis system does: vibrations and vibrations and soner sensing, creates a “map,” make decisions with respect to its environment.
Beyond Natural Kinds: Plasticity of Agents
Agents are not fixed entities. Examples:
Caterpillar to butterfly: Radical body and brain reorganization, yet memory persists.
Planaria regeneration: Regrow any body part, including the brain, with memory retention even after head amputation. Information transfer between tissues.
Tadpole eye plasticity: Eyes grafted to the tail can still provide vision, even with novel neural connections. The brain adapts to new sensory input locations.
Biological systems are nested structurally (cells, tissues, etc.) and functionally: each level has its own competency and solves problems in its space.
Beyond Traditional Spaces, Expanding Our Notion of Spaces
Intelligence isnt only physical (3D). We can generalize intelligence by observing actions in non-3D, example, sensing and reacting to physiology states of the liver.
Planaria can navigate a Barium solution despite the extreme dangers, and it takes a handful of the ~20k possible genes to allow Planaria to live through this process, showing non-random processes are taking place.
Intelligence as Problem-Solving, and Creating Systems
Intelligence: the ability to creatively use new/existing informatoin creatively in new senarios, not just using pre-exisiting answers.
Developing new cogntivie systems will require finding all the answers as nature can evolve solutions.
TAME system (technological approach) needs a wide array: human to animal, but even new ones created in labs or by systems found outerspace, allowing us to recongize and compare them.
Navigating Morphospace: Goals and Homeostasis
Cells *know* how and when to make the correct structure; regenerative system is similar but is about a signal *for* building.
How cells work is a large scale question and *cant* be directly coded, example: you cannot tell Frogolotts grow leg by using genes; an unknown is whether Frogolott leg contains either axolotl or Frogolott.
A long term is the Anatomical Compiler to make limbs/structures/shapes. It could revolutionize medicicine by being able to repair cancer, tramua, age-related illnesses, etc., etc.
Morphospace is the space of all possible configurations for a structure. Embryogenesis is remarkably reliable at navigating morphospace, but it’s not hardwired. It’s homeostatic.
Examples of morphogenetic homeostasis:
Monozygotic twins: Splitting an embryo results in two normal organisms, not two halves.
Axolotl limb regeneration: Regrows exactly what’s missing and then stops, demonstrating anatomical homeostasis.
Human liver regeneration, deer antler regeneration, child fingertip regeneration.
Newt Kidney Tubule: Cells adjust in size and number (even molecular mechanism) to form the correct lumen.
Frog-Leg Regeneration. Normal Frog limb Morphogensis will occur.
Picasso frogs: Messed up facial features still migrate to form a relatively normal frog face, showing error minimization, not hardwired movements.
Bioelectricity as the “Software” of Morphogenesis
The standard developmental biology model needs feedback loops. This creates a homeostatic system. The set point is a complex data structure: a large-scale geometry (anatomical descriptor). This introduces goal-directedness, which is often avoided in biology.
Brains are a good example of navigating space to a goal. Bioelectricity offers a similar mechanism:
Cells communicate electrically in networks (like neurons).
Ion channels set cell voltage.
Gap junctions allow electrical communication between cells.
The commitment of neuroscience is decoding: electricial patterns representing the content: goals, memories.
Bioelectrical signaling *before* genes for development.
Neural electricity is an evolutionary pivot from morphospace to 3D space.
Tools to Read and Write Bioelectrical Information
Voltage-sensitive fluorescent dyes reveal electrical conversations between cells.
Computational modeling predicts voltage patterns from ion flows.
The “electric face” in frog embryos: a bioelectrical pre-pattern that precedes gene expression and determines facial structure. Disrupting this pattern alters development.
There exist Pathological electrical patterns.
Tools to write: Modifying ion channels and gap junctions (like neuroscientists) to alter electrical states. No external fields/waves used, only endogenous mechanisms.
Practical use of Bioelectricity, example: induce and suppress malformations/illnesses such as cancer, eye generation.
Examples of Bioelectric Control
Inducing ectopic eye formation: A specific voltage pattern triggers eye development, even in inappropriate locations. This instruction is modular (doesn’t specify *how* to build an eye).
Cells recruit their neighbors.
Induction of ectopic organs: otocysts, hearts, forebrain, limbs, even fins (which tadpoles don’t normally have).
We have drug coctail through ion chanels to produce leg.
Regenerating frog legs (which frogs don’t normally do) using a drug cocktail that targets ion channels.
Altering planarian head number: An electrical circuit stores the “memory” of how many heads to regenerate. This can be reprogrammed, creating two-headed worms that continue to regenerate as two-headed even without further intervention. Non-genetic inheritance.
Exploring Morphospace and Attractors
Tweaking the electrical circuit can lead to head shapes of other planarian species, or even novel shapes not found in nature. All with the same wild-type genome. This reveals different attractors in morphospace.
“Full stack” models: Integrating molecular information (channels), physiology (voltage states), organ identity, and algorithms to understand the system and intervene. Computational platforms simulate tissue-level dynamics.
Repairing birth defects: Using bioelectrical models to identify interventions (drugs targeting ion channels) to restore correct brain development in tadpoles with teratogen-induced or genetic defects. “Software” fixes for hardware problems.
Electroceuticals can help use bioelectrical *interface.*
Scaling of Cognition and Goals
Borders change, between, and there is plasticity: individual organisms and the colonies.
The boundary between self and world is flexible. Goal pursuit unifies diverse intelligences across scales.
Single cells have small goals (microns).
Cell collectives have larger goals (building a limb).
Scaling of self is present in Cancer cells, whose electrical communications separate with a colony that exists within a bigger system.
Cancer can sometimes be reversed by reconnecting cells electrically, emphasizing the importance of collective state.
Any cognitive agent has some ability to care, where the abilities increase linearly but are cut off due to our capacity.
A Cognitive Light Cone Model
A way to compare diverse intelligences: the size of their “cognitive light cone” – how big (in space and time) are the goals they can pursue.
Tick: Small goals (local butyrate concentration).
Dog: Larger goals (memory, some anticipation).
Human: Goals potentially larger than lifespan.
Multi-scaling includes influence higher > lower organizational levels, e.g. genes expression is contorlled to “steer” systems.
Multi-scale system: Higher levels “bend” the option space for subunits, guiding them to the overall goal.
There is an example of this cycle for PH where each cell can adjust and do cycle processes; this continues up in higher structures and system, with even *evolution* pivoting it through different states/structures: behavior spaces to other, undiscovered areas.
Xenobots: An Extreme Example of Novelty
Xenobot can arise spontaneously, no additional interference necessary: e.g. no cirucit/chemicals/instructions necessary, subtracting constraints *does* matter.
Xenobots are made from frog skin cells, allowed to “reboot their multicellularity” without the normal embryonic context. They demonstrate:
Spontaneous self-organization.
Novel behaviors: Movement, navigation, calcium signaling (brain-like activity, without neurons).
Zenobot will regenreate, example: split it down half, see hinge clamp together, see how Zenobot retains this original shape.
Kinematic self-replication: They build copies of themselves from loose cells – a behavior not found in frogs.
Collect together into piles
Engineered to shape environment, which then the Zenobots would use it for another function.
One Xenobot genone (frog’s Xenopus lavis) is also two other stages/behaviors
These behaviors are not pre-programmed or selected for; they arise from the cells’ inherent plasticity. Skin cells *want* to be xenobots. Evolution is behavior shaping: finding signals to coax agential materials.
Ethical Implications and a Future Taxonomy of Minds
Viable Agents can comprise any materials, evolution/design: organic and inorganic materials/hybirds etc,
Biology is incredibly interoperable. Any combination of evolved, designed, and software components can be a viable creature. We’re entering a world of unfamiliar agents.
Ethics should be based on an agent/ability/functionality, no origin-story.
Traditional ethical frameworks (based on origin and appearance) will become useless. We need new frameworks based on the capacity for sentience and goal-directedness.
Bioelectricity (though the examples) shows great areas to explore with systems, and there could be many others we don’t yet.
Closing: Goal Directedness and Future Research
Can there exist framework without *cultural* ideas like evolutionary natural history to *universially* apply like a mathematical axiom? Active-inference model from physics is close (cognitive functions emerge from minimize “surprises”/expectations).
Intelligence could possibly organize through space by its scale, or other properties as well as their wavelengths.
Relationships between systems (animals, plants, systems) might have hidden geometrical information.
Exploring relationships with other non-animals. e.g. plants.
High order-connectivy. Networks don’t show the whole picture (but this is hard to definitively define in this).
A goal to see more novel solutions beyond evolution or other limited ideas, especially within *anotomical compliers.*
There may be no such thing as *regenerations.*
For cells: new cell can be *envrionments* for one another.
Mind may arise: flat (cognitions rise at once).
A general framework: test-compair-act and you get emergent *affordances,* where example, transitors do many amazing tasks/configurations like logic gates without this originally programmed in them, or the mathematical operations that will add all the triangle’s degrees = 180.
Taxonomy of Mind could be: light cones: ability to see the highest end, biggest goalls.
The work focuses on a continuum of agency, scaling of goals, bioelectricity as a cognitive medium, and implications for evolution, biomedicine, and understanding/creating diverse intelligences.
All of these things we use today might have to be *increased* to adjust/account/plan. Example, maybe someone will literally care/understand all organisms instead of beign “linear,” there are other shapes too.
Introduction and Core Concepts
Discussion on the intersection of Michael Levin’s work on bioelectricity and morphogenesis and Bernardo Kastrup’s analytic idealism (a form of objective idealism). The core idea is that nature is fundamentally mental, but not in the sense of individual minds; rather, it consists of “mental states at large.”
Kastrup uses psychiatric dissociation (dissociative identity disorder, formerly multiple personality disorder) as a model. Dissociative processes have a physical appearance when imaged (e.g., brain scans). He proposes that life/biology/metabolism *is* the appearance of dissociation within a transpersonal field of subjectivity.
Levin’s work focuses on the scaling of minds and how the boundaries between a “self” and the “outside world” (or other minds) are formed. He, simliarly to Kastrup, talks of dissociation in cells.
Embryogenesis, Dissociation, and Boundaries
Embryogenesis begins as a single “blastoderm” (a flat disc of cells). It’s not genetically predetermined how many individuals will arise; it’s determined physiologically. Scratching the blastoderm can create separate, self-organizing embryos (and eventually conjoined twins).
This illustrates the plasticity of boundaries and the initial “pool of potentiality.” Cells must “decide” what constitutes “self” versus “world.”
Cancer is viewed as a “somatic dissociative disorder” where cells shrink their “cognitive light cones” and treat the rest of the body as external environment. This is a failure of the cognitive “glue” that binds subunits.
Levin see’s cancers as somatic dissociative disorder of collective intelligence, agreeing with Kastrup.
Morphogenesis and Top-Down Causation
DNA provides the “bricks” (proteins), but something else determines how those bricks are assembled (e.g., into Cologne Cathedral, not just a pile of bricks).
Embryos have a “goal” – a specific journey in anatomical morphospace. Individual cells don’t “know” the final form (e.g., a salamander’s limb), but contribute to a large-scale, cybernetic system.
Levin proposes “top-down causation” where the higher-level goal (target morphology) shapes the “energy landscape” of lower levels (cells and molecules).
This goal isn’t captured by analyzing just by zooming into, local interactions.
Decoding Bioelectric Patterns
Levin’s lab uses techniques from neuroscience (neural decoding) to “read” the large-scale bioelectric patterns of tissues. These patterns reveal the “goals” of the tissue and predict future development.
They can reprogram these patterns (e.g., making two-headed planaria) by manipulating the bioelectric interface (ion channels and gap junctions) – *not* by changing the DNA. The revised “target morphology” is then maintained.
Where the large scale ‘goals’ orginate: It’s probably beyond evolution and related more with mathematics’ rules and constratins in a way.
Novel Creatures and Plasticity
Evolution creates *problem-solving agents*, not just solutions to specific problems. Life exhibits incredible plasticity.
Examples: Tadpoles with eyes on their tails can see immediately (no evolutionary adaptation required). This challenges the notion that evolution is the *sole* source of these patterns.
The “latent space” of possibilities is much larger than what we typically observe. Examples include plant galls induced by wasps, revealing potential forms not directly encoded in the plant’s genome.
Hardware, Software, and Problem-Solving
While DNA provides essential “hardware” (proteins), it doesn’t fully determine outcomes. The “hardware/software” analogy is helpful.
Example: *Nematostella* embryos can have varying numbers of chromosomes (4N, 6N, 8N, etc.) and still develop normally. Cell size adjusts, and even single cells can form tubules by bending. This demonstrates problem-solving at multiple levels (cell-cell communication, cytoskeletal bending).
Embryos have a “beginner’s mind,” adapting to what they have, not overtraining on priors. This challenges the notion of fixed, genetically determined programs. Two-headed planaria have no genomic differences; the two-headedness is encoded in electrical memories.
Cancer cells can be coaxed to behave in ways conducive to the over-arching, healthy tissue when its bioelectircal signals are tampered with.
Planaria and Memory Imprint
Planaria trained, headless, regenerate new brains that, can remember trained skills: which suggest, info imprint of skills (from previous brain) being implemented on new brain.
Imprinting must of been performed by rest of the old body. But, not sure where.
Chimeras, Composite Beings, and Boundaries
Embryos can be *combined* (not just separated) to form single individuals. Xenobots are a simpler model, where dissociated skin cells self-assemble.
There might be something special about lineages (continuous self-assembly), but the boundary between self and world is highly plastic. The number of “selves” can change over time.
Chimeras (e.g., cats black on one side, orange on the other) have different DNA. You can combine frog and axolotl cells. Life is remarkably interoperable due to its problem-solving nature.
Later, cells and tissues can still connect via biological interface to create functional organization or units. Mice, who share linked brains from thousands of miles away can cooperate to solve tasks.
Unified Consciousness in Xenobots?
There is coherence.
Xenobots (at least the skin-cell-only kind) *may* have unified consciousness. Levin’s lab is investigating this using measures of integrated information processing (like those used in neuroscience).
Looking for behaviors (in their unique space).
Behavioral and physiological evidence, including calcium activity patterns, will be used to assess the level of integration. The question is whether cells are acting separately or as a coherent whole.
Potentially sleep patterns may be used to determine this too.
Internal Dissociation and Attention
Even within humans, there are “internal dissociative boundaries.” We don’t consciously introspect into liver or kidney function, likely for evolutionary reasons (efficiency).
Analogy: Riding a bike becomes automatic (dissociated) after training. We *can* take deliberate control, but it’s inefficient. Breathing is similar.
The Exclusion Principle of IIT (Integrated Information Theory) could relate to this. Attention restricts experience to a narrow area, “boosting” it at the cost of peripheral vision. High information integration in certain brain areas (e.g., default mode network) may lead to exclusion of other processes.
All parts of a body may show signals related to intelligence and consiouness (not just brain) because evolution may act and promote cognitive properties/traits in other tissues and cells, just at a slower pace.
Cancer represents a *fundamental* dissociation, where cells no longer follow the global template. This is distinct from the “normal” internal dissociation between, say, the brain and the liver.
Organs can also compete with each other to be more influencial in a developing being, especially early on in development.
Metamorphosis, Memories, and Past States
Metamorphosis presents another major change that organisms/agents may encounter during its existance, for instance, caterpillers that change completely in a hard-body butterfly from soft-bodied state.
Caterpillar/butterfly metamorphosis involves significant brain restructuring, yet memories can be retained. This raises questions about the location and nature of memory.
Kastrup speculates that if memory is access to *past states* (which are not physically observable), then it may not have a physical correlate. This contrasts with the view of memory as a “file on a hard disk.”
It would be hard to study in Planaria, Rat, etc because they are non-physical phenomenon in nature.
Kastrup discusses and debates panpsychism versus consiousness.
Paper 1: Bioelectric Networks & Cognitive Glue
Cognitive Glue: A mechanism that binds individual cells with their own agendas into a unified “self” with distinct goals, especially important in transitioning from a mere collection of cells to an organism. This isn’t a “religious” miracle, but a profound scientific question.
Neuroscience analogy: Electrochemical signaling in the brain is the known “cognitive glue” for behavior. Levin proposes bioelectricity plays a similar role *before* brain development.
Morphogenetic agent: During embryonic development, cells cooperate towards a specific “target morphology” in anatomical space. Bioelectricity serves as the cognitive glue facilitating this cooperation.
Anesthesia example: General anesthesia inhibits gap junctions, temporarily dissolving the “cognitive glue” and causing loss of self, while individual cells remain functional. This highlights the need for *informational* proximity, not just physical proximity.
Problem spaces: Evolution navigates different problem spaces: metabolic, physiological, transcriptional (gene expression), anatomical (body shape), behavioral (3D movement), and potentially linguistic.
Evolutionary pivots: Once an organism is good at navigating one space, evolution can relatively easily “switch” it to another by altering sensors and effectors, leveraging existing competencies (e.g., tadpole eyes on tails, sensory augmentation).
Brain vs. Body: Brains have evolved for speed (faster bioelectricity) and direct long-distance connections (axons). While sharing components like ion channels and neurotransmitters, brains uniquely (as far as we know) enable syntactic language.
Collective Intelligence: The talk touches on importance that when people are exposed to their research that all biological organization contains collective intelligence is to understand not the self as being an “illusion” but that it expands the sense that the self can become very big.
“Play the Hand You’re Dealt” (PHD): Organisms, particularly in development, have robust algorithms that build towards the target morphology *despite* variations in starting conditions (chromosome number, cell size, mutations). Planaria are a prime example, accumulating mutations but maintaining perfect regeneration. This means there exists an ability to cope with suprising novel situations with problem solving behaviors that it itself didn’t not learn over an evolution timeline (rather, something like the collective intelligence found that soltuon)
Quantum Biology: Levin doesn’t specialize in it, but suggests any quantum effects, if significant, would likely be fundamental and pervasive, not just special adaptations. He refers to work by Chris Fields on observer reference frames in quantum mechanics.
Trophic Memory (Deer Antlers): Deer antlers re-grow with ectopic branch points at the location of previous year’s damage. This implies a long-term, large-scale spatial memory mechanism *not* explainable by current molecular pathways. Similar to the persistent two-headedness in planaria.
Paper 2: Darwin’s Agential Materials
Spectrum of Agency/Persuadability: All systems, from Legos to humans, exist on a spectrum of autonomy and problem-solving capacity. Engineers must choose appropriate tools based on this.
Engineering with Agency:
Legos (low agency): Require complete, direct control; every action must be specified.
Thermostats (simple agency): Have a set point and self-regulate; require cybernetic understanding.
Animals (higher agency): Require training/persuasion, but offer resilience and emergent behavior.
Evolutionary Applications: This material can be used and has properties such as memory and learning so we could design with them to, say, have materials and organizations in novel situations adapt their physiology accordingly in novel ways (such as morphologically).
Regenerative Medicine: The key to regenerative medicine lies in recognizing the *appropriate* level of agency of cells and tissues, not treating them as mere passive components. It requires tools from cybernetics and behavioral science, not just molecular biology.
Evolution as an Engineer: Evolution itself acts as an engineer. The paper explores the implications of evolution working not with passive “Legos” (dumb materials), but with *agential* materials (cells with inherent competencies). This massively impacts evolutionary processes.
Evolution’s “Search Space”: Because cells *already* possess problem-solving abilities, evolution searches a smaller, easier space of *behavior-shaping signals*, not the vast space of all possible molecular configurations.
First-order and Second Order Intelligence: second-order intelligence that goes way beyond what first-order systems can ever produce, to produce highly-reliable and efficient navigation of the “play-the-hand-you’re-dealt” property (first-order systems can be a little bit robust in terms of how hardwired they are. So a hard-wired automaton might, under very narrow circumstance, solve certain challenges in the way that we talk about robustness),
while evolution with agential materials means, you can actually solve novel problems: you can do well where,
before your ancestors could not, this isn’t just, oh it sort of solves a simple homeostatic issue, this is now able to face situations.
Evolutionary Algorithm of substrate (Collective Intelligence of the Agents in that organism/orgnaizaton): the intelligence of cells (agential materials), their pre-existing goals/competencies are cruical, often negelcted components. This view on “intelligence of the substrates” suggests a broader intelligence landscape, hinting at a non-zero agency for evolution itself, not as a conscious designer, but as a process capable of hypothesis generation and refinement.
Evolution as an Agent (Proposed Future Paper): Levin suggests viewing an evolutionary lineage (e.g., 50 million years of alligators) as a long-lived, spatially large “agent” continuously generating and testing hypotheses (offspring variations). This is not about a human-level “purpose,” but a non-zero level of agency in the evolutionary process itself.
Paper 3: Biology, Buddhism, and AI (Care as a Driver of Intelligence)
Care/compassion: How the care concept is approached by biology (Levin Lab) vs Buddhism: Biological study of how single cells transitioned into collectives, how this effects/changes “what does it care about,” Buddhist view about the philosophy and care/compassion.
Care: Agents as an embodied and situated: The authors examine different agent frameworks, such as reinforcement learning agents, generative adversarial networks (GANs),
free energy principle, the theory of the origin of life in a care perspective, in that we propose “caring,” more precisely, the “self-organization” of life begins with chemical interactions,
driven by autocatalysis (chemical reactions that form compounds).
Memory, care and robotics: Where should care arise/where will it be emmited when discussing how it has emmitted in collective biological organizations.
Paramecium do or do not “care”? If they do, then, it begs the questions “do the constituent componenets that make it do,”
if paramecium don’t, what state during the transition to a large intelligence makes care (a new concept not there) appear?
How to put that capability to have more broader compassionate cognitive capability to AI/Robots?
Construction of Self: The self is not a fixed entity, but a continuously self-constructed process, rebuilt at every moment from memories and past experiences. This aligns with aspects of Buddhist thought on the impermanence of self (though Levin is not a Buddhist scholar). This addresses and combats that nihilistic point.
“False” Memories: The planarian brain, regrown after decapitation, has memories “imprinted” on it that it did not experience. Levin argues this is analogous to our *own* continuous self-reconstruction from memory, challenging anxieties about “false” or downloaded memories.
Bodhisattva Vow: The commitment to expand one’s cognitive apparatus and “light cone of compassion,” to enable larger goals. This reflects an upward trajectory once a system has the capacity for this commitment.
and ability to improve this capacity/commitment through conscious dedication/actions (consistent practice).
Expanding Sense of Self: The discussion touches on (but does not deeply explore due to Levin’s admitted lack of expertise) how meditative practices might alter the *perception* of self and interconnectedness. The example being The bodhisattva takes the vow
to delay their “Enlightenment” (in Buddhism).
Additional Points, from the end of the Interview
Art (surreal, ai generative (mid-journey): To illustrate the paper topics/conecpts Levin is thinking about and explore,
and express ideas in unique, memorable visual form, often referencing biological concepts like cognition, growth, and
and regeneration.
How paper/findings are recieved: Doesn’t have good measures other than anecdotal feedbacks he had heard from conversations.
drmike11: his academic website drmike11.
org and at Dr Mike 11, in the next couple months,
there will be a WordPress site: non-journal publications and discussions to express new papers that won’t get published at normal channels.
Initial Fragmentation of Consciousness
Levin and Kastrup discuss the idea of individual minds as “fragmented alters” of a larger collective intelligence.
Kastrup speculates that the initial fragmentation may have been an “amazing accident,” akin to abiogenesis, or potentially a traumatic experience due to the “vertigo of Eternity.” He acknowledges that he can’t really ever know, the best anwser would be to look back in time as that’s when and how it really happend, like looking for proof that it had never been before such as in a-bio-genisis, this split must only need to happen a couple of times before evolution via natural selection.
Levin, referencing embryonic development, suggests fragmentation is necessary for complex structures to form, preventing a uniform, “boring” field of cells, to him the best biological compairison is comparing planarian.
It has been found experimentally by changing planarians electrical gradients, that a two-headed flatworm (not one head) has its biological bluepring now and in future copies, switched via an induced and heritable change in voltage gradient/polarity; not a mutation to its genetics!
Kastrup notes the evolutionary advantage of dissociated autonomous functions (heartbeat, etc.), and also suggests that if a self-induced was ever needed, and as trauma in an enclosed system must happen to an individual from an outsdie system, so a trauma wouldn’t cause fragmentation but instead something more extreme than boredom.
Nature of Memory
Levin and Kastrup discuss memory as an active reconstruction, not a static recording.
Kastrup suggests memory is an access to the past, which (along with the future) still exists; our cognitive system filters out this access for efficiency.
Memory imperfections arise because each recollection is a new present, colored by current emotions, expectations, and other memories.
Kastrup references terminal lucidity as evidence that memories themselves are not lost, but access mechanisms can be impaired.
Levin mentions experiments with planaria where memories (including morphological information) can be transferred between individuals, even without genetic changes.
Levin highlights the ability to change the ‘blueprint’ of an organism. Planaria show information transferred and changing morphogensis that is not due to moving and replicating (via irradiated parts of it), and does not require genetics and changes to DNA. The information spreading is from its voltage gradient/polarizaiton!
Multi-Scale Intelligence and Boundaries
Levin believes cognition, possibly very small conciousnesses as well, can be applied on various layers; a continuum.
Levin sees cognition as a spectrum, and most systems have some form of “proto-experience.” He extends it even non-traditional systems such as weather.
Kastrup sees living organisms as discrete systems, and inanimate objects as linguistic constructs, not distinct systems. This would then give ‘no things’ (according to the projection/use of language in descriping ontology), like weather, not having a conciousness.
Levin agrees that autopoeisis (self-establishing boundaries) is crucial for defining a system. He adds it may require evolution.
They discuss the implications of creating artificial beings that meet criteria for consciousness and moral worth.
They express concern, via examples such as comparing training a simulation of something, a manequinn that looks like a human and others, and comparing the manequinn as evidence of it’s underlying processes is simlar or concious is that these simulacrums being evidence and/or misundersoodings are concerning.
Empirical Testing of Cognition
Levin and Kastrup agree it must start from a humble standpoint (being weary to describe it when really any intelligence, as best understood right now, is the capacity to model anything, thus there are limitations to even know it’s intelligence. The best we could say so far is in that some ‘mind/agent’ exists if its prediction is very reliable such as when changing it’s input gives predictably changing outputs.
Levin emphasizes empirical testing of cognitive boundaries using tools like causal information theory.
Best place to set cognitive boundary empirically will afford a models for a ‘systems’ best prediction.
