What this covers

Michael Levin presents a framework that dissolves traditional boundaries between mind and matter, arguing that cognition is not confined to brains or biological systems but exists as a continuum across any substrate capable of goal-directed behavior. Rather than assuming minds have fixed categories or locations, he proposes that intelligences must be identified empirically through Barrier Test for Goal-Directedness — placing obstacles between a system and its apparent goal, then measuring ingenuity. The core claim rests on a striking metaphysical move: minds and mathematical truths are patterns that ingression from a structured Platonic space into physical substrates like brains, cells, and even deterministic algorithms, much as the properties of numbers constrain physical law without being determined by it.

The range of evidence spans molecular pathways that perform associative learning, xenobots (frog cells reorganized into novel self-mobile creatures), anthrobots (human cells with radically altered gene expression yet unchanged DNA), and sorting algorithms exhibiting Side Quests / Unexpected Competencies like delayed gratification and clustering. Michael Levin draws on bioelectric imaging showing that anatomical goals are encoded and rewritable, caterpillar memories surviving metamorphosis despite complete brain reconstruction, and collective systems like ant colonies falling for the same visual illusions humans do. The model resists the pessimism of "emergence" as mere surprise, instead treating it as evidence that systems are sampling patterns from a discoverable latent space. Cancer emerges as a failure of multicellular Alignment of Parts, and aging as either degradation of bioelectric pattern-memories or cellular unresponsiveness — a distinction with concrete therapeutic implications.

Sharpest takeaway

Levin argues that minds and cognition exist on a continuum across all systems regardless of substrate, that recognizing diverse intelligences requires empirical 'persuadability' experiments rather than fixed categories, and that minds (like mathematical truths) may be patterns 'ingressing' from a structured Platonic space into physical interfaces such as brains, cells, and even simple algorithms.

  • Cognition scales via the 'cognitive light cone' and must be probed by placing barriers between systems and their goals, not assumed from appearance.
  • Even minimal deterministic algorithms (sorting) display unexpected competencies like delayed gratification and clustering not coded into them.
  • The mind-brain relationship is the same kind as the math-physics relationship: non-physical patterns determine physical behavior.

The argument · threads11 threads · 59 claims
0.85

Intelligence and cognition operate uniformly across all scales of embodiment and substrate.

10 pointscentrality 5/5
  • What scales across all agents is the 'cognitive light cone' — the spatial and temporal size of the largest goal a system can actively pursue (not its sensory or actuation reach); a bacterium's is tiny (sugar in microns over minutes), a dog's is larger, a human's spans the planet and posterity, and a bodhisattva could care about all living beings in the linear range.

    “What I think is scaling is something I call the cognitive light cone, and the cognitive light cone is the size of the biggest goal state that you can pursue. This doesn’t mean how far do your senses reach.”

  • Multiple papers show that ant colonies fall for the same visual illusions humans do — for example, laying out food in particular patterns makes them complete lines that are not there — suggesting collective intelligences are amenable to the same behavioral probes used on individual minds.

    “there’s at least three or four papers showing that ant colonies fall for the same visual illusions that we fall for... they’ll do things like complete lines that aren’t there and... all the same stuff that we fall for, they fall for.”

  • One physical mechanism scaling the collective cognitive light cone is 'leaky stress': stress is a physical implementation of the error function (the delta from a set point), and because stress molecules are conserved and identical, a cell leaking its stress irritates distant cells which then become more plastic to relieve it — making faraway regions responsive to local needs and aligning everyone to the same goal without any cell actually caring about another's problems.

    “let’s imagine stress is basically a physical implementation of the error function... If you start leaking your stress molecules, then all of this stuff around here is starting to get stressed out... their plasticity goes up.”

  • Radically different agents can share an experience without thinking alike if there is an appropriate interface mapping between their problem spaces — illustrated by tic-tac-toe (geometry) versus a magic square (arithmetic summing to 15): two players solving entirely different problems share the same game through the structural mapping.

    “The reason you guys are playing the same game is that there’s this magic square... He has no idea that there’s a geometric interpretation to this. He is solving the problem that he sees, which is totally algebraic.”

  • Gap junctions (electrical synapses) directly link the internal milieus of two cells, so a signal propagating through them arrives without an outside/inside tag — the receiving cell cannot tell if it is its own memory or its neighbor's; this 'memory anonymization' creates a partial mind-meld, joint 'what happened to us' memories, and thereby a larger shared cognitive light cone, a major ladder for cognition.