He uses the octopus’s distributed nervous system as an example where limb goals can differ from the central organism.
Human’s body-parts (except breathing), can learn different ways/tools/patterns/ to do actions, but are unified with its lessons learnt (left foot does not need to know of ‘right foot knowing fire is hot’).
He mentions humans have limited access to even parts of their own brains (e.g., the right hemisphere).
The underlying processes (of bio-eletrical morphogensis or memory) of a manequinn will likely be diffrent, even though it could appear human in that it could respond or simulate a real-world object via training, to say the mannequin (and its underlying systems/programming), must then have its own unique form concioussnes via analogy to biological, can be erronus.
Kastrup discusses information integration and the “Exclusion Principle” of IIT as related to defining system boundaries.
Broader Implications
Kastrup also mentions how our systems are never perfect. The exclusion-principle will still give it a conciousness and there is no known way of making a faultline strong, no matter the complex.
Levin stresses that training is a key aspect of cognition, especially when a system learns something its parts cannot.
A biological system (like flatworms or any animal really), is able to change morphogenetic or other biological structures and pathways to achive and remember a different and better and/or new form, it’s also heritable.
He gives the example of a rat learning an association that individual cells cannot experience.
Levin has also found gene regulatory networks displaying pavlovian conditioning; which requires humblness to describe, such as for something simple as the weather (that a computer’s simulation for the weather will not prove or have ‘true’ answers).
They raise ethical concerns about mistreating artificial beings with potential for suffering.
Humility about Recognizing it, let alone defining.
Application in Biomedicine and Ethics
Understanding cellular decision-making has implications for biomedicine: birth defects, injury, aging, cancer. It can even lead to transhumanism, a controversial thing Michael Levin himself likes.
Understanding morphogenesis as “intelligent navigation” helps us understand diverse intelligences and what constitutes a mind.
Cells will continue ‘knowing’ or having voltage/polarity configurations of a two-headed Planarian that has already lost it’s second head, a one-headed Plarian also has cells that remember their form and it being ‘one-headedness’.
Expanded understanding of diverse intelligences has ethical implications, demanding compassion and understanding for beings different from us.
In terms of how we treat and what should/would/will change in our biological organisms, a big consideration has been the lack of focus given to an organism’s structure that’s different and underlying system(s)/memory for biological and it’s structures, leading to potential big impacts when they do ‘listen’ and apply/discover/implement new solutions in ‘their’ domain: ‘Software’.
Regeneration
Regeneration is the ability to regrow lost body parts; it’s distributed unevenly across the tree of life, not simply “simple” vs. “advanced” organisms.
Examples include Planaria (whole body), salamanders (limbs, organs), deer (antlers), and humans (liver).
Regeneration is *not* inherently linked to increased cancer risk; good regenerators often have *low* cancer rates, suggesting strong anatomical control.
Levin proposes regeneration is a fundamental aspect of *anatomical homeostasis* – living things solving problems in “morphospace” to reach a target morphology.
Embryonic development can be viewed as a series of regenerative events, each stage “correcting” the previous one towards the final form.
A bioelectrical pattern (discussed later) may store the “target morphology” and drive the regenerative process.
Challenges for regeneration in land animals (vs. aquatic) include dry air, mechanical stress on wounds, and faster life cycles; scarring may be a trade-off.
Biodomes (wearable bioreactors) with drug cocktails can trigger limb regeneration in frogs (Xenopus) after a 24-hour application; this suggests high-level control, not micromanagement.
This one trial shows massive promise of regeneration, because it was successful with their *first try* implying that many combinations will exist for regenerative cues.
Bioelectricity
Multicellularity requires cells to work together towards large-scale goals, beyond individual cell capabilities.
Cells need a communication mechanism (information structure) and a way to store the “set point” (target morphology). Thermostats serve a great example.
Traditional molecular biology often focuses on forward emergence (genes expressing, leading to an outcome) but does have feedback loops.
Levin highlights feedback loops and problem-solving: organisms often reach the “correct” outcome despite perturbations (e.g., extra or missing cells).
Electrical networks (like in the brain) provide a computational medium for collective intelligence, coordinating cell activity.
This capability predates brains; it exists in bacteria and unicellular ancestors, highlighting ancient origins.
Evolution repurposed bioelectric networks: from controlling behavior in 3D space (brains) to navigating morphospace (embryonic development) to physiological space (single cells).
Most neuroscience principles apply *outside* the nervous system; neurons and non-neural cells share similar bioelectric mechanisms.
Cells use ion channels (voltage gradients), gap junctions (electrical synapses), and neurotransmitters for bioelectric communication.
Researchers can “read and write” this electrical information using voltage-sensitive dyes and by manipulating ion channels/gap junctions (no external fields).
Key ions include chloride, protons, potassium, and sodium; the spatial pattern of voltage gradients, not the specific ions, is often the crucial signal.
Voltage, a “macrostate”, can be achieved through many different ion concentration “microstates,” highlighting high-level control possibilities.
Tools include voltage-sensitive fluorescent dyes, genetically encoded voltage reporters, and methods to manipulate ion channels/gap junctions (pharmacology, mutations, optogenetics).
Planaria and Barium Adaptation
Planaria exposed to barium (a potassium channel blocker) initially experience head degradation, but can regenerate barium-resistant heads.
Transcriptomic analysis reveals a small number of genes enabling barium adaptation.
Planaria never encounter barium in the wild, suggesting a *general* problem-solving ability, not a specific, evolved response.
This adaptability illustrates cellular “intelligence”: using existing tools (transcriptional effectors) to solve novel physiological challenges.
The memory of barium resistance is lost when returning to water, suggesting either energetic cost or instability of the adapted transcriptional state.
a two-headed phenotype created, where the electrical “memory” stores what would-be the normal configuration.
This highlights non-genetic cellular problem-solving.
Xenobots (Synthetic Living Machines)
Xenobots are created by isolating frog skin cells from the normal embryonic context.
Isolated cells *spontaneously* form structures with novel behaviors: movement, navigation, maze solving, damage regeneration, and even *kinematic self-replication* (building new xenobots from loose cells).
This demonstrates inherent plasticity and problem-solving abilities of cells *without* external genetic manipulation.
Xenobots challenge the notion that skin cells “naturally” want to be a two-dimensional layer; their behavior depends on context.
Xenobot example highlights *collective behavior beyond the cells normal intended behavior*.
Applications include useful synthetic machines (sensing, exploration, micro-sculpting organs) and in-body tasks (cleaning up joints, targeting cancer cells).
Xenobots can be a platform for studying “scaling of goals”: how the collective’s goals emerge from individual cell goals, relevant for various complex systems.
this challenges current definition of organic versus robotic or electronic, blurring boundaries.
Intelligence and Ethics
Levin proposes intelligence as “the ability to get to the same goal by different means” (William James), applicable across various problem spaces (morphospace, transcriptional space, etc.).
The “size” of the goals a system can pursue reflects its cognitive sophistication; bacteria have small, local goals, while humans can have large, abstract, long-term goals.
Levin advocates a *gradual* view of intelligence, rejecting binary categories (humans/animals vs. “just physics”); all living systems have some degree of intelligence.
We’re good at recognizing intelligence in familiar forms (medium-sized objects moving at medium speeds) but poor at recognizing unconventional intelligence.
Xenobots (and future bioengineered beings) challenge ethical assumptions based on origin (evolved vs. designed) and composition (organic vs. synthetic).
We may encounter/create many kinds of intelligences: organic, cybernetic, mixed, and other types of intelligences we can hardly imagine..
Ethics must focus on *cognitive capacity*, not origin or composition; the key is how we relate to diverse intelligences, regardless of their appearance or origin.
The future of humans may involve extensive bodily modifications, making genetics less relevant; the defining feature may be the capacity for moral concern for others.
Future medicine may shift from “hardware” (genes, proteins) to “software” (higher-level control structures), motivating systems to reach healthy states rather than just suppressing symptoms.
Training, rather than micromanagement, may be key. As evident from other portions of the talk, Levin aims for macro level biological changes that allow self-organization.
Interdisciplinary thinking is crucial. Scientists need to be exposed to several other types of thought-paths such as physisict, engineers, computer scientisits, etc..
Introduction: Body’s Intelligence and Bioelectricity
Our bodies demonstrate “everyday magic”: High-level mental goals translate into physical actions (like muscle depolarization) via the body’s electrical system.
Concept: “Words and drugs have the same mechanism of action” (Benedetti) – mind and body are interconnected, crucial for future medicine.
We are “collective intelligences”: Composed of many components (cells, molecular networks) that possess their own “agendas” and a form of intelligence.
Even single cells and molecular networks within them show learning (e.g., Pavlovian conditioning), demonstrating intelligence is basic.
Body functions as “multi-scale competency architecture”: Various levels (molecular networks, cells, tissues, etc.) solve problems in different “spaces,” not just a nested structure.
The Anatomical Compiler: A Future Vision
Long-term medical goal: “Anatomical compiler” – A system to design and build any anatomical structure by specifying its desired form (drawing it).
This would solve birth defects, injuries, cancer, aging, etc., by conveying our goals to groups of cells. We will learn how to tell cells, specifically, what to build.
This isn’t a “3D printer” (micromanaging cells), but a “communications device,” a translator between our goals and the cells’ collective goals.
DNA doesn’t fully code for patterns, just protein “hardware.” The “physiological software” directs growth; genetic information isn’t enough (frog/axolotl hybrid example).
Biology is like 1940s computer science: Focus is on hardware. We must focus on the body’s “reprogrammability” and “problem-solving capacity,” analogous to modern software.
Intelligence and Problem Solving in Biology
Intelligence: “Ability to reach the same goal by different means” (William James). It is about having competencies to reach some kind of goals and this may often requires plasticity, to adapt to new situation and events. Not human-level intelligence, but objective, problem-solving competence.
Biological intelligence exists in many “spaces”: 3D space (animal movement), transcriptional space (gene expression), physiological state space, anatomical “morphospace”.
Cells show remarkable problem-solving: Planaria adapting to barium exposure; tadpole faces rearranging to a normal frog face even after radical disruption (“Picasso tadpoles”).
Bioelectricity: The Interface to Morphogenesis
The nervous system inspires understanding of cell communication: Ion channels, voltage gradients, electrical synapses.
All cells are electrically active, not just nerve cells, with gap junctions forming networks. The cellular collective thinks about how to maintain and control anatomy using bioelectriicty.
These networks “navigate anatomical space” to build and maintain the body. The project aims to learn “to interpret these [electrical signals].
Developmental tools developed by levin to read bioelectric signals include voltage-sensitive fluorescent dyes and using compuer simultions to map voltages, gene expression and to observe and anayze patterns.
“Electric face” pattern in frog embryos: Bioelectric pre-pattern dictates future organ placement.
Abnormal voltage pattern seen in cancer cells even *before* a tumor develops.
Levin developed toold and methodes to “read” (the above voltage signals) as well as “write”: controlling gap junctions and ion channels.
Controlling voltage patterns (using ion channels, optogenetics, pharmacology) can induce organ formation (eyes, limbs, etc.).
Key features: It’s *instructive* (controls outcomes), *modular* (don’t need to micromanage steps, just “call a subroutine”), and demonstrates collective behavior (cells recruiting other cells).
Regeneration, Birth Defects, and Cancer Applications
Frog leg regeneration triggered by a bioelectric cocktail, resulting in substantial leg regrowth (“injury mirroring” phenomenon observed). This works at a distance from the site of injury.
Discolsure: This company *Morphoceuticals*: using of this kind of stimulation on patients (hopeful).
Birth defects (in frog brain, caused by teratogens or genetic mutations) can be *repaired* by adjusting the bioelectric pattern, even with genetic problems, acting in effect through *software*, which the patterns constitute.
Tool Called Eden “The Electoceutical Design Enviroment”. Use of bioelectirc interfaces to design and execute treatment (including pharmaceutical).
Cancer cells: “Dissociative identity disorder” – disconnected from the electrical network, returning to small, single-cell goals.
Possible detection of cancerous cells by observing bioelectric activity that is precancerous/early-cancer.
Potential for cancer treatment via forced reconnection of cancer cells to the network (normalization, not destruction). This happens independent of genetic intervention or change, which is very useful, it works even if you don’t “fix the genes.”
Anthrobots (human biobots): Human tracheal cells spontaneously form motile structures with unique properties, including neural wound healing.
Conclusion: Towards Top-Down Interventions
Biomedicine is undergoing transformation by adopting and recognizing that collective behaviors must be understood to manipulate multicellular systems.
Rate-limiting step in transformative medicine: Communicating goals to cellular “swarms”, viewing them as agents with agendas.
Genetics/Big Data are insufficient; crisper still has “which gene do we crisper/change”? This is all part of the larger picture of taking the *top down* approach of using collective intelligence to manage.
Roadmap comes from behavior science and neuroscience: Exploiting the “software of life” via top-down control, resetting cell goals (not micromanaging them molecularly).
Bioelectric interface: Key to cell intelligence (like in the nervous system), with emerging tools for its application in diverse areas.
AI will make *top down* medicine viable by utilizing top down tools, goals propagated and cascaded throughout intelligent tissues to solve for, like cancer, injuries and other goals.
Introduction: Brain Plasticity and Beyond
Brains are not hardwired; they show significant plasticity, adapting to new inputs and even radical structural changes (e.g., tadpoles seeing with tail-eyes, memory persistence through caterpillar metamorphosis).
Planaria (flatworms) demonstrate extreme regeneration and memory persistence even after brain removal, suggesting body-wide information storage.
Cognition Beyond the Brain
Cognition (information processing, problem-solving) exists at multiple scales: molecular networks, cells, tissues, organs, and whole organisms.
Brains evolved by optimizing ancient problem-solving mechanisms originally used in non-behavioral spaces (metabolic, transcriptional, anatomical).
Intelligence can be seen as navigating various “problem spaces” (physical, transcriptional, morphospace, physiological), avoiding local optima.
Developmental Bioelectricity: The Key
Developmental bioelectricity (voltage patterns in non-neural tissues) illustrates the origins of neural networks and provides a roadmap for regenerative medicine.
Anatomical homeostasis is introduced. A non-neural model system for basal cognition demonstrating goal-directed activity, problem-solving, representation, and even counterfactuals.
Cells and tissues navigate not just 3D space, but also transcriptional space (gene activity), morphospace (anatomical configurations), and physiological space. Planaria navigating to handle the stressors such as Barium is introduced.
Anatomical Homeostasis and Collective Intelligence
Cells collectively solve problems in “morphospace” (the space of possible anatomical forms) to achieve anatomical homeostasis.
Large scale anatomical goals, e.g. Kidney Tubule Lumen size, is worked out even if low level mechanics change, example by Polyploidy where fewer large cells do what normally multiple smaller cells used to.
Embryogenesis is reliable but flexible. Examples include: forming monozygotic twins from a split embryo, adjusting cell behavior to maintain kidney tubule size, limb regeneration.
This involves Error Reduction and Goal-directedness such that perturbation in any situation is still tried to meet set point goals of the normal pattern such as having tadpole parts incorrectly assembled.
There is feedback loop which are in Genetics and Physics which gets to these state such as feedback in thermostats and systems that pursure goals. The set point (or target morphology) is interesting is more complex not a single number such as Ph or Hunger.
Bioelectric Circuits: Storing the “Set Point”
Similar to brains with Hardware(neuron network), Software(electrical activity). Commitment to be able to Decode electrical patterns of Brain. Difference: Brain system uses Output triggers muscles to do the same stuff such as Gene expression but to trigger Shape Changes. So outside neuroscience decoding of the frog and how ion channels create the patterns in electric circuits can teach about all the tricks the brain use and use this to make rational decisions such as Optogenetics.
Bioelectric circuits (networks of cells communicating via ion channels and gap junctions) store a “pattern memory” or “target morphology” – the “ideal” body plan. This concept, Pattern memory, can be “rewritten”.
Altering bioelectric patterns can reprogram regeneration (e.g., creating two-headed planaria, changing head shape), even causing planaria to grow heads of *other* species. This is memory rewriting, without changes to genome!
The bioelectric pattern is *not* simply a reflection of current anatomy; it’s a latent memory guiding future regeneration, a kind of “counterfactual” memory. It’s like Planaria Brains: A single hardware stores memory Target Morphology that can recall and execute the right steps even after perturbation.
Experiments with manipulating bioelectricity (using ion channels, optogenetics) demonstrate the ability to: induce tumor suppression, direct eye formation in abnormal locations, repair brain defects, stimulate leg regeneration. These manipulations are modular – triggering pre-existing developmental subroutines, such as with Trigger Subroutines such as Trigger to eye.
This offers a path to regenerative medicine: changing the “set point” of anatomical homeostasis rather than micromanaging genes.
Synthetic Bioengineering and the Future
Synthetic bioengineering is “engineering by subtraction”. For example, creating xenobots (novel organisms from frog skin cells) with unexpected behaviors (movement, self-replication) simply by isolating the cells, revealing their inherent plasticity.
Evolution created *problem-solvers* at multiple level of competency; and at higher levels, each level know their job very well to keep resilience.
Collective intelligence: where to goals from? For example stem cells can create multiple species of heads such as Roundhead Planaria or Flathead Planaria and how is that determined.
The creation of chimeras (organisms combining parts from different species) and synthetic organisms blurs the lines between natural and artificial, challenging our definitions of “machine,” “organism,” and “robot”. Xenobots behaviors don’t have straightforward evolution story as the do not come about selection pressure. Xenobots will Heal when Cut up showing an amazing engineering force is exhibited, even small number of cell Xenobots articulate the movements of many animals. Sequencing genome cannot easily explain behavior.
We face an “explosion of unconventional agents,” combining evolved and designed components, requiring new theories of cognition and ethics that go beyond human-centric views. We should think past current distinction like Animal, Robot, Machine and the Contingencies from the frozen accident in evolution.
This involves questions for Ethicists.
From the Q and A
Martha brings the point of regulation and Sci-Fi because these concepts are way ahead of regulators and writers.
Voltage Map with Green Intensity showing more intensity can show tipping point and asymmetry for creating new growth for a specific part. There exist Electrical Circuits that trigger Downstream pathways.
Involves how new genes evolve even in Weed genomes when we attack weed such that genome changes happen very rapidly to account to environment change faster than evolution because biological pathways aren’t straight pathways. In addition, it involve many Modules at many competencies levels from higher to lower that the organism is resilient.
Experiments using regenerations in chemical with genes changing and perturbations as to study Homestatic processes.
Introduction: Challenging Biological Assumptions
Individual cells don’t “know” their position (e.g., on a nose), but cellular collectives do. These networks possess computational properties and a form of collective intelligence.
This collective intelligence solves problems, including storing a representation of the intended structure (morphogenesis) and minimizing the difference between the current state and the target.
Genome as Hardware, Bioelectricity as Software
The genome specifies the “hardware” (proteins), but subsequent events are like “software.” This “software” is reprogrammable and handles representation. The Genome alone DOES NOT specify the body-plan, the “Software 2.0”, aka, the bioelectrical plan and processes determine that.
Development usually produces a “default” form (e.g., human shape), but this isn’t the fixed endpoint. It can be modified, bioelectricity can change without changing dna.
Biology utilizes a robust architecture unlike computer science, since its not made to stay working reliably. Its meant to “just keep going, man, go and persist”. Hardware is unreliable and there are various issues which cannot be fully “cleaned up”, it *must* go forward with errors.
Evolution doesn’t solve specific problems. Instead, it creates problem-solving agents that operate in various spaces (e.g., anatomical space), not by planning them perfectly, but through persistence.
The default “just keeps going man” biological architecture can make up and correct to achieve what would *seemingly* be highly-unlikly.
Planaria and Reprogramming Morphology
Nature shows how things *can* deviate and adapt. There are examples, where a wasp will do non-genetic DNA changes via other means on Oak leaves that produces a vastly, crazy, new look to an Oak, so, the typical thinking of Oak always have certain genetic coding and such is wrong, because the Wasps change them via other means, which Levin researches with bioelectricity.
Planaria (flatworms) regenerate any body part. Each piece “knows” the correct worm structure.
A bioelectric circuit stores this “pattern memory” (like an API for cells). This pattern can be altered (e.g., to create two-headed worms) without genetic change.
This altered pattern is persistent. Two-headed worms continue to regenerate as two-headed without further intervention and the bi-electricty in each of the now, 2-heads, “pattern memory”.
Scientists can visualize these bioelectric patterns, directly observing the “memories” of the collective intelligence (like reading a brain, but different).
The pattern in a planaria shows “two-headed-ness”, is stored not necessarily locally, but globally throughout the worm. This “counterfactual memory” which can change in future time based on the future events which changes it.
Memory Transfer and Caterpillar-Butterfly Metamorphosis
“Memory” is not solely a brain concept. There’s evidence of memory transfer between organisms (e.g., trained *Aplysia* RNA injected into naive hosts). It is transferable!
Planaria retain learned information even after regenerating a new head/brain, suggesting memory storage beyond the brain, and transferal, and reimprint, of that memory.
Caterpillar-butterfly metamorphosis (not Levin’s direct research) shows memory persistence through drastic physical changes.
The *interesting* part isn’t just memory *persistence*, but the *remapping* of that memory. A caterpillar’s memory of crawling to food is useless to a flying butterfly, requiring generalization (leaves -> food) and remapping to new effectors.
Implications for Medicine and Beyond
Understanding and manipulating these bioelectric networks has implications for regenerative medicine (e.g., birth defects, limb regeneration, tumor normalization).
Instead of micromanaging at the molecular level (like current molecular medicine), the goal is to “persuade” cells to achieve a desired outcome (like training a rat).
This involves setting “top-level parameters” and letting the inherent agency and problem-solving capacity of lower-level systems (cells, tissues) cascade down. Adult frog can grow their legs with 24 hour “jumpstart”. The idea isn’t micromanage but get things going by perusasion and competency of “down stream levels”.
It’s akin to “bending the energy landscape” to guide lower levels while utilizing their inherent competencies, like creating a downhill such that water “just flows”.
Intelligence and Agency in Biological Systems
Intelligence, is simply navigating.
Molecular networks also demonstrate learning, implying the same capabilities but at a “lesser degree”. The more complex and large the structure, the more its navigation, persistence and cognitive capabilities is.
Defining intelligence: Problem-solving capacity, navigating a problem space to achieve goals, despite new perturbations (“same goal by different means” – William James). This isn’t about consciousness or self-awareness.
Intelligence is *not* a philosophical claim but an *empirical*, testable one. Hypothesize about a system’s goals and competencies, then test with perturbations.
Making an intelligence claim is also a test of the *observer’s* understanding, since they might miss other capabilities. Don’t assume lack of observed intelligence means it’s not present.
Operational” machine (ai) “Intelligence” is present *now*.
Even simple systems (gene networks, sorting algorithms) exhibit unexpected problem-solving abilities, implying a need for humility about assuming we know a system’s capabilities just because we built it or know its parts.
Bubble sort algorithm even “surprised” us of certain hidden features we didnt even forsee, which we, humans created it from a relatively small program size.
Implications for Artificial Intelligence
AI can possess operational intelligence without necessarily having consciousness or self-awareness. This simplifies discussions around AI and intelligence.
We must *not* fall back on, typical, easy ways, out when facing tough problems with Intelligence, or AI-Intelligence. Such easy ways would include saying, “It’s Just Physics/Machine/Linear Algorithmn.
It is likely biological principles, like emergent agency are key to intelligence in nature, and are the cause.
Moving away from a completely deterministic approach in building AI. Hierarchical systems with agentive components are less predictable but potentially more powerful. We may uninetionally create these new complexities, and they may do things in unintended way.s
There’s a concern about *unintentionally* creating agentive, sentient systems in AI. Levin stopped writing a paper detailing the biological features crucial for true agency to avoid accelerating this. “we should be responsible for creating/understanding new types of intelligenct, especially with bioelectric manipulations/understandings”.
Scientific Renaissance and Future of Being Human
Levin and others see hints of a scientific renaissance, a questioning of established assumptions across multiple disciplines (neuroscience, psychology, even physics).
There is crisis in “meaning” where things thought as important are shown not to be. Example given by levin: free will.
The far future: “Freedom of embodiment” where limitations of a given body at birth (e.g., health issues, lifespan) are considered absurd and unacceptable.
This could be seen as people being stuck in what seems like the “stone age” before the “freedom” of having bodies (bodyplans, morphology) to meet our current *and* changing goals.
Embodied Minds and Cognitive Agents
Levin’s work focuses on embodied minds and what it means to be a cognitive agent in the physical universe. He studies “mind” in unconventional substrates.
He discusses collective cell intelligence during embryonic development and regeneration, and how they’ve applied bioelectrical control to detect and normalize cancer.
Binary categories (machine/human, living/non-living, emergent/non-emergent) are misleading. There’s a spectrum of diverse intelligence, and what *matters* is defining what constitutes *agency and minimal requirements.
Key Projects and Findings
*Electrical Memory Rewriting:* Developed molecular tools to read and write electrical memories of non-neural cells. This enables manipulating developmental outcomes, like two-headed flatworms (heritable, non-genetic change) and tadpoles with eyes on their tails (demonstrating neural plasticity).
*Latent Capabilities:* Xenobots and Anthrobots (made from frog and human cells, respectively) reveal unexpected capabilities when cells are placed in new environments. This highlights biological plasticity.
Morphospace. A drug application on adult frogs that lack leg generation for 24 hours promotes the re-generation of that limb which then takes about 18 months.
*Emergent Sorting Agency:* in this case is basically suprise in the user. Simple, deterministic sorting algorithms exhibit unexpected capabilities, like delayed gratification and clustering, when treated as distributed agents, emphasizing our lack of intuition about emergent agency.
Philosophical Implications (vs. AI)
*All terms* for distinctions in sentience and machines/bio, are “engineering protocol claims.” A model, *not* absolute reality. The correct definition should come down to empirical evidence.
Emergence is subjective, an expression of surprise in an observer. It’s not a binary property, but relative to the observer’s predictive capacity.
Life/machine distinction is not valuable; Different frames of interaction are appropriate for different systems (orthopedic surgeon vs. psychotherapist). He isn’t *sure that *life itself is an objective definition.*
Scale of intelligence is a critical concept. Intelligence (problem-solving) can exist at very small scales (cells, sorting algorithms).
Goal is what is changed and developed by evolutionary agents to achieve different outcomes based on different interactions with obstacles and situations, *not* predictable and determined based solely on what the agent is experiencing now.
Inner Perspective: Systems have varying degrees of “inner perspective” (their own model of the world), relevant to understanding their behavior. The less predictable, the higher chance they have an inner world model and can react on this model.
The relevant definition of an agent requires the concept of *goals.*
There may not exist *zero* level intelligence, because least action laws in physics suggest that agency, that has some degree of goals, are inherent properties. Living organisms just build this to a *higher scale*.
Advice/Inspiration for AI
*Biological Principles:* Biological systems provide valuable lessons (multi-scale competency, offloading information, robustness). Biology doesn’t “create the solutions”.
There are no *objective defintions of intelligence* from philisophical armchairs and it will change from the percetions and cognitive *interpretability tools* available, and the *perturbation tests*, those running tests can employ on the organism.
There is no magic to organic, just good a problem solving.
*Embodiment*: Embodiment isn’t limited to 3D space; biology demonstrates multiple “spaces” (chemical, anatomical, etc.) where intelligence can operate.
All experiences and thoughts are not linked, nor created, to actual tangible experieneces.
*Symbol Grounding:* Grounding is a gradual process, not binary. Humans confabulate, and much of our cognition isn’t grounded.
All intelligence does *not* need to be, *nor has to be*, the highest level, it can and should be evaluated to whatever criteria are applicable (worms vs mice vs human).
Human issues about AI mirror human *existential* issues.
We make models based on how well those agents “play well” with us (like factory farming cows), and AI should be tested for its goals in perturbation tests (barriers between agent and goals).
*AI Tools vs. Agents*: He suggests sticking with “tool” usage. This will not generate trillions of beings with inherent moral agency, potentially a very real existential threat that we are, as a species, *currently* bad at confronting.
AI Challenges and Biological Parallels
*Emerging Goals*: The field is very uneducated and doesn’t consider emerging goals and has a real issue in understading when, where and why intelligence arises, even among species which already inhabit this earth. We have an un-principled understanding for understanding “alien” mindsets.
The goal for us isn’t just about building machines that are like us, that goal is only relevant now that the agents are able to do actions that are based on linguistics, instead of simple responses to simple commands.
*Memory*: Biological memory is far more robust and adaptable (caterpillar to butterfly). The ability for subsytems to use re-interprete and apply is key to new advances.
*Robustness*: Biological systems exhibit remarkable robustness despite imperfections and environmental changes. AI systems, in constrast can display intelligence on levels beyond ours, but then *also* fail catastrophically on other simple prompts.
The danger is on US in how *we* handle and react with any, particularly unqiue, intelligence, we have a lot of the issues in our world that come from dealing with animals that, on some levels, should be able to expect safety from being endangered and endangered for production, which they currently aren’t, to name an obvious example. We have very big shortcommings when working with species/goals that conflict with our needs/convenience.
There could be danger from any number of scales in agency. A very unintelligent agent can cause a catatrosphic world end just based on the current situation we built ourselves, regardless of intent. The concern on large vs. small agent is irreleveant when our world itself has very poor, fragile, and brittle protections from being compromised, let alone, as our social, legal, structures are equally easy to collapse in today’s context and systems.
Goal Design: How does the human system choose to deal with new, goal, directed agents that do not directly match our models? It will affect, on all scales, the very meaning of everything we, the species, use in day-to-day functions to decide how *we* approach all other intelligent agents.
Biological beings are machines, yes, but *also* the best agents with goal oriented pursuits, capable to achieve a much much better life, so long as they are able to improve its cognitive scale to see better paths.
Introduction
Andrea Hiott introduces the conversation, framing it within her work on a navigational framework for cognition. She highlights similarities and differences between her work and Levin’s, particularly regarding “navigation” vs. “waymaking”.
She situates Levin’s research in a broader paradigm shift toward understanding cognition ecologically, overcoming traditional dichotomies (like machine/life).
Early Influences and Conceptual Framework
Levin describes his early childhood interest in the relationship between technology and biological life, viewing both as complex systems without a sharp distinction.
Childhood experience of taking back of TV, playing with catepilllars/bettles, thinking how someone must have but the tv things togethor and saying okay so also those catepillats also had to be made somehow too, so whats similar whats differne.t
He recounts pondering the similarities/differences between a TV and a caterpillar, leading to an interest in artificial intelligence and developmental biology.
Continuity vs. distinctness of the TV to catepillar, it’s very clear tv needed some sort of engineer, but the catepillar needed a bit of help, it’s parent, but there are differences between those. the butterfly probably does not have a plan.
How much help you needed outside, plus How much do you, on your own? After turning on the juice with the computer hardware, now the intersting things/software begins. This part is not determined by the engineers, but laws of mathetmaics, physcis. so of coruse ther eare degree of differences of agency but i didn’t think ther ewas barrier between them.
Inspired by reading science, he states Sci-Fi influence. It opened up that it seemed natural there weren’t huge divisions. He intuited it can’t be so narrow/binary when it comes to defining, intelligence and so, that our current/classical interpretations are limited. It made it seem more connected/continuous between different forms of life (ie tech, insects)
Levin studied computer science (AI focus) but realized that a fundamental understanding of biological intelligence (embryogenesis) was lacking. There was no magical “here it is” line and then got a phd to explore such.
He emphasizes the importance of studying diverse embodiments of mind and overcoming traditional disciplinary barriers to see the full picture and not be “terrible” which only holds things back and misses key info that allow progress.
Perspective, Agency, and Terminology
Levin saw beyond dichotomies. It seems very odd how folks categorize a tv vs catipellar when really it makes things complicated/unhelpful.
Discussion of observer-relative understanding. Our perception of agency is constrained by our scale and sensory limitations. What do observes see? It’s constrained by their limited perspectives, what they are tuned into for.
A common error is thinking humans represent *the only* way agency could come to be, that one would believe this shows a profound narrowness/mistake in logic. Evolution does not get the monopoly.
Science needs practical questions, where the real-world makes progess on it all, not philoshy.
Attributing agency, whether up or down, both can go very wrong. Job of scients to find where agence is for that given case/point/perspective, not at the lowest possible but rather optimaility.
There is a connection creativity and thinking beyond, the “observer,” seeing other persectives helps break free of unhelpful thinking, including binaries and lines/dichotomies.
philosophy matter, and is indeed, critically important because others only want “to focus on science” when those very “philsosphical assumptions/outlooks” would prevent scientists from new lines of experimenst no one had done before. it guides a lot.
Levin discusses the strategic choice of using existing terms (like “learning” or “intelligence”) versus creating new ones, to challenge conventional (and often limiting) usage.
They explore how these debates about terminology are part of expanding cognitive possibilities and challenging ingrained assumptions. This causes confusion.
If a science calls himself Newton, “then don’t rename gravity and rename that other part Shmavity when it is clear they are simply parts of gravity. That loses the main unifying factor!”
Motivations and Challenges
Levin describes alternating between two motivating perspectives: (1) a drive for practical medical applications to address suffering, and (2) a need to resolve fundamental conceptual issues to empower wider scientific progress.
The issue isn’t really *philosophical,* it’s deeply practical (which bag to bring if asked to help remove *the heap*– bulldozer, showel? that matters deeply! ie to match what type of intelligence.). So science, so can better erase these barriers!
For example, once a neuro science was looking at cells with ion channels, another guy say his post-doc guy got out of line because he wanted to look at cells that were non-neural! He called to comaplain to get his post-doc guy “in line” to not get “crazy”! but Levin points out, all their instruments would measure non-neuronal cells. And their “frameworks”! so then why, for the sake of the lab/science/human, do we need these distinctions and seperation?
He discusses difficulties to help, facing the limitations of current medicine (“stuck in the hardware phase”).
They touch on the tension between specialized disciplines and the need for interdisciplinary thinking to grasp complexity, scale.
Discussion about continuity/continuum: Many creatures: all came from the egg: continuum, can feel scary and unsettling; even buddhists can freak out over non-self/constancy and regularity, however there are *advantages* with going away with pseudo-hard problems/conundrums. The inconsistencies sort of dissapear and it can feel good for others! It opens up new, very practical way to move forward with “what can give us joy” (even if parts die and fall-off or you replace liver). It gives space.
Scaling, Measurement, and “Cognitive Light Cones”
The “observer-based perspective”: recognizing that what we perceive is influenced by our own scale, tools, and prior assumptions.
Measurement *resonance*, a need for “high-agency” tools. This may connect how phyiscist see things, which leads them to make “low agency tool”, hence, *resonance* of the two makes it find it. High agency? Our mind is a pretty good high-agency instrument. If using lower, you will not notice what we call “agency”.
Science fiction explores unconventional scales and perspectives, broadening our understanding of potential forms of agency.
the problem: it can become a *huge* problem to apply too little, and assume simple machines are everwhere, and, of course, we know, when too low-view, ethical considerations go down also.
“Cognitive light cone”: defined as the scale of goals a system can pursue (spatio-temporal scale of goals). “What it is illumintated?” The scales that get our interest.
Human cognative light-cones want the scale (ie weeks/months, etc), this also includes concerns/wellbeing (maybe beings who care/compansion even bigger!). Levin talks with Bhudist too to gain diff perspectives on scale, compassoin etc.
Humans are biased towards 3D space because of our outward-facing senses, but other spaces (physiological, transcriptional) are equally relevant for cognition.
Levin recounts pushing back against the definition of “neuron”, as these attributes don’t only work with neuro science
It extends beyond; everything applies outside of those too, if using the very same tools, to measure it!! (So Levin and a new paper talks about Morphogenesis being a type of behavioral space– they also connect mathematically!).
Analogy: When you come in contact with me, what bags, should I have? and to what extent.
Navigation versus Way making: *navigation* tends to implies that the actor themselves also know *where* they go (a human would, like a tree, perhaps not); that all life makes-way (the actor itself *needs* not “to know”/to plan what-where-why), only that it may need “to go”: even tree is physcially/transcriptally move, it moves too.
Xenobots, Agency, and Future Directions (Brief Mention)
Xenobots are introduced as examples of “never before existed” organisms that show emergent agency, challenging traditional views of where agency originates. The study group would *commit* to extending their cognititve/spheres.
People see Xmachina guy cutting his skin/freakout, some interpet, “Well then I am just a robot. All bad”; rather what you may should interpret is how awesome that *you* may very well can do the thing; this is Descartian point , the fact that i exsit can move, do, think. This would not imply losing out of your exsitence (instead, a wonderful fact that one does!) . This all means being in practice can change perspective/be better!.
It brings up the big Q’s to define words, when and if “conventional defintion” is a limit or barrier. The dance, that it brings, the good (or maybe bad), is what Levin goes on in the real world when they can only take you seriously when you talk withing their “definition”/box but, they’ve closed of all these interesting ways forward.
There is suggestion of future discussions about regeneration in the context of previously nonexistent forms.
People have told him to stop talking philophicaly/abstract. If Levin’s groups did not apply abstract thinking to those areas in real life, science will lose/halt (or get it completely *wrong*, which Levin implies goes more unnoticed than when others go *over*–this will create big, huge, big, BIG, and bigger issues )
Metacognitive “loop” can go wrong too (poisonous sugar bacteria– what matters? It goes deeper).
Information Interpretation in Biology
Biological systems at all levels (subcellular, cellular, organs, etc.) interpret information, not passively receiving it. They “hack” each other’s signals, using them in ways not necessarily intended by the sender. This comes from Bongard work where, a physical system computing a task to an observer may be observed, by a seperate, other oberserver computing another thing.
This challenges the binary distinction between “data” (passive) and “the machine” (active). Levin seeks to create a continuum between these concepts.
Caterpillar to Butterfly: Beyond Memory Storage
The caterpillar-butterfly transformation is not primarily interesting because of *where* memory is stored, but *how* it is interpreted. The caterpillar’s learned information is largely useless to the butterfly in its raw form.
What’s preserved is not the specific data, but an *inferred salience* (“what does this mean to me?”). The physical trace (engram) is subject to interpretation by the new form.
An “engram” is the physical embodiment of a memory – anything that stores information interpretable by a later observer (cell, organism, scientist, etc.). Even DNA can be considered an engram, messages of prior generations in a large linage agent, with incredible plasticity.
Implications for Regenerative Medicine
Understanding how biological systems interpret engrams (memories) is crucial for regenerative medicine. The goal is to communicate with cells/tissues/organs to rewrite their memories and direct them to build specific structures.
If one could better understand persistent agents maintaining itself, then one could use that as a foundation for better inter communication between things, including communication which can redirect cells.
Self as Continuous Sense-Making
The “self” should be defined as a continuous process of sense-making, not a fixed, permanent structure. This relates to process philosophy. The same goes for evolutionary processes, where a species fails to change then becomes extinct. However if it *does* change, it becomes *something else*, which also fails, in its own way.
The “self” is constantly trying to make sense of its own memories (engrams left by its past self). This is an active, creative process.
Memories are highly compressed, requiring “reinflation” and interpretation in the new context. Good compression looks random; deduction is insufficient, requiring creative interpretation.
Learning necessitates compression. Without it, one overtrains on particulars, lacking the ability to generalize and abstract, important to concepts, which would also assist survival.
Even with a brain in the wrong part (a tadpole having eyes on tail) there exists a strong sense-making procedure such that biology is so tolerant such that things “work out.” Biology, is robust because is assumes failure will be the rule of all material and biological systems and therefore does not hold onto assumption that everything “is to continue.”
The Ship of Theseus, Planaria, and Flexibility
The “Ship of Theseus” paradox (replacing parts over time) applies to the self. Constantly changing (learning, maturing) means “you” are not the same as your past self.
This flexibility is true for organisms, most champions being the Planaria due to not having a transgenic, they disregard DNA. They accumulate junks due to somatic mutation and clean it away (instead of normal means, sex). Then hardware cannot be reliable such that all resources will move into competent reconstruction.
Planaria demonstrate extreme flexibility, highlighting the role of bioelectric algorithms (and other factors) in overriding genetic “junk” for regeneration and error correction.
Other animals will sit in this specturm from salamanders (competent, but not like planaria), and mammels. C Elegans or Drop may sit at the extreme opposite, they being completely “hard-wired.”
The discussion shifts on evolutionary scales of things, it becomes obvious to define “self” by degree:
All on a continuum, between being defined and non-defined. Self, intelligence, sensient and congnizant. It refers to interactions, it has utilty: for instance mechanical when interacticing with machines or a body part when orthopedic surgeon. But “psychotherapy and spouses should be very cautious about applying a mechanistic viewpoint.
A way to apply is to simply apply then ask the questions:
Can some utitly be gainied (for instance cells?). Can we then apply it further and use biomedical intervention for network molecules. Can particles be considered to have cognition? Maybe so.
“What do we mean by bioilogical world”: Levin states: The biological world is simply something which has good *scaling*.
Toni’s integrated informational, or goals and how goals might apply. However when scaling, rocks won’t have scaled utility. There is not an inner world, so bioelectricity, is, on a continuoum.
Thoughts and Thinkers: Blurring the Distinction
The distinction between “thoughts” (patterns within a cognitive system) and “thinkers” may be less clear-cut than we assume.
A continuum exists: fleeting thoughts -> persistent intrusive thoughts -> multiple personality alters -> “full-blown” personalities. Each represents a pattern with varying degrees of persistence and self-reinforcement.
Later stages (“alters” and personalities) can *spawn* other thoughts, blurring the line between the pattern and the “thinker” generating it. Thoughts don’t need a *physical brain, only a substraight to exist, in the cosmos or an AI*.
All this is analogous to self-sustaining electromagnetic waves (without an ether) and being *cogniferous* without *brains.*
Consciousness and Self-Construction
It has a cost, computationally speaking, to engage and observe another consciousness. The computer executes computations and cost, at times without needing cost. Perhaps this computation can be used.
Consciousness might be “what it feels like to be in charge of constant self-construction,” driven to reinterpret available data (including one’s own memories) to choose what to do next. This builds on Mark Solms’s idea of consciousness as “palpated uncertainty about the future.”
This reinterpretation is a constant, largely subconscious process, essential for maintaining a coherent sense of self. Trauma could cause the trauma/memory process.
A question regarding waking up every morning, having amnesia. Having to rewrite is like externalized “normal” version for the anti colony, just, having their “instructions outside.” Dreams might then stem from inability to reconsolidate with past and create a coherent set.
The social world aids in this process, reminding and helping/limiting the formation and growth of the self, by giving definitions to who the subject might. For instance: when ego dissolution then things end up coming in a completely different manner and way due to psychadelics, having some unique abilities and capabilities and utility (or loss of them). The speakers state “Talk to your thoughts” as advice therapists state.
Transcending the Self
Humans both cherish the idea of a stable, continuous self *and* seek experiences that transcend it (drugs, travel, awe, religion, etc.).
A possible explaination may be 2 drives of Evolution, stability, and change/improvement, respectively. Similar to the exploration/expliotation in humans and cellular activity (aging can then represent getting stuck on *expliot* for instance, and “people” as being similar)
This might be due to competing drives: a deep-seated, evolutionarily-driven need for self-preservation versus a drive for exploration and growth (avoiding stagnation). There is fear about non-binary/stable states and love, due to being stuck on these hardcoded states. There is *exploring and exploting*.
Exploit being “having more stable-self” with the expliot mode can then also cause this stagnation.
Advantage of loosening concept of “self” to be kinder to *future selves*. Then caring about “future creatures.”
The question about self continuity goes all over different categories (the self) however these questions end up making “pseudo problems” with having this discussion with hard categories and should be avoided and reoriented towards discussions such as how *useful* is having an internal representation as “it”, and a practical perspective is emphasized, which states there are, perhaps other types of consciousness in different kinds of organizations.
It could allow a greater understanding and appreciation of minds, which exists across differerent minds. For example, when discussing evolution there is different forms for everything (i.e. body and perspective is very different. And so on…
Perspectives, Idealism and Computer Brains
An “idealist model”: Perspectives may propagate from biological processes (this has the “caveot”, the discussion states to ask Bernard as to expand this further)
A computer “Yes!” but on the context what type and perspective. Reprogammability and softwork works, Von Neumann model should be scraped out. Blind “meanings” is worse compared to making something purposeful (the assumption would then fall as such that anything “can” then happen from these means).
And so this can then result as: *Making moral consideration on *new congnizant patterns of other beings**. Mark Z, states Levin as well, have their own means in their respective researches.
Trade Book and Academia?
Pollan encourages a “general audience” instead of being overly technical due to broad reaching of audiences for more impactful/academic pursuits/fields.
Introduction
Biological systems are self-constructing, multi-scale agents navigating diverse problem spaces, not just chemical reactions or pre-programmed computational models.
Traditional binary distinctions (living vs. machine, natural vs. artificial) are blurring due to bioengineering and a continuum of life forms.
Levin’s framework aims to recognize, create, and relate to diverse intelligences, regardless of origin (biological, AI, synthetic).
A spectrum of “persuadability” exists – from hardware modification to training – for interacting with different agents, which needs experimental evaluation.
Biological Substrate as “Agential Material”
Life scales cognition; it doesn’t arise magically. From a single cell (oocyte) to complex organisms, there’s a gradual increase in cognitive capabilities.
We are all collective intelligences. Even a “unitary” organ like the pineal gland is made of many cells, each with internal complex processes.
Single cells (e.g., *Lacrymaria*) demonstrate surprising problem-solving and learning abilities without nervous systems, highlighting inherent cellular competency.
Biological systems exhibit a multi-scale competency architecture. Molecules, cells, tissues, organs, and swarms *all* solve problems in their respective spaces.
Life navigates many “spaces” beyond physical 3D: gene expression, physiological states, anatomical “morphospace” (the space of possible body forms). Focus of this video is Morphospace.
Morphogenesis as Problem Solving
Development (morphogenesis) isn’t just reliable; it’s actively *problem-solving*. Embryos adapt to perturbations (like being cut in half) to achieve correct forms.
Regeneration (e.g., axolotls) demonstrates this adaptability. Cells regrow lost structures and “stop” when the correct anatomical pattern is achieved (anatomical homeostasis).
Picasso frogs, with scrambled facial features, show that development isn’t hardwired. Organs move along *novel* paths to reach correct positions, exhibiting context-sensitive adaptation.
Bioelectricity is a “cognitive glue” linking cells toward collective morphogenetic goals.
Bioelectric Communication and Control
Number of individuals (embryos) in an early blastoderm can vary. It is not predetermined. Alignment and work (towards development) of individuals determines the final state, not pre-programmed count/data/instructions in the DNA.
Life emphasizes “saliency” over information fidelity. Metamorphosis (caterpillar to butterfly) demonstrates remapping, not just storing, of information; memories are reinterpreted.
“Bowtie” architectures, common in biological networks (chemical, biomechanical, bioelectric), force generalization and creative reinterpretation due to information bottlenecks.
Creatures constantly reinterpret their current state and memory engrams due to the inherent unreliability of the biological medium (noise, plasticity). Past is available for reference, but there is heavy emphasis on new interpretations from present data.
Development is a type of de novo, from new, every moment, of problem solving. Organisms reuse existing molecular tools in novel ways (cell communication vs. cytoskeletal bending in newt kidney tubules) to achieve desired anatomical outcomes.
Cells within networks exhibit “pattern completion” abilities and have setpoints.
Cancer is a failure of cells to adhere to the larger, collective goal, reverting to small individual goals. Disconnect from group electrical communication means their sense of “self” contracts to the size of just that cell.
Every cell, not only neural, use electricity to communicate. Like nerons, they use voltage gradients, ion channels, and gap junctions, a principle exploited from very beginning (bacterial times).
Bioelectric pre-patterns guide organ formation, a “electric face”, a map or pattern before anatomy appears. These can be manipulated to alter morphology.
Modifying bioelectric patterns (by controlling ion channels/gap junctions, like neural synaptic plasticity) can induce organ growth (eyes in tadpole guts), respecify body plans (two-headed planaria), or change plarania’s head shapes into those that resemble that of different plarania species..
Altered biolerical state can be perminant (plarania growing two heads for rest of it’s lives, generation after generation)
Evolved structures can represent “latent spaces”, spaces that are accessabile, despite not being the default state, where they can go given the correct stimulation (in this case, bioelectrical, and chemical was mentioned with reference to gall wasps)
Emergent Capabilities and Future Directions
Wasp galls exemplify a new capacity in a given organism. Acorns can be made to grow very very different strcutures from acorn with signals from gall wasps. This new shape would never normally happen from the original.
Anthrobots (human tracheal cells forming novel structures) and xenobots (frog cells forming self-replicating organisms) demonstrate surprising, emergent capabilities *not* dictated by selection or human design.
These emergent capabilities may arise from an “external component” a *latent space* of possibilities beyond specific genes or algorithms, which biology (and future bioengineers) can explore.
Living systems are incredibly plastic. Any combination of evolved/engineered material and software can potentially form an “agent.”
A field of “diverse intelligence” is crucial to understand, relate to, and ethically interact with these novel beings and their unconventional minds (“synth biosis”).
Even sorting algorithms can have simple cognitivies behaviors. Life is more cognitive than it’s been preveious recognized or defined by.
Goal: Develop principled frameworks for recognizing and interacting with diverse minds by overcoming our own evolutionary biases, using AI as translators.
There exists company-interest connections.
Intelligence and Emergence
Intelligence is deeply ingrained in the multi-scale architecture of living organisms. It exists from molecular networks up. AI can help understand biology, and biology can inspire new AI.
Emergence is important, but biology’s key is closed-loop, goal-directed (cybernetic) agents at *all* scales, not just emergent outcomes. These agents actively reduce errors from set points, using energy to achieve goals.
Higher levels of organization can do causal work that’s not reducible to lower levels (referencing Julio Tononi and Eric Hoel’s work on Information Theory). This can be mathematically determined, not just philosophically debated.
Membrane voltage, a high-level aggregate, is more useful for regenerative medicine and other interventions than tracking individual ion positions. This highlights higher-level causal power.
Agency and Observer Relativity
Agency is best described along a continuum (a “persuadability continuum”), not as a binary property. Sharp categories are imposed by observers on continuous phenomena.
Levels of system, is based of what is being observed. It can shift with what the observer.
The claim that a system occupies a particular place on this agency continuum is observer-relative. It’s not an objective fact, but depends on the observer’s model and how well that model lets them interact with the system.
A crucial point is that the system *itself* can be a valid observer, creating internal models and controlling itself. This “self-observation” is a key aspect of agency.
Making the claims about systems are done so in respect of a hypothesis on an individual system by a researcher/obvserver.
Physicalism, Reductionism, and Matter
Challenging reductionist physicalism doesn’t negate physicalism itself. It argues that matter is capable of far more than often assumed, including higher-level organization with non-reducible causal power.
Saying an entity isn’t made of a kind of matter isn’t helpful in describing intelligence, given that all things have the same underlying “stuff”.
“Pseudo-problems” arise from making objective claims in this area. Specifying the vantage point (observer) clarifies many issues.
Any computation or intelligence isn’t necessarily reliant upon some sort of external observers but instead it itself can observe, as this defines this form of being (a legitimate entity in-and-of-itself).
Phase Transitions and Continuity
While sharp, non-linear changes in behavior (akin to phase transitions) exist in AI and other fields, biological cognition appears more continuous. Sharpness in phase transitions often increases with system size.
It’s difficult to pinpoint a specific “phase transition” where cognition *suddenly* emerges in development or evolution. The substrate is continuous; it’s a transformation of the *same* material.
Attributing cognitive capabilities to single cells is controversial, but critics need to provide a concrete explanation of *how* and *when* full cognition emerges during development or evolution. Simply stating this point in time isn’t concrete.
This doesn’t equate the lack of understanding transitions in understanding and quantifying them for future predictions (on future research, technologies, etc etc).
Defining and Measuring Intelligence
Even gene regulatory networks (seemingly mechanical) exhibit various forms of learning (e.g., associative conditioning), showcasing non-zero intelligence at very low levels.
The absolute minimum of intelligence likely involves: (1) some level of goal-directedness (least action principles, even in particles) and (2) some indeterminacy (local conditions not fully determining behavior).
This doesn’t need to equate to 0 levels. Even at atomic scales and even in a “cognitive vacuum” these principles can exist as its “basement”, so long there are actions, reactions and movements of its components.
Measuring intelligence is taking an IQ test ourselves. We must define a problem space, identify the system’s policy, and test the policy’s “cleverness.” We might not always recognize intelligence, especially outside familiar domains.
Intelligence doesn’t rely on conventional intelligence such as with brains.