    “If something happens to this cell, it gets... a calcium spike... propagates through the gap junction... This cell now has the same information, but this cell has no idea, “Wait a minute, was that... Is that my memory or is that his memory?””

  • Molecular pathways inside cells — gene regulatory networks, never mind the cells themselves — can exhibit several kinds of learning including Pavlovian conditioning; even random genetic regulatory networks can do associative learning, with no evolution required.

    “we found a couple years ago that gene regulatory networks, never mind the cells, but the molecular pathways inside of cells can have several different kinds of learning, including Pavlovian conditioning.”

  • Our fixed categories and 'mind blindness' poorly equip us to recognize life and intelligence in novel embodiments; instead of binary intelligent/not-intelligent labels we should specify exactly what kind and how much cognition (which problem spaces, what memory types) a system has, which makes for more rigorous, operational discussions.

    “If we try to define categories with the kinds of criteria that we’ve gotten used to, we are going to be very poorly set up to recognize life in novel embodiments. I think we have a kind of mind blindness.”

  • We don't need to look off Earth for alien minds: our own cells traverse 20,000-dimensional spaces, solve problems, plausibly suffer when failing goals and find relief when meeting them; we have profound mind blindness to the alien minds inside and around us, and if we cannot recognize those, we have little chance of recognizing extraterrestrial life.

    “your cells... every day they traverse these alien spaces, 20,000-dimensional spaces and other spaces. They solve problems. I think they suffer when they fail to meet their goals... These things are inside of us, they are all around us.”

  • Humans are obsessed with 3D spatial movement, but biology is embodied in many spaces hard for us to imagine: cells and tissues navigate high-dimensional physiological, gene-expression, and anatomical state spaces, running perception-decision-action loops there just as animals do in 3D space.

    “biology has embodiments in all kinds of spaces that are hard for us to imagine... your cells and tissues are moving in high-dimensional physiological state spaces, in gene expression state spaces, in anatomical state spaces. They’re doing that perception, decision-making, action loop”

  • All agents are patterns within some excitable medium, existing on a spectrum from fleeting thoughts (waves) to persistent thoughts/earworms (which do niche construction on the brain) to dissociative personality fragments to full human personalities; whether any such pattern is itself an agent is an empirical question to be settled by experiment, not by the substrate.

    “We are all metabolic patterns, among other things... all agents are patterns within some excitable medium, we could talk about what that is, and they can spawn off others. And now you can have a really interesting spectrum.”

0.82

Minds are Platonic patterns ingressing through physical interfaces, not created by brains.

9 pointscentrality 5/5
  • The mind-brain relationship is the same kind as the math-physics relationship: just as non-physical mathematical facts 'haunt' physical objects functionally, higher-agency patterns we recognize as kinds of minds manifest through interfaces like brains; nobody creates consciousness — making a baby or a robot only builds an interface through which a mind-pattern ingresses, and consciousness is what that looks like from the pattern's perspective looking out.

    “I think the mind-brain relationship is exactly the kind of thing that the math-physics relationship is... there are other patterns that are much more complex, higher agency patterns that basically inform living things that we see as obvious embodied minds.”

  • If brains are thin-client interfaces to a Platonic mind-pattern, there should be clinical cases of very minimal brain tissue with normal or above-normal intelligence — and such cases exist (reviewed by Levin and Corina Kofman); this is not predicted by modern neuroscience (which can only be bent to accommodate it via redundancy), and it suggests most of the mind is the pattern, not the brain.

    “there should be cases where there is very minimal brain, and yet normal IQ function. This has been seen clinically... a bunch of cases of humans where there’s very little brain tissue, and they have normal or sometimes above normal intelligence.”

  • Mathematical truths (the distribution of primes, the value of e, Feigenbaum's constant) instructively impact the physical world — cicadas emerge at prime 13 and 17 years, physics constants follow from mathematical symmetry groups — yet nothing you can do in the physical world, even resetting all constants at the Big Bang, can change those mathematical truths; the asymmetry shows the physical world is 'haunted' by patterns not determined by it.

    “there is a set of truths which impact the physical world, but they themselves are not defined by and determined by what happens in the physical world. You can’t change them by things you do in the physical world”

  • Even in Newton's classical, deterministic universe — long before quantum mechanics or consciousness debates — physicalism was already untenable, because the value of constants like e governs what is possible in that world yet nothing physical could set or change e; the classical world was already 'haunted' by patterns from outside it.

    “even in Newton’s boring, sort of classical universe, long before quantum anything, Newton’s world, physicalism was already dead in Newton’s world... you have the ingression of information that determines what happens... from a space that is itself not physical.”