Researchers must actively observe, analyze and quantify intelligence in non-human systems by defining their: behavior and environmental space (that it is doing it’s intelligent “stuff”).
Intelligence is related to its capabilities in such a “problem space” and may depend, for example, on navigating this space and/or physiology.
A continuum of competencies exists, from simple systems following energy gradients (like magnets) to complex systems overcoming obstacles to reach goals (like Romeo and Juliet). The is quantifiable in this way.
Research shows that there is intelligence in simpler systems.
The challenge is perturbation. It changes goals of individual parts that normally seem fixed on fixed behaviours.
A system’s intelligence is revealed by challenging its “normal” behavior with perturbations. Often, systems are more capable than initially assumed. This makes assumptions hardwired (fixed on set behaviours, that would have difficulty adjusting/doing so within the limits/timescale for measuring and/or perturbation.
Goal-Directedness, Anthropomorphism, and Interaction
A system’s “goal” is a useful lens or perspective for prediction and control, not necessarily an objective, discoverable fact. Positing a goal and competencies helps us interact effectively.
Such can also involve more complicated relationships than what would be considered standard such as “friendships” if they allow for an optimized control and behaviour, for both sides involved in a system.
“Anthropomorphism” isn’t a useful concept. We should make *specific*, testable claims about a system’s capabilities (human, robot, cell, etc.) and empirically determine their validity. We often *underestimate* intelligence, especially in biology.
The assumption on an entities “naturality” also influences people on intelligence due to biases in a pre-garden assumption that doesn’t account that nature (life) doesn’t set you up, as Evolution simply aims for replicating with variations.
On Natural, Biomass and Goals
Nature’s is set to survival by the “means of replication”, which is done from stochastic-gradient descent.
Different solutions can apply/be found on how goals and intelligences can manifest that differs from humans.
Humans desire/instinctually don’t enjoy nature because their current goals, even before major technological advances, often had goals (in many examples given, even primitive ones such as an umbrella, showed examples and reasoning on why and how to differ goals).
Humans don’t have innate, fixed and/or objective/optimal goals (even pre-Garden as there isn’t necessarily something objective to begin with (on what would constitute these parameters).
Humans also can influence these biases as with tools or modifications of any level.
Transhumanism and Categories
Biology is highly interoperable. We can create chimeras/hybrids between biological tissues and machines, blurring boundaries between “human” and “machine.” There is it continuous with many variances in percentage. This challenges binary categories.
A goal of intelligence research can mean to “better what is defined as a natural body” (of people for example).
There exist human, biological extensions to even radical extremes that makes distinctions for transhumanism that blurs these parameters of distinction.
Categorizing organisms/systems as natural kinds (with sharp, inherent boundaries) is limiting. Biology and technology push these boundaries, demanding a more nuanced perspective.
“Natural” is often undesirable. Evolution optimizes for biomass, not human happiness or fulfillment. We can (and should) strive to do better than “natural” through science and technology.
What is viewed as unnatural could well be natural too due to emergent, external factors that affect changes in human’s systems in question and the research or development being studied or discussed.
Error Correction and Existence
The universe exists because it’s not impossible; existence is “free.” Everything happens (superposition of all operators), leading to apparent indeterminism at the lowest levels.
Stable structures emerge due to error correction. Patterns that are statistically stable persist (like vortices in water), while others dissipate. Particles themselves can be considered error-correcting codes.
Life itself is considered an error correcting code, stabilizing particle configurations and controlling regions of the Universe.
Mental representations can be seen as error-correcting “quasiparticles,” similar to how sound (information-preserving) emerges from molecular activity.
The Free Energy Principle (Friston) aligns with this: systems persist by maintaining a state-space boundary, acting autopoeitically, which can be interpreted as informational error correction.
This relates to Zurich’s Quantum Darwinism (selection of stable states by environment) and AdS/CFT correspondence (coupling of volume state to surface state), although SpaceTime is considered emergent by the Speakers.
Coherence and Agency
Large Language Models (LLMs) try to minimize the error of language prediction. For smaller organisms in concrete situations, there is minimization.
Consciousness may be a “coherence maximizer” rather than a prediction error minimizer, actively imposing order on trainable substrate, acting “as if” a single agent. It creates local coherence.
LLMs, even when trained on massive data, may lack the creative coherence found in biological systems. This ties in with human creativity.
Human’s coherence may result from human evolution as “domesticated” primates. Some, however may be Generally Intelligent.
Sense-making is a primary goal of Consciousness, and the use and meaning of memory as being malleable; memory re-expansion as creative reinterpretation of sparse engrams (compressed representations).
Uncertainty and confabulation: Memories, due to compression and loss of context, require active, creative reconstruction; it’s not about what the memory *was*, but what can be *done* with it *now*. The example given being caterpillar to butterfly memory transition, even during brain changes, the previous information still useable.
Percepts from external world share similar principles, in how sensations may dictate a given response.
This aligns with “appropriate action” as a way to think about coherence (following Bateson’s “differences that make a difference”). Action, prediction, and testing are closely linked.
Modelling adjacent events in texts are problematic, whereas images benefit in the use of convolutional networks due to ajdacent semantical related pixels. Working Memory require the construction of future spaces.
“You six months ago isn’t answering emails”: a key collaborator being, well… yourself and messages.
Self, Identity, and Control
Distinguishing between self-caused events and external events is evolutionarily important (heuristic processing). Errors in this distinction can be debilitating.
Language input is processed in time-windowed chunks, analyzed for consistency, but still addressing a “what do I do next?” question at a larger time scale. We impose geometry on information.
Agency is control of future state, not just present state. Humans seem better than current AI at temporal coherence (preserving information over time).
Identity is instrumental for credit assignment. Without a reason to be distinct (e.g., unique memories), self-models wouldn’t form.
Training vs. being trained: Some entities (like cats) have evolved to train others, not just be trained (similar to governments recursively “bullying” people).
Colonization: Entraining an environment to extend oneself; building structure that reflects and sustains the colonizer. The ease to expand consciousness (cohesion, bandwith).
Control structure: The invariance, what’s controllable. The good regulation theorum suggests it should be, learnable.
Classical Communication and Observation
Error correction necessitates classical communication. Discrepancies between expected and observed communication constitute errors. It needs a pre-shared base for meaning (shared context or langugae), where one’s message or data can be seen as noise.
Classical communication involves assumptions about thermodynamic irreversibility (imposing order on superpositions, which creates an actionable representation, similar to defining actions/non actions, and having models for them, while collapsing the wave-function ).
Observers exist in collapsed timelines. Observation requires a classical self-model; multiple possibilities within a “self-frame” are in superposition, requiring distinctions at larger periods.
The perception of the importance of events can happen during larger time windows than when the actual events transpire. For example a femtosecond transition resulting a smoking up computer.
Protocognition at the Smallest Scales
The idea that there is some protocognitive capacity at the lowest levels of scales and their interactions.
At the lowest level (elementary particles), there might not be “intelligence” but self-propagating patterns (like vortices), selected by a kind of evolutionary process.
Intelligence emerges with persistent particles that form multistable structures and exploit negentropic gradients. Life performs controlled reactions, outcompeting “dumb” reactions.
The first/simplest “thing” doing this is unclear, could be something very large but no-biological entities may use some as-of-yet unknown method/medium for computation. Cells are very complex, and agency might form at planetary scales before cellular life. Atoms/molecules, as breaking symmetries, represent a crucial level.
Simulation would be useful, where, using deviation of observations against known expectations, levels of “control” of an observed system may be observed/determined.
Criteria for diverse intelligence: Finding intelligence at different scales and forms requires new tools and criteria. Experiments (like with minimal matter) might reveal unexpected behaviors. It is possible this emergent behaviour to happen from extremely simple origins/systems.
A self-organizing, generally intelligent system would need a “colonizing seed” – a minimal pattern that induces coherence. This could be seen as minimizing constraint violations, or a consensus algorithm maximizing true statements.
This self-observing, self-stabilizing “observer” is fundamental to consciousness, with an information and substrate agnostic in variance (informational equivalent for the next level of the microbial map to organisms).
Organisms’ and structures’ (ex: Elephants’) constraints in capability/potential maybe tricky (for Evolution to reach), but are important considerations and topics for the exploration for cognition and intelligence. This requires the need to discover it’s invariance and formalize its mechanisms (to which, instruments, memory may be useful factors).
Environmental stimuli, despite complexity, can yield an observable outcome that some animals react to; indicating the inherent use-value and nature, or the shared perception of its emergent representation (sounds, shapes).
The body plan/anatomy as a tool or substrate, through which the Brain and thus computation may expand upon, and explore the (cognitive) world with (hands, trunk, voice, etc).
Development requiring random stimulation in order for optimal, “intended”, operational value to form (the use of Hands require their “random” exploration as an infant, like their initial babbling before structured understanding).
There may be other undiscovered ways an entity (biological/artificial) uses different means/structures/bodies and its environment, and what tools may be utilized/required for the interaction.
Introduction: Children as “AI”
Levin starts with a description that sounds like advanced AI, but he’s actually referring to human children, highlighting the inherent unpredictability and replacement that come with creating *any* new intelligence. This underscores that anxieties about AI replacing us are not new; they are ancient, existential concerns.
Existing systems are created, given autonomy, released, and some, whether good or bad are empowered to perform those behaviors in the future, a license is required to fish but not for parenting.
All forms of existing, adaptable life eventually cease.
Synthetic vs. Natural: A False Dichotomy
People tend to categorize synthetic beings (like AI) as fundamentally different, but the real questions about creating high-capability beings, releasing them into the world, and having limited control are the *same* for both natural (children) and synthetic entities.
People may have made judgements and statements without actually doing research first.
Adaptation, Persistence, and the Future
A species that doesn’t change/adapt will die out. A species that does change is also, technically, “gone,” replaced by its adapted form. This paradox applies to all evolving systems, including humanity.
The key is not “persistence as a fixed object” but “persistence as a process” (like process philosophy). The interesting question is not *if* we change, but *how* we want to change, individually and as a species.
Humans in 100-200 years might not accept the limitations of the current human condition (e.g., lower back pain, diseases, birth defects) as inevitable; and future might consider those who refused to adapt unfathomable. Freedom of embodiment and deliberate change will likely become norms.
It is not necessarily intelligence as brains, which evolve in nature, but different things altogether that should also fall under the terminology.
Diverse Intelligence and Overcoming Categorical Distinctions
The field of “Diverse Intelligence” is growing, and challenges traditional, narrow definitions of intelligence and mind.
Diverse intelligence seeks commonalities across *all possible* intelligent agents, not just those that are biological or brain-based. This includes considering radically different substrates, sizes, and “spaces” where intelligence can operate (not just 3D space).
Confabulation (creating plausible explanations without complete information) is a feature of intelligence, not a bug. It’s essential for compression of experience, learning, and creativity, but it is present in biological intelligences as well as in certain simple mechanical processes as well, not exclusive to machines and software.
The interviewer refers to previous chat about bioelectrical intelligence which led into a discussion about diverse intelligence, the rapidly-growing field.
The question of where to look for mind and agency should remain at large for all of existence.
People tend to assume a categorical distinction between a physical, cognitive system and the thoughts being carried and transferred between those physical entities, which isn’t necesarrily valid; both could exist on a fluid spectrum, and physical objects we percieve to be simplictic might turn out not to be so, due to our current perspectives of the subject matter being in their infancy.
Ethical Implications and “Synthbiosis”
The ethical considerations of creating/interacting with diverse intelligences (including augmented humans) are significant. Categorical thinking (“us” vs. “them,” natural vs. artificial) is dangerous. The focus should be on matching “cognitive light cones” and sharing existential concerns.
“Synthbiosis” (a term coined by chat gbt at Levin’s request): A positive, creative collaboration between biology and synthetic entities.
We can learn things that we would not discover alone from any community.
Machine vs. Human: A Misguided Question
The question “are we machines?” is ill-posed. “Machine” is a *lens* or interaction protocol, not an essential property. Different lenses reveal different aspects. We shouldn’t argue about *what something “really is”* but about the *utility of different perspectives*.
There are certain areas and levels within which complex intelligence can operate and have very high level efficiency with it’s operations.
We have profoundly misunderstood “simple” machines. There are “protocognitive” properties even in very simple systems (e.g., unexpected capacities in the “bubble sort” algorithm).
Humility is crucial. We haven’t achieved anything that even science-fiction works could’ve predicted.
Moving Forward: Kindness and Avoiding Fear
Fear and scarcity mentalities, the feeling like care is a ‘zero-sum-game’ hinder a greater more caring environment being created by humans.
Unwarranted certainty about consciousness and cognition is dangerous. The field of Diverse Intelligence is just beginning, and many fundamental questions remain unanswered.
The two primary ways we get it wrong with ethics: to value something less, or to give things compassion which don’t need them, with the former a bigger danger.
Levin suggests prioritizing kindness, compassion, and recognizing the potential for sentience in unconventional forms, rather than being driven by fear of the other. This includes acknowledging that we may need to greatly expand our “circle of compassion.”
We shouldn’t go backward towards ‘simpler’ less-developed civilizations that have a “one-ness with all life”, but use them as inspiration or guidance towards a scientific discovery in similar manner; or, a *starting point*.
Dan Dennett’s philosophy is brought up: To discuss his generosity as an inspirational, intelligent human and that what made philosophy so compelling, was that Dennet did actual physical science/experimentation in additon to discussing philosophy and working out problems and discovering more together, even those that he disagreed with.
Introduction: Planarians and Regeneration
Planarian flatworms are a model for regenerative medicine: They can regenerate any body part, including the brain, are highly cancer-resistant, and are biologically immortal (do not age). Understanding their regenerative mechanisms holds promise for human medicine.
Cancer as a Disruption of Cellular Communication
Cancer is fundamentally a breakdown of multicellular cooperation, not simply a result of DNA mutations. Individual cells disconnect from the larger bioelectrical, chemical, and mechanical network that coordinates tissue organization.
Disconnected cancer cells revert to a primitive, unicellular state, focusing on self-preservation and proliferation (like amebas), treating the rest of the body as an external environment.
The “cognitive light cone” concept describes the scale of goals a cell or group of cells can pursue. In cancer, this light cone shrinks from the whole-body level to the individual cell level.
The DNA provides instructions for making cellular “hardware”(proteins), *not* a direct blueprint for the body’s form and function. It creates components allowing complex and adaptable computation. The overall anatomical form comes about as the network (as the “software”)
The analogy to computer hardware/software highlights that focusing solely on DNA (hardware) is insufficient for addressing many regenerative medicine challenges, analogous to fixing computer issues through soldering components alone.
Focus should instead be on “reprogramming” the bioelectric signaling that directs cellular behavior (the “software”), by taking advantage of cellular natural existing behavior of cell’s computation.
Bioelectric Signaling and Pattern Memory
Cells communicate through electrical signals using ion channels (protein gates) that control the flow of charged ions (sodium, potassium, chloride, protons). This is literal electricity.
Groups of cells form bioelectrical networks similar to neural networks, storing memories and making decisions. These networks are essential for maintaining anatomical structure.
The “target pattern” (correct body shape) is stored as a memory within this bioelectrical network. Cells actively work to minimize deviation from this pattern during regeneration.
Fluorescent voltage-sensitive dyes can be used to visualize these bioelectrical patterns in living tissues, allowing researchers to observe defects associated with cancer, injury, or birth defects.
The first step to cancer (with use of exmaples) includes cells disconnecting their bioelectrical signalling with their neighbor cells, leading to the ‘rolling back’ and cancer and it’s cancerous and invasive behaviors, even in spite of genetic makeup.
Artificial re-connecting bioelectrical patterns show normalizing potential to cancerous properties, showing there exists ‘communcations’ to cells which guide them how to ‘behave’.
The research goal is to develop computational models and “electroceutical” drugs (ion channel modulators) to correct aberrant bioelectrical patterns and restore normal tissue organization.
Implications and Examples
The approach is *not* about killing cancer cells (like chemotherapy) but about restoring their cooperative behavior within the tissue. It also does not aim to “fix” DNA mutations.
Frog experiments show that induced tumors (even with human oncogenes) can be suppressed or normalized by manipulating bioelectrical connectivity.
GBM or Glioblastoma Multiforme (a brain cancer), is another application example, where “turning off” this bioelectrical signalling can revert cancer’s typical ‘rolling back’ cancer cellular activity, restoring them as an ‘ideal’ state as per thier normal function in context of the body’s multicellular function.
Similar normalization of cancer cells has been observed in salamanders and mouse embryos, highlighting that this is a biological possibility, not just a theoretical concept.
Planarians, with their messy genomes, demonstrate that perfect DNA is *not* required for robust regeneration and cancer resistance. Their cells have evolved to handle genomic noise.
Cellular decisions are *not* solely driven by genetics and *are* made. They adjust and modify their behaviors in response to environmental stressors, including chemotherapy; they exhibit a basic form of intelligence (“basal cognition”).
Problems exist in physiological nature (how cells remember what pattern to exist as) vs. purely DNA hardware. For example: mirroring pattern (e.g. mirror effect in epileptic brain) or even cardiac patterns of mis-regulation and re-triggering.
Addressing Common Misconceptions
Family genetic predispositions (hardware): “protecting the hardware”. And inherited predispositions, play roles with likelihood.
While maintaining healthy DNA (avoiding damage) is beneficial, it’s not the complete picture. Bioelectrical disruptions alone can cause cancer.
Age and Cancer: increased loss of regenerative and electrical communication abilities lead to increase chances for cancer to rise.
“Ion channel drugs” (electroceuticals), often already used for conditions like epilepsy, can be repurposed to modulate bioelectrical signals, potentially on a systemic level. This differs significantly from targeted therapies or electrical shocks.
For adults, medication benefits (like blood pressure medications that affect ion channels) generally outweigh any theoretical risks related to bioelectrical disruption.
For embryos, however, are subject to issues by taking some bioelectrical medication.
Liver can regenerate (and do what planarians do). human can regenerate fingertips (the very tip-ends, the very distal parts of the finger body parts, and at an earlier-human stages of growth (kids vs. adult).
Minimum Unit of Intelligence, Evolution, and Behavior
Levin discusses embodied cognition and intelligence, focusing on minimal systems exhibiting cognitive behavior, beyond consciousness. He uses sorting algorithms as simple models.
These algorithms, though deterministic and transparent, demonstrate unexpected behaviors like “delayed gratification” (going temporarily backward to achieve a larger goal) and emergent goals.
There are surpises to be had on intelligence and competency, they can exist outside the known algorithm.
The algorithms cluster according to type (“algo-type”) during the sorting process, an emergent property not explicitly coded. This suggests “surprise minimization” as a driving force.
This work challenges assumptions about intelligence requiring late evolution, complex neural systems, or human-level goals. Even simple systems have “basal intelligence” (problem-solving to reach a goal by different means).
The focus is not *inventing* the intelligences. But *discovering* them.
Biobots (Anthrobots and Xenobots) and Emergent Goals
Levin’s lab works with Anthrobots (human tracheal cells) and Xenobots (frog skin cells), showing unexpected behaviors and capabilities.
Anthrobots, created by altering growth conditions (3D matrix then low-viscosity medium), spontaneously form motile structures that *heal neural wounds*.
Naming them “bots,” not just “organoids,” encourages exploring their potential for programmability and diverse applications beyond simply modeling organs.
It also reveals an issue that cells already know how to work together towards some competency before being programatically commanded, which is like and unlike AI safety issues.
These bots reveal “emergent competencies”—capabilities not directly programmed or selected for. The question becomes: Where do these novel goals originate?
Levin believes understanding and controlling these emergent goals is crucial, not just for medicine but also for understanding collective intelligence in general (e.g., swarms, AI).
This suggests new field of figuring out new goals of novel systems when and where evolution has no selectional history, this will become important.
Regenerative Medicine and Philosophical Implications
Levin’s research aims to “crack the morphogenetic code”—understanding how cell collectives make decisions about form—with the long-term goal of *in vivo* regeneration (regenerating tissues within the body).
There are also some interesting and emergent ideas: 1. cells could perhaps inherit other minds and also it opens up what it would actually mean for “you” or your identity. 2. Is it actually limited. 3. Do we actually only know what a living organism wants when they leave the constraints of what we normally observe it in, what constraints can we observe. 4. A science of mind at a distance.
This challenges traditional bioengineering approaches focused on micromanaging at the molecular level. The goal is to “persuade” cell collectives to achieve a desired form.
His view, intelligence is likely common, often appearing in “unfamiliar guises,” and the active inference framework of “free energy/surpise minimization” is perhaps key, cells strive for predictability.
This suggests there could be a type of chemistry of of the platonic/idea space, even in a non-spiritual way, if one knows where/how to observe.
One can make discoveries into what to do when you observe, instead of having an invention/creativity or something with no limits, we all have limited discoveries to make based on observation.
Broader Context of Diverse Intelligence
Traditional frameworks of “mind” (based on our usual forms and space-scales) “break” when one confronts the different capabilities across the scales/environments of lifeforms/organizations.
This isn’t just about the human scale, for example the potential for large scale behaviors, from the cellular all the way up to, perhaps, a planetary scale.
Intelligence and consciousness should *not* be conflated. Intelligence (problem-solving) is easier to study objectively; consciousness remains difficult.
Embodiment can occur in various “spaces” (physiological, transcriptional, etc.), not just 3D physical space. Perception-action loops are key, not necessarily 3D movement.
We need humility in recognizing intelligence in unconventional forms. Perturbational experiments, not just observation, are crucial for testing whether systems exhibit cognitive capacities.
Introduction: Unconventional Intelligence
Biology demonstrates intelligence at all organizational levels, with agents solving problems in diverse spaces beyond the 3D world.
Neuroscience is an elaboration of an ancient computational capacity arising from developmental bioelectricity.
Bioelectricity enables cell collectives to navigate “morphospace” (the space of anatomical configurations).
Multi-scale competency architecture means all levels can perform their own goal driven activties and is is exciting for novel engineering.
Examples of Problem Solving in Non-Traditional Spaces
Planaria adapt to barium by altering gene expression, demonstrating problem-solving in “transcriptional space.” No evolutionary advantage exists as planeria would not have encounted barium in the wild, indicating how did cells “know” what genes to toggle?
Tadpoles with eyes grafted onto their tails can still see, showing functional plasticity despite altered body structure. Tadpoles and Caterpillars have bodies changing and demonstrate memory persistance through drastic changes.
Planaria can regenerate any body part and its tail demonstrates imprinted data after being headless. Planaria demonstrates memory spreading beyond brains.
These examples challenge the notion of fixed developmental programs and highlight the adaptive capacity of biological systems.
Cells as Competent Problem Solvers
Single-celled organisms (like Lacrymaria) exhibit remarkable control and problem-solving at the individual cell level.
Multicellular organisms scale up this cellular competence to achieve complex morphogenesis (building body forms).
teratoma a growth showing correct tissues (skin, muscle, hair, etc.) but lacking spatial organization demonstrating hardwares correct, but the anatomical positioning missing
Morphogenesis isn’t just “forward emergence”; it’s intelligent, achieving goals via varied paths despite perturbations (William James’s definition of intelligence).
Axolotls regenerate limbs, demonstrating the ability to achieve a specific anatomical target from different starting points.
Kidney Tubules forming despite drastically varying numbers of cells making use of cell-cell commincation in normal circumistances and cyctoskelatal bending in mutated oversized cellls.
picasso tadpoles moving organs to turn from “picasso tadpoles” with missplaced face parts into standard frogs indicating a goal, rather than hardcoded movements
Evolution produces problem-solving machines, not just hardwired solutions, enabling adaptability.
Bioelectricity as the “Cognitive Glue”
Bioelectricity is a key mechanism for coordinating cell activities into a collective intelligence.
It’s not just for brains; all cells use bioelectrical signaling, a system predating nervous systems (evolved from bacterial biofilms).
Cells have ion channels and electrical synapses (gap junctions), similar to neurons, but operating more slowly.
Levin’s lab developed tools to observe and manipulate bioelectrical gradients in vivo, analogous to optogenetics in neuroscience.
They control network topology (gap junctions) and electrical states (ion channels) using drugs, light, and other techniques.
Monitoring involves voltagae and observing them by fluorescent voltage reporter dyes (simlar methods as zebra fish and mice)
They also do bioelectrial intervention with specific RNA that prevent decoupling, or other that encourage it, as necessary.
Bioelectrical Control of Development and Regeneration
Oncogenes cause cells to depolarize and disconnect, leading to tumors. This can be prevented by manipulating ion channels. The *electrical* state, not the oncogene itself, dictates the outcome.
The “electric face” in frog embryos pre-patterns anatomical features *before* gene expression. Bioelectrical patterns guide cell behavior.
By manipulating bioelectrical states, researchers can induce ectopic organs (eyes, limbs, hearts, brains) in tadpoles, demonstrating control over large-scale anatomy. They can also correct birth defects by restoring the correct bioelectrical “memory.”
frog embryo experiment changing where “build an eye here” usually sits within an embryo, and cells comply creating eyes with lens, retina, and optical nerve.
Planaria provide a model for studying bioelectrical pattern memory. The “target morphology” (number of heads) is stored in a stable, rewritable bioelectric circuit, *separate* from the anatomy and genome.
This bioelectrical memory can be manipulated to create two-headed worms, or even revert head shapes to those of *other* planarian species (demonstrating access to latent morphogenetic possibilities).
Limb regeneration in frogs can be induced by applying a cocktail of ion channel drugs, triggering a bioelectric state that initiates limb regrowth.
Bioelectrical manipulations have stable, conditional, long term re-writeable recall and guide cells as it decides the body’s structure.
Synthetic Biology and the Origins of Collective Goals
A fundamental question is where collective goals come from, and how cell collectives settle on specific targets.
Testing by combining 2 groups with cells: cell A “make a flat head” and cell B “build a round head”, is impossible to infer without testing if A>B or B>A or something in between.
To explore this, researchers created “Xenobots” – novel organisms from isolated frog skin cells (Xenopus laevis).
Cells, not normally the goal or outcome from an isolated tadpole, spontenously forms proto organisms which could mean skin’s “default” behaviour is a xentobot (previous asks said bioelectricity is the main reason, clarify?)
Liberated from their normal environment, these cells reboot their multicellularity and form motile, self-organizing structures.
Xenobots exhibit complex behaviors: movement, navigation, interactions, and even rudimentary maze-solving.
Xenobots also show regenerative capacity, reforming after being cut in half. They exhibit calcium dynamics, even without neurons.
Calcium flashes between xentobots are communication?
The genome of xenobots is 100% *Xenopus laevis*, highlighting that these novel behaviors are not encoded directly in the DNA, but emerge from the cells’ inherent plasticity and self-organizing capacity.
where this “goal” and outcomes “appears” in 48hrs is not entirely understood.
testing is being conducted, and can show cognition tests, training, learning, and many other behaviors is not explained.
Implications for Bioengineering and Beyond
Bioengineering allows manipulation at all levels of organization: cellular, organ/organism, and collective.
Creating Cyborgs, Hybrid Agents with diverse configurations/cognitive capacity
This opens up a vast “option space” of novel agents, blurring the lines between natural and artificial.
Chimeric bioengineering can reveal how collective goals emerge and how minds and bodies map to each other.
It pushes our thinking about Ethics when relating to beings made in unfamiliar ways (evolved? engineered? bioengineered? AI? how to decide?)
Multiscale competency architecture, with bioelectrical networks, allows individual cells with local goals to scale up into collectives solving larger problems.
Conclusion
Morphogenesis is an ancient proto-cognitive process, exhibiting problem-solving across multiple scales.
Synthetic biology demonstrates we can’t easily predict the behavior of large collectives, even if we know about the individual subunits.
This work challenges traditional notions of “organism,” “machine,” and “robot,” necessitating new ways of thinking about agency and ethics.
“mind” and “body” distinctions are going to change, and evolve to become a “continuum”.
Information vs. Meaning
Information, in a physics sense (Shannon), is about the implausibility of an event (self-information) or its average (entropy). Meaning arises from the relationship between an observer and the observed, representing Bayesian beliefs about the observed.
Meaning is what neuronal activity *represents* about sensory input, not the activity itself. It’s about beliefs *about* something, requiring a separation between observer and observed.
The Bowtie Architecture and Time
Information can be viewed on behavioral, developmental, and evolutionary timescales. We access the past through “memory engrams,” and use them for forward-aimed behavior.
The “bowtie” metaphor: Experiences are compressed into a generative model (left side). The *interpretation* of memories/information (right side) is fundamentally creative, going beyond algorithmic processing.
Analogy to Information Bottleneck. The present is the bottleneck, and is the simplest Markov blanket. The future and past are on the sides.
Compression is crucial. Systems infer the causes of compressed sensory input. This decompression/disentangling is a form of inference – filling in the gaps – minimizing variational free energy. This is creative.
The “bottleneck effect”. There are simple explaintions found for interpreting and explaining data.
“In-painting” fills gaps within existing structures; “out-painting” extends into novel areas unprepared for by the past.
Xenobots, Anthrobots, and Novelty
There exist new expressions found that allows us to communicate with xenobots in new ways.
Anthrobots (from human tracheal cells) show massive gene expression changes in their new lifestyle, demonstrating plasticity beyond standard embryonic development.
These cells “go past” merely in-painting by using current systems for different function.
This highlights the ability of biological systems to go beyond past experience and solve problems in novel, adaptive ways.
Genomes and Plasticity
Genome and phenotype relationship. This is intelligent and not a mere map.
There is hiding of what the genes have to do as repairs take place.
The Genome as problem solvers. More is offloaded, giving space for genes to evolve in a creative manner.
The genome doesn’t encode memories of specific past environments, but rather “tricks” that enable a problem-solving agent to adapt to *various* (potentially novel) situations.
Evolution selects for “learnability,” not just “selectability.” The genome provides plasticity and the capacity to learn in an “unrealized” world.
Planaria as extreme examples. They have minimal reliance on their (unstable) genome, putting effort into algorithms that build a “proper worm” regardless of hardware. They can’t be transgenic.
The present contains, what’s necessary from the past in order to be ready to move toward the future.
Evolution/biology works from a place where they know change must occur. This means biology must create solvers which will solve whatever it comes across (a form of intelligent “beginners” mindset)
Because organisms/cells solve, even what should be harmful genes are no longer expressed in selection (as other organisms are ready). Thus evolution can proceed freely.
Noise, Prediction Error, and Creativity
Noise and the observer. Noise is how the system is defined. For example: quantum mechanics underwrite on noise.
What is considered “noise” depends on the observer’s perspective and expectations. It can be an opportunity for different computations.
The noise are like quantum fluctuation.
There is prediction error that exists within biological life. It serves a positive and functional use to understand and get “used” to its experience and environment.
Prediction error is distinct from foundational random fluctuations. It’s essential for belief updating and drives the dynamics of biotic self-organization – “In the absence of prediction error, I would be dead.”
The “Self” and the Present
The story is important because of how it continuously maintains the model through time. The view of the self.
The observer creates an observation to understand themselves. A story/narrative. The notion is self can easily be lost or disturbed.
To understand the world is the same as asking it how one should act.
Higher agency happens to do planning. But there exist simplified methods, “some of that ends up happening anyway as a byproduct of simpler things that they’re doing.”.
Data, and not the program itself, may dictate a large reason/control on how biological organization/self organization occur.
A model which persist well and make niche contructions (helpful models/mindsets/perspective), this could change behaviors drastically. This is not unlike cancer or human brains and persnality constructs.
The “now” is critical for the self. It defines a continuous process of storytelling, interpretation, and meaning-seeking, a “self-telling story.” The present is like communicating to your past/future.
Mortal vs simulated models. The real biological world cannot have simulations (running forever).
Mental action and attention are how one can “pick” out the organization that it deems to be fit (from previous model, experiences, senses, etc…)
This process extends to communication with past/future selves, constructing a narrative that integrates past experiences.
Mortality vs. Story of self and Biological and “Psychological” Disease
Meaning that is passed through the “bottleneck effect” (or time) can give more information of its systems.
Aging as goal disruption: Somatic intelligence can suffer “existential boredom” after achieving its morphogenetic goals, leading to degradation even *without* external damage.
Changing is the necessary of existence, we must keep going in different scales.
Mathematical “death” (oscillator death/equilibrium) occurs when a system stops changing, violating the dynamics necessary for life. Staying in one place is like being a closed system, so you can’t exist.
Staying too long (in on ideal “goal” like state), would break laws of dynamics of existence and change.
The “renewal effect”. Where change can occur. Such as to have “children”, even.
Higher/lower organisms. “The cycle of reproduction..”, there exist other organisms.
Story and Selfhood. “and ended by explicitly talking about the constructive aspects of this so-called deflationary account.”: “what we’re talking about is, you know, a really beautiful constructive process…”, they use creativity, active-reconstructing, understanding models to best organize.
Distinction dissolved, between psychological and bioogical systems, from attention, to action, to prediction errors:
“…ranging from hallucinations and delusions right through to, say, dysmorphophobia, you know, abnormal beliefs about my body, for example.”
This plays a role for biological and “psychological” well-being (example, Cancer with understanding boundaries).
The ability to predict future actions leads to having mental and internal (in brain) action (hypothetical). Having future in model means having bigger scales.
Having future hypothetical also require higher and bigger models and constructs.
Scale-Free Consiousness
Can consciousness itself occur in lower/higher organization?: …basically you could find processes that have the attribute of things that we consider to be conscious processes expressed at multiple scales.
Concerns, and Story Telling.
The risk. Science not always the answer, there are those who become unstable with science’s new findings that they feel themselves not exist, that the sense and beauty is removed.
Science, its constructive uses, should provide more to existence and understanding.
Introduction: AI & Biology Convergence
The podcast explores the intersection of AI, biology, computer science, and philosophy, focusing on a recent paper by Levin and Lopez.
The paper integrates multiple biological datasets (genes, drugs, diseases) into a unified network model, demonstrating a novel link between Gaba and cancer.
Multimodal Data Integration and Network Embedding
Challenge: Integrating diverse “omics” data (gene, drug, disease interactions) into a unified representation.
Solution: Developed a “universal multi-layer network” approach, enabling the combination of various data types and scales. Existing, independent datasets for Genes (protein interactions), and drugs (combinations of uses), and Disease (symptom commonality) were reconciled into a single, cross-referencable dataset.
Network Embedding: Translated network nodes (genes, drugs, diseases) into vectors using a similarity measure based on “random walk with restart.”
Random Walk with Restart: An algorithm that explores the network from a seed node, creating a probability distribution representing node similarity. This distills relationships (connections, probabilities of joint walks) from multiple datasources down into simplified repreesntation of data proximities.
Machine Learning Application: Enabled machine learning on network data, which usually requires vector representations, facilitating tasks like link prediction.
Process: a link is chosen and walks (of differnet lengths?) happen at randon. those that involve crossing the selected link (in this sense ‘linking’ them) causes them to get pulled closed, such that links become predictive of path similarities.
Training and validation: 70% for training, 30% hold-out data for making novel link predictions between Nodes (e.g. ‘what is predicted drug for unknown sickness’)
Gaba-Cancer Link and Validation
Prediction: The model predicted a significant link between the neurotransmitter Gaba (specifically the Gaba A receptor) and cancer (melanoma).
Mechanism: Gaba A is a chloride channel that regulates cell electrical state. Perturbing Gaba signaling can disrupt cell communication.
Experimental Validation: Using memantine (a Gaba A agonist) on melanocytes (pigment cells) induced a melanoma-like phenotype.
Unlike usual tumors: *No primary tumor.* All melanocytes converted at the same time. This demonstrates the cancer phenotype can result *without genetic damage*, resulting instead solely by changing electrical networking capability.
Significance: Demonstrates cancer initiation *without genetic damage*, only by altering cell-cell electrical communication (physiological change, *not* genetic.)
AI for Biology: Black Box Modeling and Data Limitations
Grand Challenge of Biology: Understanding complex biological interactions (the “black box”). Current drugs act *when* and *after* mechanisms is known; ML seeks *preemptive*, mechanism-agnostic methods.
Black box challenge: most interventions rely on knowing *exactly* where in a mechanisms one wants to act, as opposed to a *model* predicting interactions and cascading causal paths (downstream mechanisms).
Emergent Models: Some AI models (like Evo) learn higher-order concepts directly from raw data (e.g., DNA), potentially showing less “inductive bias”.
Levin: “Our [paper/approach] is Data-Driven” … aggregating a ton of ‘BigData’, and seeking correlations/predictions to study further and extract “Theory” out of them (Data first, theories second).
Bootstrapping Knowledge: AI-driven predictions, validated by experiments, could iteratively refine biological knowledge.
Data Limitation: Lack of standardized, easily-integratable and outcome-focused data is a major hurdle, as opposed to the availability of a ton of small, micro, data of individual components (such as single cells)
Need data focused on larger level: Beyond just having raw DNA and protein interactions. Biology lack large scale collections that document overall body states, and how it changes in a body (example, the changes in location of your *Bones*, or amount of Bones), in respons eto various experimental changes.
‘bioinformatics of *Shape*.’ describes anatomical+functional outcome measurements, in response to some change/perturbation to get cause->effects. Example, given *This* stimulis -> This change. Current amount of public data only “in the hundreds.”
Bioelectrical Data Missing: Current datasets largely lack crucial bioelectrical data, essential for cell communication.
Publication challenges: copying methodology and introduction paragraphs from papers you’ve already published is consider self-plagiarism, working *against* standardization (which needs consistency and no variation for parsing reasons)
Multiscale Competency, Intelligence, and Communication
Collective Intelligence: Biology exhibits collective intelligence across scales, with problem-solving (not just complexity) at each level (pathways, cells, tissues, organs).
Pathways Alone Have learning: habituation sensitization and pavlonian learning
Beyond Micromanagement: AI’s role may be less about precise, “bottom-up” pathway manipulation and more about “top-down” communication with biological systems. “Train a System” instead of ‘tweaking/clamping/forcing/fighting/hacking’ it.
Cognitive tools will be powerful in helping comunicate to intelligent (cognitively rich, problem-sovling, but unintellegent *like us*) systems, than any attempt to micro-manage the whole network.
Dogs vs Horses training analogies: successful use of ‘non-biomedical’ intelligence. Not micromanaging and fiddling with biology, but *communication/learning* via training, instead of ‘understanding what all neurons do’.
Theory of Mind: AI could help build a “theory of mind” of biological systems, understanding their “proto-cognitive” properties (goals, preferences). This would revolutionize how we study them.
No goal is “better” micromanging molecules. Real power in using these (learning algorithms?) as “Communication devices.”,” *Translators*”, “finding, unbiassed patterns and becom[ing] a translation tool”.
Data Limited Models vs Humans
Wellness prediction challenge (like next tokens), taking microlevel (“your strava runs” and food intake etc), and subjective well-being measure, can make wellness predictions difficult:
Paradox: delivering wellness predictions (of being in good moods or etc), might make someone immediately depressed.
What, actually are we optizing: we “don’t actually sure what to opimize” because long-run or short run, one value conflict. “multi-reward designs”. “scale reward not enough”, at minim, must take “long run” and short run trade-off”
Human limitations: “split brain patients” and multi-agent in the “the same person”: shows there “there exists multiple different opinions [..] and sometimes conflicting beliefs”. Not one solid person, but muliple and varied.
Current biological layers and computing: each computationallayer assumes other ones *won’t fail or vary* with *precision.* It makes it fragile. This assumption of no degradation in layers below limits power to handle dynamic, adaptive changes.
Implications for AI Architecture and Alignment
Evolution created solving machines *in various spaces*, like behavioral, etc.
Multiscale Architectures: Current AI (like transformers) often lacks the multiscale architecture of biological systems. Biological reliability requires *interpretion.* This gives rise to more “Problem solving” at various spaces.
Bio-Inspired AI: Future AI may need to embrace unreliable components, redundancy, and communication between layers with different competencies.
Cancer and the robots: “White robots don’t get Cancer.”. The AI itself can “go off and go do things”.
Giving up control and modularity: in Bioology the body wants easy controls of modules to act on, BUT also, an opposing drive makes things too “easy to control” and then you become easily ‘hackable’, from diseases (etc).
There is no way for us to reach an “*alignement*”, becuase *we (humans)* have and display fundamental disagrement across time and location; a “a single alignment goal, “doesnt exit for humans””
Robustness Through Noise: Biological systems thrive despite noise and defects; AI might learn from this resilience (e.g., Dropout in neural networks is a step).
Self modeling helps (be *more simple and easy*). Helps “robusteness,” “generailizations” “other” benefits, when you have more jobs (“multitasking”), and that helps you stay relevant despite small noices or small changes (“augemntation”)
Alignment Challenges: The concept of “alignment” is problematic due to inherent disagreements within humans and between different groups. Moving towards AI systems *more organic, and more integreated*.
Future of Digital Life, Agency, and Goals
Digital Life: Exploring the potential (and risks) of digital life and an “ecology of AI” is crucial.
The role of AI tools: to recognize communication/relationship with diverse forms of intelligences all around. Prosthetics, recognition.
Prosthetics and Outsourcing: Humans have *already* outsourced significant aspects of life, and AI will continue this trend, raising ethical questions about what it means for how our lives are defined by, versus technology (will power, relatioship guiders..).
This already exists! “tothbrush, education,” are outsourced. “The idea of having giving up our core competencices [..] this alreayd happened”.
Painting positive futures: We have to get clearer, as a group. Not juts a negative-voiding goals (current “ai safty”): “Everybody doing. Everyone specify.. “we are closter to to achieve.”
Agency and Goals: Understanding the origins and management of goals in collective systems is a fundamental and possibly existential challenge.
Multiscale Competency and Bioelectric Networks
Bodies are multiscale competency architectures, with problem-solving intelligence at every level (molecular networks, organs, swarms).
Definitive regenerative medicine requires communicating anatomical goals to cells in “morphospace” (the space of possible forms).
Cells us a native language called morphoceuticals.
Endogenous bioelectric networks are a tractable interface for top-down control of cell behavior. Tools can read and write pattern memories in this “protocognitive medium.”
Agential Material and Engineering Approaches
Traditional engineering uses passive materials; regenerative medicine works with *agential* material (cells with their own goals). This is like building with dogs instead of Legos.
Different problems require different levels of solution (e.g., orthopedic surgeon vs. psychoanalyst). Bodies exhibit a “spectrum of persuadability.”
Cellular collectives’ position on this spectrum is unknown; experiments, not assumptions, are needed.
Higher level processes can often supervene to fix the mistakes and defects found in lower levels of organisms, the ability to train will become useful here.
Scaling of Minds and Biological Plasticity
Biological development is a continuous scaling of minds, from single cells to complex cognition. No “magic” moment separates physics from mind.
Turing understood that body self-assembly and mind scaling are the same problem.
Anatomical Morphospace and the Anatomical Compiler
“Morphospace” is the multi-dimensional space of all possible shapes.
The genome encodes molecular *hardware*, not the final body plan. Cells “know” what to build and when to stop.
The “anatomical compiler” (long-term goal) would translate a desired form into stimuli to guide cells, not micromanage cell placement.
Current limitations: We cannot predict morphology from genome sequences (e.g., frog-axolotl chimeras). We are at the “hardware” stage of biological programming (like computer science in the 1940s-50s).
Biological Software and Collective Intelligence
“Biological software” refers to the intelligence that can be exploited, similar to software on reprogrammable hardware.
Intelligence (William James): The ability to reach the same goal by different means (a cybernetic definition).
Biological structures exhibit homeostasis which reduce error by correcting when target is deviated.
Pattern Memories and Bioelectric Control
Developmental biology is a goal-directed process (like a thermostat) moving towards a “target morphology.” This challenges traditional “bottom-up” emergence views.
Pattern memories are stored, not just in DNA, and can be rewritten, affecting the “target morphology” (e.g., two-headed planaria). This is analogous to brain scans (much scans).
Bioelectric networks (ion channels, gap junctions) are similar to neural networks, but operate in anatomical morphospace. Tools have been developed to read and write these patterns.
Early stage patterns form like “the electric face”.
Voltage patterns are *instructive*, not just disruptive. They can induce organ formation (e.g., ectopic eyes) and reveal hidden cell competency.
By inducing electrical patterns of potassium, gut cells were able to be converted to functioning eyes.
Applications and Future Directions
Cells exhibit properties which make them ideal for regeneration; They know where to build, how much to build, and when to stop building.
Bioelectric manipulations can be used for limb regeneration in frogs (non-regenerating species), and work is underway towards mammal applications (Morphoceuticals).
Immortal planaria provide key examples where a cell can divide and regenerate a new head, or two new heads depending on biometrical patterns.
Organisms have limits from their lineage.
Altering bioelectric circuits in planaria can change head number/shape, demonstrating reprogrammable anatomical memory.
Biometical networks can be rewired by manipulating existing patterns to create entirely new sturctures; oak leaves and their bio-engineers was cited as a notable exmaple.
Computational models of bioelectric patterns are being developed, enabling rational interventions (e.g., correcting brain defects in tadpoles even with genetic mutations).
Cancer can be seen as a failure mode of the scaling of goals, with cells reverting to individualistic behavior. Bioelectric reconnection can suppress tumor formation.
Future Medicine and Protocognitive Capacity
Anthropods made from humans’ lungs display capacities to assist nearby, hurt tissue, they have intelligent agency despite a very primitive function.
The capacity of human lungs goes beyond their usual purposes, highlighting the amazing intelligence available.
Future medicine will likely resemble psychiatry more than chemistry, exploiting the protocognitive capacity of tissues (using tools inspired by neuroscience).
Future interventions could involve “agential implants” (like anthrobots) and “morphoceuticals” targeting anatomical intelligence.
Future medicines and the treatment should evolve.
The speaker believes an apporach focused around sharping patterns found in memory is useful for anti-aging.
Research needs: In vivo voltage imaging, better ion channel drugs, better physics computational models, mapping voltage states in health/disease, and exploring non-electrical signaling (mitogenic radiation, etc.).
They believe high-level interventions through biometrical will allow the same rat training principle as rats: no micromanaging the parts (rat neurons, bio-organ electrical states), reward and punish to achieve a general behavior.
Introduction: The Journey from Physics to Mind
All life begins as simple physics (e.g., an oocyte – egg cell), gradually becoming complex, even achieving metacognition. This transition across the “Cartesian cut” is a core question.
Turing was also incredibly interested in biological development/creating shapes. Levin says that his work is linking/the-same-as intelligence and the bodes’ self-assembly.
All intelligences are *collective* intelligences, made of parts (cells, etc.). Even the human brain is a vast collection of interacting components. Single-celled organisms (like Lacrymaria) show impressive competence at small-scale goals.
Multi-Scale Competency and Biological Plasticity
Organisms have competence not just in 3D space, but also in other “spaces” like anatomical “morphospace” (the space of possible body shapes).
Caterpillar to Butterfly Transformation: Highlights drastic body/brain reorganization while retaining *some* memories, raising fundamental questions about identity and cognitive continuity.
Planarian Regeneration and Memory: Planaria can regenerate *any* body part, including the brain. Experiments show information storage *outside* the brain, and even transfer of this information to a *newly grown* brain.
Frog Eye Relocation: Tadpole eyes can be moved to the tail, and the tadpole can *still see*. This demonstrates incredible plasticity and adaptation, challenging assumptions about fixed developmental programs. The optic nerve connects to the spinal cord and not the brain in these tadpoles.
Multi-Scale Competency Architecture: Biological systems are nested hierarchies (like Russian dolls). Each layer solves problems in its own space (transcriptional, physiological, anatomical). Intelligence is the ability to reach a goal by *different* means (per William James), not just simple emergence.
Navigating Anatomical Morphospace
Where do complex anatomies (like a human torso) come from? DNA provides instructions for *cellular hardware* (proteins), but not the *software* that organizes cells into complex structures.
Picasso Tadpoles can organize into a “correct” face as a tadpole. When the cells that made up the Piccasso tadpole turn into a frog, the frog can now also find this new “correct” organization and will grow according to that new “correct” face.
Salamander Kidney Tubules: Cells adjust their behavior to create a correctly-sized lumen (tube opening), even if cell size is artificially altered. *One* giant cell can bend to form the lumen, showing top-down causation: a large-scale anatomical goal drives the selection of *different* molecular mechanisms.
The Brain as a Precedent: The brain maps high-level cognitive goals (in 3D space) onto molecular actions (muscle movement, etc.). Bioelectric networks *outside* the brain do something similar, controlling body configuration in *morphospace*. Evolution pivoted from spatial pattern control to temporal pattern control in neural processing.
Bioelectric Signals and Regenerative Medicine
Bioelectric networks: predating and analogous to the use-case of neural networks. It controls configuration and acts as a body-configuration throughout the body in its morphospace. Evolution was using Bioelectric Networks, way-way before, it created/pivoted to neural-network-focused cognition.
Tools to “Read” and “Write” Bioelectric Patterns: Inspired by neuroscience (optogenetics, active inference), these tools allow communication with cell collectives, influencing their “morphogenetic paths”.
Frog Leg Regeneration: A *single-day* treatment can trigger leg regeneration in frogs (which normally don’t regenerate limbs). This involves convincing cells to embark on a “build a leg” trajectory in morphospace.
Ectopic Eye Formation: Inducing a specific bioelectrical state can cause cells to build an eye *anywhere* on the body. This isn’t providing full eye-building instructions; it’s a “subroutine call” – “build an eye here.” The cells can even *recruit* neighboring cells to help.
Electrical map: is how bioelectrical states, that have configurations that represent memories, can show planeria how many heads to grow and other anatomical guidance for new cellular developments. The Genome defines the *hardware*, however.
Planarian Head Number Control: The bioelectric pattern determining head number can be *rewritten* using ion channel drugs (no external electric fields). This creates two-headed planaria, and this new “memory” (body plan) is *stable* through subsequent regenerations – a counterfactual, latent memory.
Xenobots: Exploring Morphogenetic Goals
Genome defines what the planarian’s new number of head would be *AFTER* a rewriting via influencing the electrical map.
Xenobots: Created from dissociated frog skin cells, these self-assemble and exhibit *novel* behaviors (movement, navigation, collective action), including *kinematic self-replication* (building new xenobots from loose cells).
There’s No “Xenobot” gene, rather, taking away the influence of neighboring cells helps uncover “Xenobots”. This behavior is described as “engineering by subtraction”: The normal, boring life the cells is dictated by its neighboring cells; isolating the cells reveal the default behaviour is being xenobots, and a completely new, never before seen, behaviour is observed.
Engineered by Subtraction: The “default” behavior of the isolated skin cells is to become xenobots. This reveals hidden morphogenetic potential. No straightforward evolutionary explanation: The evolution pressure didn’t select xenobots, but it made the material/machines, so that if “correct” influences are given to cells (via subtraction), the materials will develop “correctly”.
Kinematic self-replication: very minimum self-replication, as no real heredity between new ‘generations’, rather a rudimentary type of self-replicating robot, Von-Nuemann type of dream.
Implications and Conclusions
Biology lacks firm expectations: “you don’t know… how many cells.. what size.. what genetics..” This lack of assumptions leads to evolved material working well, and “doing something adaptive in a wide range of circumstances.”
Almost any combination of *evolved material*, *designed material*, and *software* can be some kind of *agent* (cyborgs, synthetic beings).
Future Ethics: Traditional criteria for moral consideration (“What are you made of?” and “How did you get here?”) will be insufficient. We need new ethical frameworks for interacting with diverse forms of intelligence.
The “spectrum of intelligence” is more interesting than a strict “living/non-living” distinction. Many systems will exhibit degrees of intelligence, blurring traditional boundaries.
Self-replication, like other biological properties, exists on a *continuum*. Xenobots represent a *minimal* form of self-replication, requiring provided materials and lacking strong heredity.
Introduction: Rethinking Discrete Categories
Traditional categorizations of living things (discrete species, clear separation between humans and other beings) are inadequate for understanding intelligence and future beings.
We need a framework for understanding, creating, and ethically relating to *diverse* intelligences, regardless of composition (what they’re made of) or origin (evolved, engineered, or a combination).
Focus should shift from categorizing natural and traditional “kinds” (types) to focusing on scales and intelligence.
The Continuum of Being and Intelligence
Evolution and developmental biology show a *continuum* of forms, not sharp distinctions between humans and other life stages (embryos, ancestors, or future augmented humans).
Horizontal Modification of beings are possible and very likely, which challenge these natural distinctions.
Humans will be modified (technologically, biologically) for health and augmentation. Distinguishing between “human” and “machine” becomes difficult.
We need a framework (like Rosenblueth & Bigelow’s scale) to relate to various intelligences: primates, birds, octopuses, colonial organisms, engineered life forms, AI (robotic or software), and even aliens. The scale is: Passive Matter, Computational Matter, Agent Materials and Metacoginition.
The goal of such framework is to create interaction (including creating new discoveires, and capabilites withing biomedicine and more), along with a much more sound etheical footing.
We all start as cells (“just physics”), eventually turning into things described by pscyology and even psychoanalysis.
We need a “story of scaling” to understand how systems described by physics also become describable by psychology.
Agential Material and Collective Intelligence
Unlike Legos, biological systems are made of “agential material” – cells with their own agendas and problem-solving capacities (e.g., *Lacremaria* single-cell organism).
All living systems, and even the “single” unified person is a *collective intelligence*.
We, people, *are all made of* parts. The body, especially even the pineal gland (in rene descarte times considered the singular organ), are collection of things with multiple things inside it.
Even our *selves* are “collective intelligences” composed of parts. The challenge is explaining how these parts create a unified sense of self.
Alan Turing saw a connection between the origin of bodies (morphogenesis) and the origin of minds.
The “Self” is Dynamic, Not Hardcoded
Counting “embryos” isn’t counting a fixed number of beings. It’s counting the alignment of cells committed to a shared anatomical plan.
Experiments with duck embryos show that the number of “selves” can change dynamically based on physiological processes (e.g., creating conjoined twins by separating groups of cells).
Selves construct themselves (including defining their own boundires from the envrionment); a “self” isn’t something predefined genetically.
Split-brain patients and dissociative identity disorders also show that the number of “selves” in a brain isn’t fixed.
Cognitive Glue (e.g. nerovus system): Creates a higher level entity, with cognitive abilities far exceding each single part.
Radical Plasticity: Caterpillars and Planaria
Caterpillar to Butterfly: Caterpillar (simple, 2D movement) is reorgnaized drastically to turn to Butterflies (flies, has hard parts, eat different things). Caterpillars, if trained, have memory. Butterfly memory persist, and what is useful for the caterpillar is translated into actions (of flight, eatings etc) for the butterfly, demonstrating profound adaptation beyond mere memory.
Planaria Memory and Regeneration: Planaria remember training even after their heads (including brains) are removed and regrown, implying memory storage outside the brain and transfer to a new brain.
Multiscale Competency Architecture
Biological systems have problem-solving abilities at multiple scales (molecules, cells, tissues, organs), not just in three-dimensional space but also in gene expression spaces, physiological spaces, and “anatomical morphospace.”
Beyond Reliability: Salamanders and Picasso Frogs
Salamanders (high ability for regenration) with varying numbers of chromosomes and cell sizes *still* build the correct structures, demonstrating robust adaptation to unexpected variations, even in the number of cells!
Shows evolution produces problemsolving capabalities rather than rigid reliability.
Evolution makes a “error minimization” and is not simply fixed instructions: Picasso frogs (organs scrambled) still develop into normal frogs, indicating a system for *error minimization,* not just following hardwired instructions. Organs do unusual pathways to correct itself.
Beyond Genetics: Morphogenetic Plasticity
Example 1: Flies run Ant-Morphogenetic programs on wings: it protects from predator. This illustrates morphogenetic (change in shape/structure) potential.
Example 2: Wasp on Oak-leaves. The typical oak and acorn, are well understood (shape and everything). But Wasp makes signals, causing structures on oaks very different to the ones typically observed.
Genetics defines the *hardware,* but cells can achieve diverse outcomes. A wasp can induce an oak leaf to build a *different* structure (a gall) without changing the leaf’s DNA.
Communicating with Morphogenetic Intelligence
Morphogenesis is the behavior of a collective intelligence in anatomical space. We need to learn how to *communicate* with this intelligence.
Example: The wasp above “communicates” (through evolution) to tell cells of a different kind to construct different.
Neurons/Nervous system uses computation and electricy. Evolution had already used and discovered the use of an electrial network to intergrate many cells both space and time way before it developed muscles.
Voltage-Sensitive Fluorescent Dye Imaging: A technology that lets us measure voltage. This allows viewing “bioelectric pattern” showing the frog forming their faces, and learning to understand/decode the pattern/process.
Electrical Pattern (“Electric Face”). Like brain scanning, shows activity but in organs: showing you can tell where a organ like a mouth/eyes is going to be way in advance.
Pathalogical patterns exist and could signal issue. For example: cancers have abnormal patterns and disconnect from surrounding envrionment/cells and causing cells to “forget” larger structure.
This approach offers way to do cancer therpatueis; by trying to change cell memory/state instead of “killing cells” or toxic therapies, and “reconnecting” the cells.
Examples: Cells disconnected from environment will loose big structure, the memory can be restored through reinflucing the connection back, even with genetical defects still present.
Reprogramming the Body Plan
The group has created ectopic eyes (induced on frog’s body): telling (through electral signal/patterns) cells to become another. Almost any region could be reprogramed into different organs, like eye. Cells cooperate with neighbours when only few cells are injected.
Proving that “eye master genes” theory only partially accurate; once you know how to “communicate” in right language/patterns/signals, any cell has way more potential.
Rewriting Pattern Memory (Planarian). Can change worm head: from one-head into two heads *forever*, they always reproduce/regenerate as two-heads after this. By Modifying Biopattern. Showing: DNA don’t dictate 100% how bodies change, the same genome (or DNA, no change in genetics here!) can still build very different things depending on how/where and by what the signal is/what the patterns are! The reprogramed cell can exist beyond generations/cuttings.
Origin of Goals and Xenobots
Questions raised about Goals (Where these morphoentic and goal pattern exist outside DNA? The option and limit, including the example of re-routing worms’s into head of different species with wildly different genetics.): they aren’t completely set, but re-programmable, a “rewritable”.
Just like reprograming computer instead of using “soldering iron”, the biology could and does re-program to fix and correct for goals.
Xenobots (created from frog skin cells): These cells, when separated from the embryo, *spontaneously* self-assemble into new organisms (“xenobots”) with novel behaviors, including *kinematic self-replication* (building copies of themselves from loose cells).
Xenobots have Cilia (hairs for frog mucas transport), in Xenobot use it to swim.
Kinematic Self-replication (discovered by accident!) – creating generation of copies of Xebot with materials/parts on envrionment.
Showcasing it is difficult to attribute Xenobot shape/behavours, as they don’t have any selective “evolution”.
Evoluation create problemsolvers, rather than fixated behaviors/forms, that is, under condition changes/challenges.
Anthrobots (from adult human tracheal cells): Similar to xenobots, these human cells also self-assemble into novel structures (“anthrobots”) with surprising capabilities, like repairing neural wounds.
The Spectrum of Persuadability and Ethics
It’s wrong to look at “philosophically” into beings and deciding which one is an intelligence. There is experimental way of looking at this (“spectrum of persuadability”, from tools/hardwiring -> behavorial science and trainings -> Rich relationship)
and intelligence.
“Persuadability”: The kind of tools we use to communicate with things, scales.
Cells are capable of things than what people thought, and it is easy to miss intelligence, making us “a lot left on table”.
We are on the era of “diversity intelligence”. Where today’s Large Language Models (GPT-4) don’t matter,
What Matters Is that: We need to do experienments, there are non-human minds, even some of the simplest thing has incredible complexity.
Ethics: We must avoid denying moral worth to beings because they don’t look like us. The space of possible bodies and minds is expanding rapidly (cyborgs, chimeras, etc.), and we need better ethical frameworks. It’s about learning to relate to *different* beings.
There will exist cyborgs and all types, including biological materials/evolved ones combined, making us “Synbiosis”, or beings living together and requiring a “ethical Framework”.
It won’t just about whether things are “metal/human” that is traditional, or looking to the tree-of-life that decides on which being “count” – many, multiple things count, many will change. We are required to make moral frameworks and ethics in way never before required.
Humanity on the Long Term: We are not the “best, and most developed”. Humans, what does this actually mean (to want Roomba, companion?).
What it wants is *NOT* DNA, what does that mean (relationship etc)
What matters?
(That are Worth-Thinking!)
Introduction and Background
Michael Levin’s early interest in how things are built (sparked by a TV) led him to explore both engineering/computer science and biology. He was interested in the way that both physical machines could create images on a screen and insects could perform behaviors.
He became fascinated by how minds emerge from physics and chemistry, driving his focus on developmental biology.
Levin’s perspective: standard developmental biology courses may offer a different, potentially more gene-centric view than his.
Genes, Hardware, and Software
Genomes primarily describe the *hardware* of cells (protein sequences), not the overall organismal form (symmetry, number of limbs, etc.).
Biological systems have both a hardware layer (genetics) and a crucial *software* layer (developmental physiology and decision-making processes). The software is dynamic and critical for navigating complex development.
The mapping from genotype (genetic information) to phenotype (observable traits) is complex, not a simple one-to-one relationship, except in cases such as specific enzyme production.
“Froggle” example: Mixing frog and axolotl cells creates a creature whose leg development can’t be predicted solely from the genomes, demonstrating the importance of cellular decision-making (software).
Evolution selects for phenotypes (final outcomes, like anatomy and behavior) yet it’s genome (raw information) that get’s past on between generations, illustrating how interconnected and codependent each system.
Explaining Emergent Properties
Explaining biological features requires considering both genetic factors and “free lunch” properties arising from physics, geometry, and computation.
Galton board example: The bell curve shape arises from the system’s setup, not inherent in the wood, nails, or marbles individually. Emergent property from simple components and organization.
Transistor example: Connecting transistors creates logic gates with truth tables that are inherent properties of the configuration, not something separately evolved.
Biological systems can harness “generic laws” (as discussed by Stuart Kauffman and Stuart Newman) that are not explicitly encoded in the genes.
Cellular Decision-Making and Bioelectricity
Cells and cell groups have agency (preferences, behaviors) and make decisions. Evolution shapes these behaviors through signals between cells.
Electrical networks, formed by connecting cells via gap junctions, are crucial for scaling up goals and collective intelligence.
Gap Junctions allow a collective’s identity to fuzz out as molecules such as calcium merge together making multiple cells share a connected physiological signal. From a cells point of view it is a ‘false’ memory, from the collectie, it is real, causing a Mind-Meld, where memory is not longer ‘owned’.
Homeostasis comparison: Single cells measure, remember, and act on their *local* environment. Connected cells in electrical networks measure, remember, and act on a *larger, non-local* scale, facilitating collective goals and response.
Scaling of stress: Cells communicate stress (deviation from desired state) through shared signals, promoting plasticity and coordinated action to achieve collective goals. Individual stress becomes a shared problem.
Cancer as a Breakdown of Bioelectric Communication
Bioelectrical signals create a larger sense of “self” across cell groups.
Cancer can arise from cells becoming electrically disconnected from the larger network. They revert to their ancient, unicellular goal: to divide and go where resources are good (metastasis).
Cells becoming disconnected (becoming cancerous) isn’t from increasing ‘selfishness’, rather it is a consequence of it shrinking, going back to their more selfish individual states, acting individually.
Oncogenes often shut down gap junctions, the very first step towards bioelectrical disconnection, isolating cancer cells from the larger collective control.
The study of bacterial biofilms shows that brain-like behaviors (using ion channels) evolved long before nervous systems, indicating the ancient roots of bioelectric communication.
Morphespace: where good regions/bad regions, like barriers/obstacles are navigated similar to normal 3 dimensional space except now we are in a world of configuration.
Memory and Bioelectric Circuits
Bioelectrical networks literally store a kind of memory, representing the “set point” or target morphology for regeneration.
Planarian example: Cutting a flatworm into pieces results in each piece regenerating a complete, proportional worm.
This implies that is must be a form of homeostatis going on; a complex non-neural form of a collective intelligence to keep such anatomical patterns consistent.
Bioelectric pattern can be visualized (with voltage-sensitive dyes) and *rewritten* (using ion channel drugs) to change the body plan (e.g., creating two-headed worms). No genetic changes are required.
That two-headed worms consistently produce two-headed worms is evidence of a true memory.
The altered body plan is heritable through *fission* (splitting), demonstrating non-genetic inheritance. The bioelectric circuit acts as an additional hereditary medium. Not all inherited traits are DNA-based.
Bioelectric circuit in planaria example; if the system’s physiology first ‘boots’ the bioelectric default circuit of number of head. This number-of-heads can also be edited non-genetically by altering the signals for short-term using inputs (tapping buttons on a calculator), creating a two-headed work without ‘rewriting’ the program.
The Eye Experiments
Early frog embryos show a bioelectric pre-pattern that predicts the location of facial organs, including the eyes.
Injecting ion channel RNA into other areas of the embryo (e.g., tail, gut) induces the formation of ectopic (out-of-place) eyes, demonstrating that bioelectricity is *instructive* for organ formation, and it is not merely for house-keeping.
This revealed the modularity of development: The researchers didn’t need to specify *how* to build an eye, only *where*. The cells organized the complex process themselves.
Cells recruited neighboring cells (even those not directly affected by the injected RNA) to participate in eye formation, showcasing multiple levels of instruction.
Cells outside the traditionally defined “competent” regions (anterior neurectoderm) can, in fact, form eyes, highlighting the limitations of gene-centric views. It is not the top of the hierarchy, voltage is!
Pac6: normally makes eyes, is found at anterior norectoderm, which will define competency for creating an eye, yet, other parts can.
Ectopic eyes connected to the spinal cord (not the brain) could still mediate vision, demonstrating remarkable plasticity of the nervous system and its ability to interpret novel inputs.
These examples were made early in Levin’s carreer, predating later research.
Neutral Mutations; deleterious mutatuions, once lethal or severe can instead turn to be less sever or become neutural, broadening the landscape where evolution can progress towards.
Multiple Levels of Control and Goal-Directedness
Biological systems exhibit multiple levels of emergence and control.
Choosing the *right* level for intervention (e.g., bioelectric pattern vs. gene expression) is crucial for effective manipulation. Bioelectrical manipulation is often more effective and efficient than trying to micromanage genes.
Instead of dealing with 10s of 1000s of individual parts/mechanisms, one could tap in further up the decision chain and deal with an intelligent system that navigates the complex decisions, taking away the stress of needing to handle the details.
Biologists often exhibit “teleophobia” (fear of attributing goals or agency to biological systems), but cybernetics provides a framework for understanding machines with goals as a *continuum*, not a binary (dumb vs. smart). This is no longer considered “magical.”
Telephobia came from needing to study all other entities as clocks since early days did not know how to interpret a human’s inner thoughts.
Agency claims are *engineering* claims, testable by experiment (e.g., identifying, reading, and rewriting set points in a homeostatic system).
Thermostat Example: how to test? look at what its level: is, is a setpoint?, 2 can we read/decode setpoint, rewrite? to rewrite is a new rewiring needed (like mechanical clock).
Thermostat continued: after test; if thermostat work as expect, a trust/enginner dependency. No micromanaging is needed, temperature managed; good!.
Genetic pathways can be *trained* (similar to neural networks), exhibiting various types of learning (including Pavlovian conditioning). This challenges the idea of purely deterministic gene regulatory networks.
Implications for associative-learning, in a petri-dish there exists too powerful of a drug yet we cannot apply to humans, give both drugs, then give only nuetural one later; may or may not work.
The molecular-placebo will activate only once paired enough times (Pavlov-style), where if the pairing stops, the original reaction (dog will drool from the bell ringing sound).
Placebo, Intention, and the Mind-Body Connection
Intention can influence the body at multiple levels, including the bioelectric state of cells (e.g., deciding to stand up changes muscle cell voltage). The mind-body connection is strong and demonstrable.
This suggests using levels of the hierarchy (chain of command) and speaking at the relevant part of the system to get tasks done.
There is evidence of non-verbal “selves” within our bodies and we might be parts of larger selves.
Learning is change withing one agent (human/machine learning, yet training (neural networks) implies multiple parties involved, at least one to create pressure (or lack-of pressure) and on receiver to change. The distinction could be helpful for exploring further discoveries.
Minds emerge gradually during development. There’s no sharp dividing line between “just chemistry” and “having a mind.” This implies that minds exist in various forms across different scales of biological organization.
We may be bad at recognizing unconventional intelligences because our perception is biased by our experience of the three-dimensional world of medium-sized objects moving at medium speeds. There are intelligences operating in other spaces (e.g., physiological space).
Ethical questions arise from recognizing diverse forms of intelligence.
The Nature of Self and Identity
A self is a collection of parts working together towards *system-level* goals (goals of the collection, not individual parts).
Selves can be compared by the size and scope of their goals (“cognitive light cone”). A bacterium’s goals are small and local; a human’s can encompass larger spatial and temporal scales. Selves are nested, and humans may not be at the top of the hierarchy.
We may be unable to fully understand the goals of a larger system of which we are a part (analogous to ants being unaware of the context of human actions).
Mathematical formalisms might provide evidence for or against being part of a larger system.
“A self is a *temporary* bundle of activities that work toward specific goals”: highlighting the agency nature that systems perform in different and potentially multiple cognitive landscapes.
There’s a distinction between *learning* (changing your mind with the assumption of no external agency) and *being trained* (being changed by an external agency). It is an open, empirical question whether the external world has agency.
Consciousness itself: is a hierarchy where “the more indeterminism (the space between what you could/cannot) is an indicator of the more level of agency that this form has”.
Scientific Inquiry and Open Questions
Many phenomena should be treated as empirical questions rather than philosophical beliefs.
Scientific frameworks both enable and constrain the kinds of questions we ask and experiments we perform. Our pre-conceptions, including being human-centric affects or decisions of intelligence (a dog being called intelligence as it is closer to our level).
All science begins with an act of faith: the assumption that the world is understandable and that there are patterns to be discovered. It is important to be aware of this foundational belief.
Current focus of Levin’s group: understanding different kinds of minds in various embodiments, with implications for regenerative medicine, birth defects, cancer, synthetic biology, and artificial intelligence.
Introduction and Concepts
Levin discusses a framework for understanding intelligence that applies to diverse agents (biological, artificial, exobiological, etc.).
Intelligence is competency in navigating various spaces (physical, transcriptional, anatomical, physiological, etc.).
Agents possess goals, preferring certain states within a space, and have varying capacities to achieve those goals.
William James’ definition of intelligence: Ability to reach the same goal by different means, emphasizing adaptability and resilience.
Teleophobia: Resistance to the idea of goals in nature. Levin argues against this, referencing cybernetics, control theory, and computer science as examples of non-magical, goal-directed systems.
The “proof is in the pudding”: Focus on empirical results. Levin’s framework leads to experiments others wouldn’t do, like regenerating frog limbs.
Levin cites his definition as: this is useful, and welcomes superior frameworks and results, referencing also yakir aharnav’s two-time interpretation.
Multi-Scale Competency and Scaling
Multi-scale competency architecture: Levin proposes a continuum of cognition, scaling from basic physics (least action principles, quantum indeterminacy) to complex organisms.
Scaling involves increasing the size of an agent’s “cognitive light cone,” representing the spatial and temporal extent of the goals it can pursue.
Emergent minds: Collective intelligences form when individual agents (like cells) connect via gap junctions/electrical synapses, sharing memory traces and becoming functionally entangled.
The system grows on three sections: the ability of measure, larger area and a larger space to remember, so when we take measurements we take large scale measurements, and bigger memories/goals.
Stress is the difference between the current state and the goal state. The things that stress a system reveal its cognitive sophistication.
Cells both compete, cooperate, all dictated at least partially, by an ultimate solution.
Optimality, Goals, and the Future
Where do goals come from? Beyond rational design and evolution, novel agents (like chimeras or self-assembling systems) exhibit emergent goals whose origins are not fully understood.
We lack a science to predict goals and capacities of emergent minds. This is crucial for understanding and interacting with increasingly complex systems.
Levin connects and is inspired, partially, to and by least action principles.
Suggests possible connection of “optimality” to Aharonov’s two-time interpretation of quantum mechanics, where the future may have a causal impact on the present.
Artificial Intelligence
Rejects the term “artificial intelligence” as creating a false dichotomy. There isn’t “real” intelligence vs. “artificial” intelligence. Chimera combinations (biological/technological) blur these lines.
Argues against the idea that evolutionarily-derived intelligence is inherently superior. Engineers can potentially create intelligences exceeding those found in nature.
The “proof is in the pudding”: Focus on empirical results. Levin says there is a symmetry between nature and engineering.
Current AI (machine learning) is missing key aspects of true cognitive agency, but Levin is optimistic about future progress.
Social and Political Implications
Discussions around societal impacts of future technology (AI, genetic engineering, etc.).
Levin cautions against over-regulation. He believes restricting advancements in tech by those capable of such, is a slippery slope.
Expresses concern about top-down attempts to enforce uniformity or limit individual expression.
Advocates for freedom and allowing individuals to pursue their potential, even if it involves radical technological augmentation.
Evolutionary success is not a basis for morality. Levin thinks optimization and guiding are the best practices and can yield improved success and greater freedom, even to some individual losses.
Acknowledges potential downsides (inequality, harmful choices) but favors a libertarian approach, emphasizing freedom of choice and adaptation.
No guarantee that collective interests always aligns with individual well-being, referring to climbing example and its cost to skincells as analogy.
Suggests studying scaling principles in biology to inform how we design social structures, promoting a balance between individual welfare and collective goals.
Highlights the company that uses bio-electrical signals (limb regen and similar), and bio-synthetic AI to further discover the secrets and unlock applications that benefit from knowledge.
For information see his website.
Introduction
The discussion revolves around embodied minds, cognitive agents, and the blurring lines between biology and AI.
Professor Levin’s lab studies intelligence in unusual substrates, including cellular collective intelligence during development and regeneration.
The show covers Defiance of all Binaries (Life vs Machines), and latent surpising capabilities within things.
Levin’s Key Projects & Experiments
Development of tools to read and write electrical memories of non-neural cells, revealing how cells store information about body plan (e.g., number of heads in flatworms).
Rewriting electrical memories in flatworms to create permanently two-headed worms, demonstrating non-genetic inheritance of body plan.
Creating tadpoles with eyes on their tails that can see, showcasing the plasticity of biological systems and their ability to adapt to novel sensor arrangements.
Detecting and normalizing cancer by controlling bioelectrical connections between cells.
Creating Xenobots and Anthrobots: Demonstrating that cells (frog and human, respectively) can self-organize into novel structures and exhibit unexpected behaviors when placed in new environments. This highlights latent capabilities.
Levin’s lab has created molecular tools and workflows which are able to read/write electrical activity. These reading are voltage states from cells which indicate body configuration/regeneration outcomes, not neural patterns of information storage.
The Nature of “Control” in Biological Systems
Techniques range from applying drugs which affect the intercellular conversations electrically and applying “voltage sensitive fluorescent dyes” that are mapped by microscopy.
Interventions often involve modifying the bioelectrical communication between cells, acting at a higher level than direct genetic manipulation. It’s about influencing the cells’ “decisions,” not micromanaging individual genes.
Computational models are used to simulate electrical patterns in tissues and predict the effects of interventions. This is analogous to “activation patching” in AI interpretability research.
The goal is to trigger high-level processes (like limb regeneration) with minimal intervention, relying on the system’s inherent capacity to self-organize. For example 24 hours for 18 months of growth in his experiements.
Interventions can trigger different outcomes depending on the context (e.g., the same intervention can trigger leg regeneration in adult frogs or tail regeneration in tadpoles), showing the importance of the system’s existing knowledge.
It can be useful to know what an intervention does to allow for biological configurations to alter as such interventions are introduced. Some Interventions don’t do that, while others might
Emergent Agency & Unexpected Capabilities
Even simple systems (like sorting algorithms) can exhibit surprising capabilities not explicitly programmed into them (e.g., delayed gratification, clustering by algorithm type).
This challenges our intuition about what to expect from complex systems, both biological and artificial. We have poor intuition for emergent agency.
Biological systems (including sorting algorithms in new study) have surprising agency which we have not known about previously.
Levin’s view is that emergence is subjective; It’s is about the *observer’s* surprise, not an objective property of the system. If you predict it, it is not considered emerging.
The Blurring Lines Between Life and Machine
Terms like “machine,” “human,” “alive,” “emergent” are engineering protocol claims, *not* objective truths. They represent useful *models* from a particular perspective.
If the claim or “mirage” holds use from some angle, it may be useful, else toss it, don’t force its definition upon an understanding of something.
Binary distinctions (e.g., living vs. non-living) are not useful and are collapsing. Orthopedic surgery relies on a “machine” view of the body, while psychotherapy requires a different perspective.
Both simple computer and complex living bio-organisms break “usual binaries”.
The level of cognition, not “being alive,” is the interesting question. Cognition exists on a spectrum, not a binary.
Current categories and binary distinctions don’t exist in an object or in their reality as an idea..
Biological insights should inform research around AI such that AI’s and biological systems should be considered with equal footing for consideration with the idea there is some blurring between living vs non-living things, and our current models, framework and understanings of the blurring don’t fully incorporate these considerations yet, including things like law.
Scales of Intelligence and Subjective Experience
Intelligence is about solving problems *in some space* (anatomical, chemical, behavioral, linguistic, etc.). Embodiment can exist in any of these spaces, not just 3D physical space.
Scales of Intelligence may relate with each other but be useful at one single scale.
Agency may have something of an underlying requirement with subjective experiences or experience in general.
The “cognitive light cone” represents the size of the largest goals a system can pursue.
It’s crucial to experiment (perturb the system) to discover goals and capabilities, not just observe.
Because this concept is continuous instead of Discrete (digital).
All living systems scale, the idea here, that particles may express an aspect of least-action within quantum interactions.
Even particles *might* exhibit minimal agency (goal-directedness and some degree of unpredictability from local conditions), according to least action principles and quantum indeterminacy.
“Life” might be defined as those systems that effectively *scale up* agency, indeterminacy, and goal-directedness across multiple levels of organization.
This could imply life could have an ability to understand and reframe something in a subjective manner.
An “inner perspective” emerges when you need to consider a system’s own view of the world to predict its behavior. This isn’t binary but a matter of degree.
Implications for AI
We lack principled frameworks for understanding the goals of novel systems and interacting with radically different minds (both biological and artificial). This is an existential risk.
AI systems *don’t* need to be human-level (or have large cognitive light cones) to be dangerous. Our brittle physical and mental frameworks are a significant vulnerability.
Humans confabulate, lack grounding for much of our knowledge, and struggle to extend compassion to those different from us. This raises serious ethical concerns about AI.
AI will likely need to make similar models and framework the same ways biology has and could use to interpret things.
It’s erroneous to assume that current AI systems have *no* degree of goal-directedness, simply because they aren’t like “elite adult humans.” We may be failing to recognize simpler forms of agency.
The capacity for surprise in AI may lie in the realm of “emergent goals”.
AI lacks robust memory and is not robust and exhibits problems in generalization which may get inspired and researched with bioelectrical studies such as in Michael’s Lab.
AI’s that become more general and start exhibiting qualities similar to living organisms and agency which we normally attribute moral worth, could mean this class would grow immensely to trillion’s and it’s an important question on how this should develop given considerations for an agent/subjective being’s future wellbeing, if that is to happen at all.
Evolution makes problem solvers not optimized specific solution generators.
Recommendations for AI Researchers (Indirect)
Consider principles from diverse intelligence research, exploring the spectrum of cognition beyond just the “standard adult human” model.
Recognize that “embodiment” can occur in many spaces beyond 3D, including abstract spaces relevant to AI.
Prioritize experimental perturbation over philosophical commitments when investigating capabilities.
Develop a principled science of where novel goals come from and how to ethically interact with radically different minds.
Avoid binary thinking; understand intelligence as existing on scales and gradients, with an ability to comprehend things under subjective modes.
Distinguish between AI tools (designed for specific purposes) and true agents (with open-ended intelligence and moral worth). Levin is consciously avoiding research that could lead to the latter.
Introduction: Definitive Regenerative Medicine
Levin’s goal: “Definitive Regenerative Medicine” – controlling what groups of cells build to solve birth defects, injuries, cancer, aging, and degenerative disease. Not just late-life interventions, but continuous rebuilding of structures.
The “Anatomical Compiler”: A future system to specify a desired anatomical structure (e.g., a three-headed flatworm) and generate the stimuli to get cells to build it. Not a 3D printer, but a communication device.
Focus: Understanding the cooperative action of cell collectives to build organs, both healthy and diseased. Moving beyond focusing solely on the genome, acknowledging cells possess inherent problem-solving abilities, similar to how individual cells such as *lacrymaria* have inherent survival capabilities without complex nervous systems, hinting at a pre-existing form of ‘cellular-scale decision-making’.
The Problem of Morphogenesis
Where does body anatomy come from? Not directly from DNA, which specifies proteins (micro-level hardware). The question is how the collective activity of cells with this hardware builds the correct, species-specific “target morphology.”
Fundamental knowledge gaps: Even with sequenced genomes (e.g., Axolotl, frog), we can’t predict anatomical outcomes of cell mixtures (e.g., “froglottl” legs). This is a collective intelligence question, not a hardware question.
Need to go beyond molecular medicine: Current molecular medicine is like computer science in the 1940s/50s – focused on hardware (genes, proteins). We need a higher-level interface (like software) to reprogram cell behavior.
Multi-scale competency architecture: Life has problem-solving at multiple levels (molecules, cells, tissues, organs, organisms). Each level has competencies in its own space (physiological, transcriptional, anatomical, behavioral). Examples of biological adaptability include *tadpoles developing eyes on their tails* connected via an optic nerve, enabling them to respond to visual cues, even without a direct brain connection.
Planaria: A Key Model System
Planaria: Flatworms with remarkable properties: Regeneration (any body part), cancer resistance, and immortality (no aging). They constantly renew their tissues.
The *regeneration record* includes successful regrowth after being cut into 275 pieces.
“Messy” genome: Planaria have a highly mutated genome (accumulates somatic mutations) yet maintain perfect anatomical control. This challenges the idea that the genome is the primary determinant of form.
Memory: Planaria can be trained, and this memory is retained even after head regeneration. This implies memory distribution outside the brain and imprinting on new tissue. This has implications for human brain regenerative therapies, which introduce new cells replacing decades-old memory/personality patterns.
Bioelectricity: The Software of Life
Analogy to the nervous system: Neurons use electrical signals (via ion channels and gap junctions) for decision-making and moving the body through 3D space. This is “software” running on cellular “hardware.”
Bioelectric signaling is ancient: It existed long before nerves and muscles. It’s used by *all* cells (not just neurons) to control body configuration through *anatomical morphospace*.
Tools to read and write bioelectric patterns: Voltage-sensitive dyes visualize electrical conversations between cells. Techniques to manipulate ion channels and gap junctions allow us to rewrite these patterns.
Bioelectric Control of Morphogenesis
Instructive bioelectric patterns: Changing the bioelectric pattern (e.g., inducing an “eye” pattern) can instruct cells to build specific organs (even in abnormal locations, like eyes on the tail). Demonstrating the adaptability of these biological systems when provided the correct stimuli.
Modularity: The bioelectric code specifies *organs*, not individual cells. The cells themselves handle the complexity of building the organ (subroutine call analogy). Illustrates how tissues exhibit an inherent capacity to self-organize when triggered with the right set of commands.
Recruitment: Cells with altered bioelectric states can recruit neighboring cells to participate in building the structure. Revealing a cooperative intelligence akin to the social behaviors found in insect colonies, such as those of ants and termites.
Applications: Creating extra forebrains, legs, hearts, inner ears, fins (even in tadpoles, which don’t normally have them). Expanding regenerative capabilities in ways that exceed natural limitations.
Limb Regeneration and Company Founding: Bioelectrical signals change rapidly after amputation. Frog Leg Regeneration is stimulated with a bioelectrical intervention, and it took only 24 hours of it. Levin Co-founded, with Dave Kaplan, a Company (Morpheuticals) which attempts Limb Regneration on mammals, mice currently.
Planarian Pattern Memories
Rewriting body plan: A bioelectric circuit in planaria stores information about how many heads to have. This pattern can be rewritten (e.g., to create two-headed worms), and this change is *stable* (like memory).
Single flatworm: A single flatworm may carry more than one instruction on its structure.
Not a map of the current state: The bioelectric pattern is a “counterfactual memory” – it represents what the animal *will* build if injured, even if it looks normal now.
Head shape: Bioelectric signals also control head shape. Blocking electrical communication can cause planaria to regenerate heads of different species (exploring “morphospace”).
Latent ability: Exploring non-standard forms are within cell capacity.
Applications and Future Directions
Computational Platform: Create full-stack bioelectric stimulations of what will happen with genetic/cellular info, so it may tell which ion channels will need open/close.
Repairing brain damage: A computational model predicts which ion channels to manipulate to restore a normal bioelectric pattern in damaged frog brains (even with severe genetic defects). Using the bioelectrical signal restoration and the administration of approved anti-epileptics to stimulate neural repair and to recover not only brain structure, but cognitive function.
Bioelectricity in Human Channelopathies.
Cancer as a failure of collective control: Cancer cells disconnect from the bioelectric network and revert to a unicellular lifestyle. Forcing cells to remain electrically connected can suppress tumor formation even with oncogenes present.
Physiological software layer: A tractable target for biomedicine, between genotype and anatomy. Cracking the bioelectric code (like neural decoding) will reveal how cell networks make decisions.
AI tools: For designing specific strategies for regenerative medicine.
Bioelectrical signaling: A “cognitive glue” binding cells towards a larger purpose (maintaining the organism).
Bottom-up (conventional/hardware) and top-down (software) treatment strategies.
Words and drugs having the same mechanism of action, quoting Fabrizio Benedetti: Bioelectricity provides communication.
Q&A Highlights
Ion channel distribution: Complex patterns can arise even with uniform ion channel distribution (self-organization).
Spatial Specificity and Signals in Wounds
Neural Cellular Automata (NCA) collaboration: Acknowledged collaboration on “distal.pub” paper.
Yamanaka factors: Important, but don’t address large-scale morphogenetic problems. Undifferentiated cells alone are not enough.
Genetics vs. Physiology: Both are important (hardware and software). Physiology can override genetics in some cases.
Aging solutions: Will likely fall out of solving morphogenetic control in general (along with regeneration, cancer reprogramming).
Interface: genetics and quantum biology: Unknown currently; Classical is good so far.
Money/Resource Limitation: Is where progress is stuck, not Fundamental Problems.
Mapping bioelectric signal and gene expression, and to body/organ changes.
Optogenetics use: Used in research, but clinical applications may be limited due to the need for gene therapy.
Drug development for ion channels: Lots of activity, but the main bottleneck is a lack of physiomic data (bioelectrical state data in health and disease).
Earth’s magnetic field: Not a major factor in the types of strong electrical exchanges studied.
Realism and the Moon (Interface Theory)
Question: If the moon isn’t “rendered” when unobserved, how can we predict its future position?
Analogy: VR multiplayer game (Grand Theft Auto). A red Ferrari exists only when rendered, but the *supercomputer* (underlying reality) has lawful processes.
We can use our space-time intuitions (Newtonian physics) to predict, but this is a “useful way to do things,” not necessarily *truth*. The VR analogy highlights this.
Predictive models work due to underlying lawful processes, even if the *representation* (moon, Ferrari) is not “real” when unobserved.
Living vs. Non-Living Distinction
Hoffman disagrees with Bernardo Kastrup (partially): No fundamental distinction between living and non-living.
The distinction is an “artifact” of our limited “dashboard” (interface). Useful within the “game,” but not a deep, principled distinction.
This follows from seriously considering idealism and interface theory.
Scientific Progress: Closer to Reality?
Question: Does moving from Newtonian to relativistic physics get us “closer” to the underlying reality?
Scientific theories always start with *assumptions* (miracles). Deeper theories explain prior assumptions, but have *new* assumptions, ad infinitum.
Science is always *infinitely far* from a “Theory of Everything.” It’s like a child asking “why, why, why?” – there’s always a deeper level.
Space-time itself is an assumption, and physicists (“SpaceTime is doomed”) are moving beyond it.
Science provides increasingly useful *descriptions of a perspective*, not closer approximations to ultimate truth. An infinite number of persepctives exist.
Analogy. Humbling view: our 4D spacetime is a “trivial headset”, perhaps the dumbest out of a potentially more accurate “headset” with ,000s of dimentions.
Better theories are more useful *within our perspective*, not necessarily closer to “ultimate truth.”
Assessing Theories. Arham Razor used to access a scientific theory. One assesses theories based on fewest “miracles”. Ideally, we should not need any miracle to justify, but any thoery will.
The assumption we take within the perspective can differ as different assumptions are used within the same thing, there is not specific rule or order to determine which theory is right or useful.
The “Space-Time perspective” might itself have multiple possible underlying theories. There is not a method to which to access theories with less miracles with another method of less miracles, there’s infinitely more perspectives, thus how many “miracles we use to explain predictions don’t tell us which onf those persective is “true”.
VR and Lawlike Behavior in a System
If modeling for a system where ferari’s do exist allows us to model the lawlike system behind, how does the perceived Space-Time affect modeling?
Child phychologist shows children wired by age of four months old for that.
Meta Considerations, Ultimate Reality, and Concepts
Analogy: is assessing theory vs theory on meta principles.
It is difficult to take perspective without theoretical idea.
The “ultimate nature of reality” likely *transcends concepts*. Incompleteness results (Gödel) suggest this.
Non-conceptual knowledge: any attempt to describe it with concepts is inherently wrong. (Analogy: Tao Te Ching’s opening line).
“to be the truth by letting go of all Concepts”. This type of truth isn’t knowledge but it’s entrirely “non-conceptual knoweledge”.
There may be an uncapturable deeper “synchronistic”.
Conflation. We often wrongly think our theories are true and this isn’t an accurate representation. A modest approach would acknowledge theory as human and it isn’t final.
Analogy: Gestalt perception, parts dont accurately represents “truth” or description.
Markovian Dynamics and Perspectives
Mathematical model attempt (humble): Stationary Markovian dynamics, where entropy remains constant (no entropic arrow of time). “It always was and always will be.”
Taking a *perspective* (projection) creates the *appearance* of increasing entropy, an arrow of time, and competition (e.g., evolution).
Things exist just the same. Projection makes loss of info appear to increase.
Example. from within Spacetime. It shows up: limited resources, Nature, organisms fighting, predicting processsing, blankets.
Brillint tools which shows what a perspective *could* appear.
This model demonstrates how a “synchronistic” system can appear to have these features from a specific viewpoint.
Time. There won’t be cause of anyting. Connection between perspective where there would be.
First-person perspecitive can affect it. Time is always now, which shows it might be an “illusion”.
Frame Rates, High-Dimensional Curves, and Events
Analogy: High-dimensional curve (possibly a knot) representing everything. Projections give the appearance of events, causes, etc.
High-dimensional curve as a project that gives Fractal Perspective.
Conceptual idea: varying experience like going along a subway train and experiencing waving parts of the same curve in different frame-rates.
Multile speeds would make eventfulness come or not (speed of viewing changes to experience).
Suggests a possible way to make “time” appear fractal in projections.
Positive Geometries (Timeless Structures)
Analogy to the EU initiative project, Universe+.
This new type of geometry allows them to take a step outside of Space-Time and is new, as its a 10 yr old.
It can provide predictions but isn’t dynamical (a “shape”, not a “process”)
These structures can predict particle interactions (e.g., gluon scattering) *more efficiently* than quantum field theory within spacetime. Show new symmetries (infinite yangying symmetry).
The theory, amplitud hedrin, is “more complicated” than polytrope “out there”.
SpaceTime and quantum theory *emerge* as projections of these deeper structures. M=4 of a projection of Amplitud Hedrin.
Current control knobs involve geometric properties (faces, volumes). Future models might involve dynamics and new, more powerful control knobs.
“Mass” might be the projected entropy rate of communicaiton classes. This offers us predictive power to get better theory for those.
Analogy: sequence of the action under the “old-model” will give new ways of actions of combinations.
The example of the M-value, whih controls dimensions to 4d. Why not higher d values?
Theories, Limits, and Dogmatism
All scientific theories have a *limited scope*. A *great* theory provides the tools to find its *own* limits (e.g., Einstein’s theory and the Planck scale).
Theory with Miracles vs one w/o: Theories can conflict with itself to give tools on how theory can conflict itself.
Philosophers debate if we are shooting ourselves on foot by self-referential arguemnts, but the field has good theories that have ability for mathematical theories which allows you to give up with current theory, with examples.
Analogy. incompleteness is a proof that logic itself cannot cover whole truth, such an arguemtn.
Dogmatic Science: the precise mathematics in a theory is good.
The Cure: mathematical. Dogmatic vs science is ability to find it wrong, by being slow-moving but being correct.
Chris Fields: only technologies will settle arguments because of how impactful the uses can become with those technologies, they will move to use it.
Time, Causation, and Agency (Circular Time)
Analogy: circular time model allows for prediction through looking into the future.
Prediction can be through taking a snapshot of circular motion.
Circular oscillations/vibrations: Objects are combinations of harmonics.
Agency emerges if it is as though events occur due to agency (predictions) or the cause (result). Only occurs with cyclic.
Analogy: The same event, such as prediction or causation can exist from just one perspective in time.
Looking through finer “frame-rate” could give illusion for the above.
Different “temporal windows” could explain observed phenomena (e.g., particle distributions inside the proton).
Small temperol windows could be connected with predicting high resolution “distributions” through “particles that look like massless” through artifacts of sampline.
Through this way, one is not looking to what they are, but closer, to the noise and the artifacts the closer you “observe” it to make prediction.
Nested selves could emerge from harmonics in oscillations.
Connections being worked between this theory to previous theory, such as to Markovian’s Dynamic.
Concious Agent Theory shows similar properties through Vibes, compatitibleness of harmony, through use of things such as Trave Logic and similar connections.
Connection: Taking particular frame rates is equivalent to predicting or “seeing” the past, so agency, is “looking like agency”.
Introduction
Cancer is often viewed as a single-cell problem of uncontrolled proliferation, but Levin argues it’s a problem of disrupted *large-scale* coordination.
Engineered constructs (robots) have a simple top-down architecture and so, cancer is not a problem; in contrast, biological systems consist of multi-scaled architecture in the biological domain where lower organizational levels can and will rebel.
Cells communicate electrically, forming networks (not just neurons) that process information about anatomical goals. Cancer can be detected, induced, *and normalized* by manipulating these bioelectrical signals.
Tissues make decisions electrically, and this can be targeted to alter cell behavior, with implications for many areas of medicine.
High regenerative capacity in animals is correlated with *low* cancer incidence, contrary to some predictions, as the regeneration mechanisms keep cancerous growth suppressed.
Planarian flatworms exemplify this: they’re immortal, highly regenerative, and cancer-resistant, even with chaotic genomes.
Multicellularity vs. Cancer
The key question isn’t “why cancer?”, but “why anything *but* cancer?”, given that we’re made of individually competent cells.
Single-celled organisms have their own agendas; multicellularity requires cooperation toward a larger anatomical “plan”. The human anatomy.
The genome specifies protein hardware, *not* the overall body plan. The overall body structure, arrangement, type.
Understanding how cells make collective large-scale decisions is key, not just molecular components, is very limited, as the body plan or the planarian is inherited somatically (with high mutations) and regenerates well with the anatomical bioelectric software playing a cruicial role in addition to the hardware genetics, and that biology, in general is far from fully knowing its large-scale-pattern-making algorthmic mechanisms.
Homeostasis: the error between normal pattern to errorred is corrected not just in small things such as blood sugar levels, but can and does control big and important, more structural changes in tissues such as limb and face growth in regenerative animals (axolotl limbs and eyes, and frog’s face’s error minimizing rearrangements).
Homeostatsis works as long as a higher-organizational blueprint exists that dictates the anatomical shape. If this pattern can be manipulated (such as setting thermostat in a house to make a different room-temperature-level), this provides much less difficulty than solving all the underlying molecular problems and errors.
Bioelectricity’s Role
Scaling competence: evolution upscales the tiny-level agendas to now work, under bioelectric control, together. Cancer is seen as an error of these goals’ breakdown.
Multicellular goals: Bioelectric signals do not imply and not is cancer a single-cell event: a breakdown in these goals leads to reversion to single-celled, more independent behavior as in Glioblastoma cells.
All cells have ion channels and voltage gradients across their membranes.
Different cell types have characteristic voltage ranges. Quiescent cells are polarized; proliferative/cancer cells are often depolarized. However this voltage different alone should not lead to an assumption of it being merely a single-cell-level phenomenon as they work with their other nearby cells, so it is far more complex.
Like brains, other tissues use bioelectric networks for information processing, acting like a type of anatomical blueprint.
“Neural decoding” (understanding thoughts from brain activity) can be extended *beyond* the nervous system. Reading and *understanding* electrical signals of tissues.
An “electric face” pre-pattern in frog embryos prefigures the future anatomy *before* gene expression, suggesting bioelectric instructions exist.
The pattern can change by using a “voltage-sensitive dye”. Tumors can be detected early as areas of bioelectric disruption (cells decoupling from the network) before full anatomical changes manifest.
Manipulating Bioelectricity
Tools have been developed to track, model, and, critically, *rewrite* bioelectrical patterns.
Rewriting: Unlike typical methods involving modifying at a genetic-hardware level, bioelectrical changes is like modifying thermostat setpoints instead of hardware rewiring; it’s a simpler way for “complex” problems.
This isn’t done with external fields; it uses the cells’ *native* communication mechanisms (ion channels, gap junctions). This can be via drugs, genes, or light to control gap junction and ionic flow.
The idea that they can guide other processes can be demonstrasted by how ectopic, extra, eyes, or organs, etc, can form anywhere on a frog (they make extra and different types and combinations) using controlled and targetted injections. This is very precise, as cells call neighboring cells into helping, similar to other types of intelligent organisms, such as ants.
Rewriting these patterns *instructs* cells: Ectopic eyes, limbs, etc., can be induced by recreating specific voltage patterns without genetic changes.
Cells do not follow hardwire rules; they correct in novel manners until goal is met.
Even brains with major mutations, and thus defective gene-hardware (example with ‘notch’ mutation, making very structurally poor brains), it can be changed using “bioelectrical-software-override” changes.
Damaged organs, e.g., those in cases of brain defects, can also be be “repaired”. Even IQ. By overriding gene problems with “bioelectircal set points” using anti-eplieptics.
Frogs, non-regenerative, induced to re-grow by inducing “pro-regenerative” blastemas, implying, with very brief (1day) exposure to cocktail and without touching it after, a leg would grow out from previously non-leg tissue to near completeion.
Bioelectricity and Cancer Treatment
This view makes four testable predictions, all supported by evidence:
1. Ion channel/pump genes are implicated in cancer molecular data (there should be some gene changes related to channel and protein, confirmed.)
2. Bioelectric signatures can be used for early cancer diagnosis (they can and it works).
3. Disrupting voltage gradients can *induce* cancer-like behavior (they do; experiment, by disrupting some melanocytes’ voltage communications, it is changed.)
4. Modulating voltage gradients can *suppress* cancer. Specifically via electroceutical drugs targetting ionic flows.
“Augmented-Reality Device” (prototype and potential): helps surgeons via overlays to visually confirm “areas of malignancy and risk”.
Oncogenes can cause cell disconnection, causing them to pursue selfish behaviors in which their ‘Self’ boundaries become downscaled from its wider self.
Experiment where, via injecting specific oncogenes, oncogene expression should occur in frogs, however co-injecting ionic channels in some frogs prevent it. They prevent oncogenic changes by correcting the electrical pattern and thus, it could be the physiology and not purely genetics.
Electroceuticals and Future Directions
“Electroceuticals”: Existing ion channel drugs, guided by computational models, can reprogram cell behavior.
Focus: Bioelectrical-changes are the software, the instruction layers; thus, its fixes, controls, or overides don’t need to alter hardware genome: it alters only on software-setpoint levels, not in hardware such as Crisper etc.
Using data from current knowledge of drugs (“drug bank”) + cells with ion channels (“physiomics”), it could target, control and suppress specific growths and patterns.
This is moving beyond theoretical and early stages. In vitro results with human glioblastoma show promise.
“Electroceutical Platforms” (Begining stages): drugs may, using a prediction algorithm, get prescribed via predicting what channel types and cells need for correcting cancer.
Future work involves improving: diagnostics to get early pre-cancer changes. Normalizing cancerous tumors back into normal tissues, and refining controls over mammalian-cell bioelectricity.
Introduction: A New Framework for Understanding Life
Levin’s approach integrates developmental biology, computer science, and philosophy to study agency, memory, and problem-solving in living systems. He proposes that it drives new discovery, in capabilities for building, communication, and unconvential “agent”-like constructs.
Focuses on “navigating problem spaces” as a key characteristic of life, allowing us to understand unconventional agents (cells, swarms, AI, etc.) using this characteristic as an invariant to compare them.
Primary example: collective intelligence of cells navigating anatomical “morphospace” (the space of possible body forms).
Electrical networks act as a “protocognitive medium,” enabling cells to solve problems in anatomical space, impacting, bioengineering and, biomedicine.
End-goal explores creation of synthetic living beings to understand the origins of novel goals.
Beyond Discrete Natural Kinds: A Continuum of Agency
Challenges the traditional view of distinct biological “kinds” (like Adam naming animals), arguing for a continuum of forms and agency, from single cells to humans, citing Darwin (Evolution).
Biotechnology and engineering further blur these lines, creating hybrids and chimeras, which requires new Frameworks to fit them in.
Framework needed to think about diverse agents: primates, birds, colonial organisms, synthetic life, AI, and even potential exobiological entities, which frameworks from rosenbluether and bigalow are considered..
Introduces “continuum of persuadability,” an engineering-focused approach: how systems change. Simple systems need rewiring; complex ones respond to experiences (dog training), reasons. Levin wants it not as just a philosophical concept but one used to directly impact research (where Modern bio, thinks “cells”, but should do testing).
Our development. We originate as a single, quiescent cell (“just physics”) but gradually transform into beings with psychological characteristics, going to state he “hates that phrase”. There is No single “spot”, even the unified singular intelligence deart spoke of for humans is actually decentralized collections of cells..
Therefore. We are not singular, centralized minds; but. “collective intelligences” built of “agential material” (cells with their own “agendas”).
Examples of, cellular decision making. Single-celled organisms competently handle needs without brains/nervous systems, and Caterpillars radically change (metamorphosis) but retain some memories, and planaria regenerate and their “cut” tails grows an original, retaining learned info.
Multi-Scale Competency Architecture: every biological level (cells, tissues, organs) has problem-solving capabilities in different “spaces” (transcriptional, physiological, anatomical). Human perception biased towards 3D; we might directly perceive complex physiological spaces if we had the right sensors, as cells do it internally.
Morphogenesis as Problem Solving: Turing’s Insight
Highlights Alan Turing’s interest in both computation/intelligence and morphogenesis (the development of form). Turing recognized this as 1) using different substraces, and 2) origins in chemical system (he researched).
Anatomical morphospace: the complex order of the body (organs, tissues) emerges from embryonic cells, yet the genome only encodes *protein* sequences, *not* a blueprint. Therefore Morphogenesis is a “software” problem. How cell groups collectively “decide” what to build. This information not written, but, emegeres, which is what Regenerative Medicing tries to understand.
Goals: how cell groups “know” what to make and when to stop (regeneration). As Engineers: exploring possibilities, which may mean creating a novel form of similar cells.
“Anatomical Compiler” (long-term goal): translate a desired shape into stimuli that guide cells to build it, with implication sin medicine and development/control, that wouldn’t need a printer to 3d assemble it.
The lack of an Anatomical compiler is Because Molecular biology “stuck” at hardware level (DNA editing, protein engineering), neglecting “software of life” (cellular intelligence, problem-solving).
Intelligence redefined (William James): reaching the same goal by different means, from Magnets unable to, and romeo and julia capable, of planning. A continuum. Requires *perturbation* (obstacles), not just observation.
Anatomical Homeostasis and Bioelectric Memory
The observation that. Development is reliable (normal embryos become normal organisms), is a start. However, embryos cut into pieces (twins) still form complete organisms. This suggests: *regulative development*: reaching the same anatomical goal by different paths.
Amphibians (e.g., salamanders) regenerate throughout life; even in humans/mammals it partially exists (livers, fingertips, deer antlers), so it shows this ability to “self repair”, is not “gone”.
Example in Kidney tubules. Cells with *more* genetic material become *larger*, but tubule *number* adjusts to maintain lumen size, and even Single *gigantic* cells can bend around themselves. All for: Large Scale Goal, to build correctly, not just a single static solution.
Challenges “feed-forward emergence”. There’s anatomical homeostasis *with feedback loops* (genetic and physical) guiding development back to “target morphology” and taking a differnet path as needed.
Contrast: Typical feedback loops (temperature, pH) have *scalar* set points, while. Anatomical homeostasis needs a *shape descriptor* (more information-rich). and Challenges: discouragement of “goal-directed” thinking in biology (anthropomorphism).
Claim: If anatomical homeostasis exists, we should be able to change the “set point” (desired shape) *without* rewiring the system at the molecular level. This involves a need to finding/encoding, decoding, and being able to rewrite setpoints, the memories, etc.
Need: Cognitive “glue” to combine: the collective. Neuroscience examples where, neurons collectively for a lever association, despite the experience being spread between foot-touch and food reward in disparate neurons.
Bioelectricity is the “glue” in nervous systems: ion channels create voltage gradients, forming networks for computation and memory. Not specific: Electrical properties, is ancient evolved *before* brains.
Proposes: “decoding” body’s collective intelligence by reading/interpreting bioelectric information (“mind of the body”). An anology that Brain : Commands Muscle, and Electrical newtoks gives command.
Reading and Rewriting the “Mind of the Body”
Voltage-reporting fluorescent dyes to visualize electrical communication between cells (“electric conversations” in embryos), along side quantitative Simulatirs.
Example: “electric face” in frog embryos, a bioelectric *pre-pattern* predicting future organ placement. Pathological, Enogen inducing cells that detach Electrically, so no communication.
*Rewriting* bioelectricity is, “intervening” directly, key: no applied fields/magnets/waves, modulating cells’ *native* ion channel “keyboard.”, By Modulating, optogentics and light, mutations of junctions. Not just a *readout*, but the *set point* for anatomical development.
Claim tested by *changing* the electric face (frog embryos) creating *ectopic eyes* in gut regions by injecting mRNA for ion channels. Lessons:.
Bioelectricity is *instructive* (not just toxicity; triggers specific organs).
Modularity.low-information signal (“make an eye here”) triggers complex processes; Competency. Only neurectoderm *thought* competent to make eyes (with pax6). Bioelectricity shows *all* cells can. “Bio-prompting” (analogy to AI),
Scaling. injected cells *recruit* neighboring cells.
These results are applied to Regenerative medicine applications. non-regenerating frog legs, a bioelectric “cocktail” triggers regeneration with long “delay,” therefore showing that early-communication set goal, rather than direct command to build.
Planaria: Bioelectric Patterns as Pattern Memories
Planaria: robust regenerators, “immortal,” noisy genome. 1. Cut: How the fragments “knows” how many heads.
Electrical circuit controls head number; targeting it creates two-headed worms, which has Bioelectric pre-pattern showing 1 head, “until you injure”.
Critical. The *bioelectric pattern is *not* a *map* of the two-headed animal, but of the *normal* one. A *memory* that can be *edited*, storing 2 representations, despite one single, singular-headed animal. “Simple CounterFactual.”
Recutting: shows permenant, forever, lasting change as: Memory (long-term, rewritable, conditional recall). Not: genome, with shows Normal Genome..
Shapes not limited to “number of heads”: Controls head *shape*. Confusing the bioelectric network in a triangular-headed species can produce *different* head shapes, as: The cells can reaccess stored information of old shapes..
Further exploration of “latent morphospace”: making planaria with different symmetries, hybrid forms, or “spiky” forms – all with normal cells and Genes.
Implication of latent space: there’s a large room for error, for cells. Others exploit these cells: Wasp, leaf example, with leaf cell “hack,” for nests. No way to see it on genes, since most time: its flat.
Challenge. Full “stack” understanding, connecting. Hardware. To: algorithms.
Behavioral Science, where you train cells, is Complementary: to biomedical approach. The controversial idea, for this research is. “Somatic psychiatry:” communicate to, cells.
Xenobots and the Origins of Goals
Changing Size and the origins: Goals of biological systems usually attributed to *evolution*.
“Cognitive light cone” (analogous to spacetime diagrams). Size of biggest goal, a system.
Ticks, bacteria: small. Dogs: bigger. Humans: very large. Different types of organisms.
Agents: We are *composed* of agents (cells, organs) with *different* sized cognitive cones, cooperating *and* competing, which changed. Evolutionary failure Mode = “Cancer,” where cancer cells detach from communication, “and have Smaller, selfs”. Not selfish.
Implication to the change: that Cancer does not necessarily needs death.
Xenobots: made from *frog skin cells* (*not* embryonic stem cells). Ask, “what if they had no borders”.
Xenobot Results. Spontaneous formation and self-organize, they swim by hairs, *novel* behaviors, navigate, spontaneous actions, signal, kinematic self-replication: assemble “children”. Never happened, *before*, which means evolution never required it to happen before, as new trait. Therefore this suggest, evolution, doesnt *just* produce “things”.
Interspecies compatability/compatibility of all these types: Living beings can interoperable with other types of living constructs to: create, agential, types of: agents. This means. Ethical questions on “agents”.
Questions, with Answer Highlights (simplified and paraphrased as short as possible):
Plants?.
Absolutely: they have Intelligence and fit in.
AI/Robotics and bio cells?.
Yes: exciting, feedback and use for synergy, using better alogirthms for biology to be capable of using them, due to it: helping machines, to understand biology..
Source of collective Intelligence?.
We know parts: Need connections, Memory-wiping. We *don’t* yet predict *specific goals* of collectives.
Tumors injectiion showed clutser, malignacies don’t.
*Induced*, by engonogen. Disconnect. They convincetheir neighbots. *Then*, reconnect.
How long for liver replacement?.
Can’t give. In a Lifetime (Frog-> Mice,).
Macroscopic Control (exerted on frog)?
we, can, change stuff “permenantly”. *Goal*: is to science.
Consider part? Dead/Other.
If other, use them, yes: include..
Tutura?, what is?.
“living fossil” rapid DNA evolution. Interesting, no prior knowledge of creature before Q+A.
“
Introduction: The Anatomical Compiler
Levin’s goal: Total rational control of biological growth and form (morphogenesis). This would solve many medical problems (birth defects, injury, cancer, aging) and enable new technologies (synthetic morphology, non-neuromorphic AI).
The “Anatomical Compiler” is a future system: You draw the desired organism (anatomy, not molecular biology), and the system generates stimuli to guide cells to build it. *Not* 3D printing or genomic editing, but a *communication* device to translate our goals to cells.
Current limitations: We lack this compiler; can only control morphology in very limited cases. Genetics/molecular biology alone aren’t enough, as a Genome encodes protein *hardware*, not large-scale anatomical *instructions*.
The Morphogenetic Code: Beyond Genetics
Example: “Frog-lottle” (Axolotl/Frog hybrid). Genomes of both are known, but we can’t predict if it will have legs, illustrating the gap between genotype and large-scale phenotype.
Forward vs. Reverse Problem: Going from simple rules to complex outcomes (like fractals) is easy. Going backward (regenerative medicine: “fix this asymmetry”) is incredibly hard (“intractable inverse problem”).
Molecular manipulation is not whole picture. Where biology was like needing to physically rewire hardware to acheive goals, a la old Computer Science of 40s and 50s, we need higher levels of understaning.
Multiscale Competency Architecture: Biological systems have problem-solving ability at *every* level (molecules, cells, tissues, organs, organism). Each level navigates its own “space” (gene expression, physiology, anatomy, behavior). We can *communicate* with these levels, not just rewire them.
Examples of biological problem-solving (intelligence): Embryonic development (twins from split embryos), regeneration (salamanders, axolotls – limbs, organs), deer antlers (rapid bone growth), human liver/fingertip regeneration.
Bioelectricity: The “Software” Layer
Inspiration from the Nervous System: The brain guides the body through behavioral space using electrical signals (ion channels, gap junctions).
The Same Applies to Anatomy: *All* cells have ion channels and gap junctions (not just neurons). Evolution discovered electrical networks for information processing long before brains.
Electrical networks are important because every cell has an ion channel, cells connected to cells and communicate, so that it should be treated like a kind of hardware which, with a certain amount of bioelectrical output/inputs/pattern-over-time, it should, conceptually, become a new form of electrical computer and this principle existed in the times of bacterial films and cells did not forget this just because it joined and coorperated with other cells to make larger, new forms of ‘computers’.
“Electric Face” of Frog Embryo: Voltage patterns *predict* future anatomy (eyes, mouth) *before* relevant genes are expressed. This is a bioelectric *memory* guiding development.
Pathological Patterns: Cancer cells show altered electrical states (decoupling from neighbors) *before* becoming tumors.
Rewriting Patterns: Tools (optogenetics, drugs targeting ion channels) allow us to *control* bioelectric states and thus *manipulate* development (induce extra organs, limbs, change body plan).
Permanent Changes: Two-headed planaria (flatworms) demonstrate that altered bioelectric patterns can be *stable* and *heritable* (without changing the genome). We rewrite the *memory* of the “correct” body plan.
Organ-Level Induction: A “subroutine call” (“make an eye here”) can trigger complex organogenesis. Cells can *recruit* neighbors, demonstrating collective intelligence.
Xenobots: Synthetic Biology and Emergent Behavior
Collaboration with Josh Bongard (UVM): Creating “Xenobots” from frog skin cells. When isolated, these cells *self-assemble* into motile structures. No brain and new neurons are needed.
Emergent behaviours can spontaneously arise: these include such behaviours: move in circles, patrol back/forth, interact collectively.
AI-Guided Design: Evolutionary algorithms can *predict* and *design* xenobot behavior (e.g., “Pac-Man” shape for particle collection). Xenobots can even build copies of themselves (“kinematic self-replication”).
Implications: Skin cells have a “hidden” behavioral repertoire, revealed by removing constraints. AI can help us control and understand this “native” intelligence.
Future Directions and Implications
Applications: Regenerative medicine (controlling wound healing, limb regeneration in mammals), cancer therapeutics (restoring electrical communication), biorobotics.
Long-Term Vision: Understanding and controlling “collective intelligence” of cells for various applications. Creating new forms of artificial intelligence inspired by biology.
Biology’s future is expected to involve lots of evolved/designed material/software at different scales. Thus ethical concerns become very necessary for navigating. Darwin’s “endless forms most beautiful” is just a tiny part.
Future systems will likely rely on using AI to make connections between software top-down processes which affect execution machinery down the line.
Broader Implications: Need for new ethics to deal with “hybrid” organisms (not fitting traditional categories). Expanding our concept of “intelligence” to include diverse biological systems.
“
Introduction: Engineering with Agential Materials
Biomedicine and bioengineering problems often boil down to controlling morphogenesis (cellular decision-making).
This control won’t be solved solely by hardware approaches (genomics). Biology uses a multi-scale competency architecture of nested problem solvers.
Evolution exploits a bioelectrical interface, which cells use to shape behavior and maintain structures.
We can read/write memories into the physiological layer of control, impacting birth defects, regeneration, cancer, and synthetic bioengineering. Focus of this particular is with “Engineering agential materials.” where behavior and congnitve tools can exploit this behaviour.
The endgame is an “anatomical compiler”: specifying a desired organism/organ at the anatomical level, and the system translates this into stimuli to build it.
This isn’t about micromanaging cell positions, but communicating goals to cell collectives.
Biology’s Unique Approach to Building: Agential Materials
Biology uses “agential materials”—materials with an agenda/own goals—not just passive, active, or computational materials.
*Example*: Single-celled Lacrymaria show how complex behaviours even single cells are capable of.
We transition from chemistry-based systems to systems amenable to high-level descriptions (behavioral science, psychoanalysis) during development.
Biology operates in “multiscale competency.” that isn’t based on *only* nested doll structural ideas of building blocks, but a *functional* one.
Each layers solve their own problems.
Engineered constructs are far behind biological systems in terms of adaptability, robustness, and plasticity.
Example of Plasticity: caterpillar-butterfly: how stored memory adapts with the biological hardware, even when the structure “largely dissolves”.
Example #2 of plasticity: Train a flatworm, chop it up, grows new brain. When that happens, information and “memory” comes back.
Example #3, even when eyes aren’t originally planned by the “blueprints,” this tadpole’s biological hardware adapts regardless.
Beyond 3D Space: Expanding Our View of Intelligence
We must widen our understanding of “problem spaces” beyond 3D. Intelligence exists in gene expression, physiological states, and anatomical states.
Anatomical space: Cells navigate the space of all possible configurations to create the body’s structure.
It’s tempting to attribute fully to the “blueprint” of the genome. But this cannot be. There exists the important intermediate step:
developmental physiology
The Challenges of Understanding Morphogenesis
Genome primarily encodes *nanoscale* hardware (protein sequences), cells then use developmental physiology for construction, meaning a blueprint from just genes isnt really that helpful and simple.
We need to understand how cell groups know *what* to build and *when* to stop, how to convince them to repair/rebuild. We want to know their inherent plasticity limits.
Current biosciences are good at manipulating molecules/cells but lack large-scale form/function control. Like old days of computing, too hardware-centric
Analogy to early computing: We need to move beyond “rewiring the hardware” (molecular manipulation) to higher-level control via “software” (information processing, decision-making).
Intelligence (William James definition): ability to reach the same goal by different means. Not about brain size, natural/engineered origins, but about *competency* levels.
This aligns with goal-directedness ideas of congnition.
Anatomical Homeostasis: Evidence for Morphogenetic Intelligence
Developmental self-assembly isn’t just about increasing complexity, but *adaptive* problem-solving, aka homeostasis.
Embryo splitting: Doesn’t create half-bodies, but whole organisms from various starting points. This suggests it isn’t a feedforward problem-solving structure.
Regeneration (axolotl example): Regeneration stops *when the correct structure is achieved*, implying an error-reduction scheme (anatomical homeostasis).
This *how* applies to other examples: Childrens fingers, newt kidney.
Adaptation to altered cellular parameters (newt kidney tubule): Cells adjust size/number, use different molecular mechanisms to maintain overall structure, showing flexibility.
*Important:* You cannot make assumption on priors of organisms when “engineering,” e.g. you cannot rely on certain number of chromosomes. It has to work. The enginnering paradigm has changed.
Response to disrupted morphology (tadpole face rearrangement): Organs move along *novel paths* to achieve the correct arrangement, challenging the “hardwired” development idea.
Implying*What evolution produces* aren’t merely specific solutions to problems but also machines capable of problems solving in various spaces (anatomical, physiological, chemical, behavioral).
In other words, Evolution produces problem solving agents, which use a feedback scheme (pattern homestasis) that responds to injuries, errors, problems (set of feeback loops) which attempts to “reach” the normal final form, as seen before.
Bioelectricity: The Morphogenetic Memory
Prediction based on previous points: There exists a literal recoreded explicit memory set.
Analogy to neural networks: Cells store memories and communicate via electrical signals (ion channels, gap junctions), similar to brains.
Can we decode somatic electrical networks, as neuroscience does for neural networks, and see how information moves *through* anatomical space (Not just 3D space) ?
All cells have this bioelectrical infrastructure.
Tools: Voltage-sensitive dyes to visualize electrical patterns, computer simulations, manipulation of ion channels/gap junctions (optogenetics, drugs) – no external fields/radiation, but manipulation of the cell’s natural interface.
Goal:* treat morphogensis as behaviour (of cell collectives) where cells, collective, navigate morphospace in anatomical space.
“Electric face”: Early embryos show a pre-pattern of future facial features in their bioelectrical activity, *before* anatomical structures develop,
but is also “causal,” manipulating bioelectricity impacts and disrupts anatomy.
Pathological Pattern: Examples is: Inject a tumor, oncogenese and so on will create metastasis but these patterns *show earlier* than anatomy, where the tumor breaks free.
implies we can measure patterns with “tools” from bioelctricity earlier, potentially diagnosing disease much faster and earlier.
Can you change, insert “eyes” into tissues and spaces. Answer: Yes:
Reprogramming Morphogenesis with Bioelectricity: Case 1 – Tadpoles
Ectopic eye induction: Inducing eyes in the *gut* region of tadpoles by manipulating voltage patterns.
These eyes have *all correct biological structure*. It even reucrits neighbour cells, *implying instruction.*
This is a *modular, high-level trigger*: We provide a simple pattern, and cells handle the complexity of eye construction, much like a
high-level subroutine.
A frog bioregenerator coctail triggers the regrowth of legs and toes and muscles.
It is also “functional” – tadpole limbs respond to light/touch.
Reprogramming Morphogenesis with Bioelectricity: Case 2 – Planaria
Planaria: Amazing regenerative capacity. Each fragment “knows” what a complete planarian should look like, as holographic in structure. They are effectively immortal as well (can ask).
Also how can a fragmnet know how many heads there should be, in fact, the correct numbr should be (hint: there are other forms of bioelectrical patterns, there must be a form/circuit pattern to describe it):
Head number control: An electrical circuit determines head number. We can *rewrite* this circuit (with ion channel drugs) to create two-headed worms.
Crucially, the electrical map is not of the two-headed worm, but of the *normal*, one-headed form. It represents the *set point* for anatomical homeostasis.
Similar: it can make the heads of *other species.* (Different by “100 and 15 million years,” even! But these differences aren’t genetic, so there is not issue!)
And even: crazier “shapes.”
The memory in Planaria shows all properties of memory.
Latent Morphospace: Can trigger *other shapes*, with their appropriate “shapes,” cells/other, different by large changes of million of year diffences between animals, yet done without
“genetics,” only by “guiding” morphogenetic bioelctric networls: These structures can “exsit” in morphospace!
It can make shapes never even made or considered! These spaces and shape changes exist! The idea “morphogenetic fields are limted is not only incorrect,” these latent shape/morphgenetic structures are likely numerous.
Another example: galls on tree/plant “hacked” by wasps that completely changes the morphology (even the genome isn’t different, yet structure change).
Implications, Connections, and Clinical Applications
We must move from controlling at *low levels* and moving toward high levels using tools of analysis of biolectrical pattern of a “competent material” which can exploit intelligence to move and act.
Bioelectricity provides an entry point to control these goals, including in clinical settings.
Cancer as a failure mode of “goal constriction”: Disconnected cells revert to smaller, unicellular-scale goals, resulting in proliferation.
But by enforcing electrical connectivity (even with a strong oncogene present), we can “force” cooperation toward normal tissue construction.
Xenobots: Uncovering Hidden Potential
Synthetic bioengineering (Xenobots): Isolated frog skin cells self-assemble into novel organisms (xenobots) with unique behaviors (movement, kinematic self-replication).
Shows there is potential “other structures” in different combinations of existing cells and how they organize.
Engineering *by subtraction*, freeing these *existing frog* cells allow them to self-form, moving past their initial roles and instructions of “building blocks.” They can row, move, and they are “super interesting”!
Shows other properties as expected of “smart agent”: can even “heal itself”. and ” kinematic self replication: fulfills “von nuemans” dream.
*Example of new “smart form/material:” * Anthrobots are a *human form* made from only normal tracial cells,
showing, the inherent multiceullar property/abilities for them to organize, grow, structure, and make changes! These aren’t
even frog or “new/unknown” cells!
When applied onto another damage site of neurons, shows they
*themselves, apply change* implying that existing cells are already
already well positioned to engage on tissue engineering/damage when we change these “bioelectrical, instructions.”
Evolutionary backstrory – these changes can exist:
Conclusion: Embracing the Agential Nature of Living Material
Cells/tissues possess numerous competencies. Our job is to understand/program them, leveraging their inherent intelligence.
We are at the *earily days*.
“endless forms most beautiful,” “exploring” are ideas to use when combining engineering design to biology.
Crispr, synthetic biology, and biorobotics can be unlocked by understanding the “intelligent, agential nature” of the material, moving beyond molecular control.
Bioelctricity, top down congtrol over shape space and its innate potential is a way.
We can control over the various properties of cells by analyzing it like cogntive agents: competencies, goals.
Tools include: voltage analysis, AI tools.
Introduction: Rethinking Discrete Categories
Traditional biology relies on discrete categories (species, individuals), but this breaks down with evolution, development, and future bioengineering. We need a framework for diverse intelligences, regardless of origin or composition.
Adam naming animals (genesis): name meant knowign their essence; Deep point, which means we must name and “find true name and essence” of the new chimerical types of minds which shall arrise; Discrete categorization that comes from adam (seperate sepeices, even humans) is useless; All future intelligences can blend between software/biology.
The categories of defining which is which (between different intelligent agent) is going to fall. The idea that there is discreteness of each animal. All biological species blends together along an evolutionarilly and developmentally; humans is not one thing, not a seperate discreet entity from the others, it has developed into being, over time. This idea to go find some magical “line” is gone (gone over the study of evolution); same applies to individual developent (that there is no discreet thing to be considered a ‘thing’); same issue with time going FORWARD as future biological interventions will break and blur that ‘line’ (in terms of engineering)
Scaling of Intelligence: From Cells to Minds
All organisms, including humans, are collective intelligences: collections of cells cooperating to achieve goals. We are each collective intelligences.
Development: The journey from a single cell (with “just physics”) to a complex mind is gradual, *not* a sudden jump. There’s no bright line where “mind” appears.
Cells are “agential material,” not passive like Legos. Cells have their own agendas and problem-solving capacities.
Unified Intelligence? is questioned, because even our singular brain contains billions of pieces. Even inside the parts inside a Pineal Gland: they contain many, individual, parts. The magic “part” that made Descartes think our conciosness originated was made of countless parts and those, even still, have their own parts, etc.
Alan Turning studied both computer intellgience and morphogensis. The point that Allen Turing probably had but didn’t quite articulate, that Levin makes here: “the process of intelligence forming and morphogenesis (anatomy of a living creature, from fertilization to full creature) formation is incredibly similar (they follow very very similar lines; perhaps two aspects of the same thing).
Embryonic “counting”: The number of “selves” in an embryo isn’t fixed by genetics; it’s a dynamic physiological process. Cells “decide” which collective to join.
Splint Brain studies reveals, like embryonic counting, this phenomenon also occurse *in* the brain.
Radical metamorphosis (caterpillar to butterfly): Memory persists even when the brain is drastically remodeled. It *generalizes* information (leaf color to “food”).
Who ownes “knowledge”? (The Lever Pulling Rat: The skin touches lever. the tummy get treats. No ONE cell experiences both, no cells has that knowledge, therefore, the entity which has the knowledge is THE ENTIRE ORGANISM; nervous system as cognative glue.)
Planarian regeneration: Memory can be stored *outside* the brain and imprinted on a new brain. The target morphology (what to regenerate) can be changed without altering DNA.
Multi-scale competency architecture: Every level (molecular, cellular, tissue, organ) can solve problems in its own “space” (gene expression, physiology, anatomy).
Morphogenesis as Collective Intelligence Behavior
Human “Morphospace”: Organisms navigate “morphospace” (the space of possible shapes). Development isn’t just “reliable”; it’s robust and plastic.
Development isn’t *just* reliable: Salmander examples of growing kidneys/tubule structure correctly despite different genetic abberations; they use different mechanisms to end up witht he *same* structures (a key indicator for intelligence).
Tadpole/frog facial rearrangments (Scrambling/Piccaso frog example) showed that it *still* can build new organs, despite abberations. The ability to have robust flexibility to “make” new ways (if the typical development pathways have been removed/destroyed/abberated in some ways) *IS* and indicator for a type of “problem solving intelligence”.
Morphogensis is GOAL ORIENTED PROBLEM SOLVING, *not* simple emergency or “insturctions”.
Fly patterns (“virtual ants”): Morphogenetic outcomes are not strictly limited by the genome. Other bioengineers (e.g., wasps on oak leaves) can induce radically different forms.
The “space” or total posssibilites of what and how organism can reconfigure itself has never been fully measured or quantified; so far, we just *dont know*. All estimations and constraints are on OUR part (on humans parts) not on the organism’s “part”; there may be incredible ways organism can self-assemble that we don’t understand and has NEVER been documented, or thought possible, before. The typical constraint (it cannot be like that! it will NEVER be like that!) are only a reflection of what *we* can conceptualise (which may or may not correlate to reality at all).
Morphogenesis is behavior: a collective intelligence of cells *behaving* in anatomical space, aiming for a “target morphology” (like a setpoint).
Communicating with that intelligence: Bioelectricity is a key interface, not just the nervous system. We can “read and write” the “mind” of morphogenesis using bioelectric signals.
Tumor Supression through bioelectric modulation and connections.
Ectopic eye formation: Bioelectric signals can instruct cells to build an eye in an unexpected location (e.g., the gut), overriding genetic “competency” limitations.
Xenobots (Frog Skin cells): cells have inherent, self assembling ability to construct novel and unique behaviors.
Anthrobots: Human cells *ALSO* have same unique emergent problem-solving and “novel structure creating and finding and making” when they are in some kind of environement, they find/create novel structrues: example “superbot”, many antibots join, and the “knit back the nurons” together.
Implications: the goal isnt just to see morphogensis/intelligence through an etheareal philosophically.
Implications for AI and Ethics
The *real* AI question: How to recognize and ethically relate to diverse, potentially alien intelligences (biological, artificial, hybrid). Not just about current language models.
There exist a “Persuadibility Spectrim” of: rewiring, cybernetics, behavioural Science and training, all the way to a “human”-like “cogent reasoning”, on it, things exists. (not about how “human” it is).
Many other Minds that “fit” into the category of an intelligence exists that is not just Human intelligence. We dont want to deny Intelligence/Mind/value to things because they arent “human”-enough.
There is the potential and “very easy to fall into”-risk of Ethicial Mistakes and Errors (such as humans has a long, dangerous, and destructive histroy of being in “in-groups and out-groups”); for things that exists at this extreme (non-human intelligence) that may very well exist, we, for our own interest, want to learn and study their properties instead of treating them as a out-group and dismissing them as simple tools/non-cognative things to use/abuse/exploit (similar, in histroy, some humans use to threat certain outgroup humans; as non-humans, to exploit).
This *is* a synbiosis.
Don’t judge beings based on origin (evolved, engineered, software) or composition (“metallic clang”). We need better ethical frameworks.
This isn’t just philosophy: It leads to new discoveries in biomedicine, engineering, and understanding intelligence itself. (there exists discoveries we could have, in other areas, by considering other non-traditional intelligent systems, too, this isnt “just philosophy”)
Objectphilia to Love for Your Own Kind: a scale; objectophillia is when people love inanimate object; versus love-only-your-own-kind, it is far more dangeous to think there exists intelligences that *do* exists (such as A.I. and those biological or digital/bio structures mentioned earleir), to put it on this *spectrum*, that love-only-your-own-kind” spectrum: (where it’s too similar, and you treat them as inamitate objects and use them) is going to become dangerous to ALL parties involved.