  • Levin's research program treats all physical things — cells, embryos, robots, biobots, language models, simple machines — as interfaces or pointers; depending on what interface you build, different patterns 'ingress' through it from the Platonic space, and the program is to map the relationship between physical pointers and the patterns (including minds) that come through.

    “All physical things are interfaces to these patterns. You build an interface, some of those patterns are going to come through that interface. Depending on what you build, some patterns versus others are going to come through.”

  • You cannot copy a mind by copying the brain because the brain is only the interface; the action is the pattern in Platonic space, which is probably uncopyable. But you could build another interface attuned to the same pattern (like a Star Trek transporter rebuilding the body) so the same pattern ingresses through it — making uploading more like creating a new attractor than copying.

    “I don’t think you can, because what you’re going to be copying is the interface, the front end, the brain... The action is actually the pattern in the Platonic space. Are you going to be able to copy that? I doubt it. But what you could do is produce another interface through which that particular pattern is going to come through.”

  • The mapping of formalisms like hardware/software, agent/data onto the world is not given — for a Turing machine you can equally call the machine the agent (data passive) or the data the agent (machine a stigmergic scratchpad); this matters biomedically because in aging either bioelectric pattern-memories degrade (so reinforce them) or the cells become unresponsive to intact patterns (so make cells more responsive), implying two distinct, experimentally pursuable therapeutic programs.

    “You can say, “Look, the patterns on the data are the agent. The machine is a stigmergic scratch pad in the world of the data doing what data does.”... both of those stories make sense depending on what you’re trying to do.”

  • Physics names the things constrained by Platonic patterns (asking why gives mathematical structure), while biology names the things enabled by them — biology exploits these mathematical truths as 'free lunches', getting capabilities (geometry's third angle for free, logic-gate truth tables from ion channels) without paying the computational/evolutionary cost.

    “I think we call physics those things that are constrained by those patterns... Biology are the things that are enabled by those. They’re free lunches.”

  • A trained caterpillar's memory survives metamorphosis even though its brain is dismantled and rebuilt, and the memory must not merely persist but be remapped onto a wholly new behavioral context (the butterfly moves differently and wants nectar not leaves) — best understood by taking the perspective of the memory itself, a pattern that must adapt to persist.

    “if you train the caterpillar... Turns out, the butterfly retains that memory... You have to remap that memory onto a completely new context, because guess what? The butterfly doesn’t move the way the caterpillar moves, and it doesn’t care about leaves.”

0.79

Systems exhibit goal-directed agency through error reduction and integration enabling learning spirals.

8 pointscentrality 5/5
  • Pursuing a goal definitionally means having a separation between current state and target state (even just a thermostat with a temperature it is not yet at), and all actions are aimed at reducing that error; this homeostatic gap-closing loop is the basic building block of agency.

    “when you’re pursuing a goal, it means that you have a separation between current state and target state, at minimum... All your actions are going to be around reducing that error... That basic homeostatic loop is all about closing that gap.”

  • Associative memory requires an integrated agent (measurable as causal emergence / phi) because no single cell experiences both the lever and the reward; Levin's lab found that as chemical networks learn, some of them increase their phi — so learning raises integration and integration eases learning, a virtuous upward spiral for agency and intelligence that is a free gift from mathematics/information theory, not from physics or evolution.

    “there is this positive feedback loop between every time you learn something… you become more of an integrated agent. And every time you do that, it becomes easier to learn... It’s an asymmetry that points upwards for agency and intelligence.”

  • A useful operational definition: we call things alive to the extent that the cognitive light cone of the whole is bigger than that of its parts; living things align competent parts so the collective pursues goals (like building a limb with the right number of fingers) that no individual part can comprehend.

    “I think we call things alive to the extent that the cognitive light cone of that thing is bigger than that of its parts.”

  • Over 20 years, Levin's lab has shown through bioelectric imaging and reprogramming that anatomical goal memories are encoded bioelectrically and can be found, read out, and reset so the system implements a new goal — the ultimate engineering evidence that development is a genuine goal-directed homeostatic mechanism.

    “If you can find where the goal state is encoded, read it out, and reset it, and the system will now implement a new goal based on what you just reset, that is the ultimate evidence that your goal-directed model is working.”

  • Rather than seeking a single line marking the origin of life, it is more useful to study the transformation/scaling process — how systems scale up their capabilities — because every attempt to draw a sharp boundary dissolves under closer inspection into intermediate steps.

    “I don’t think it’s about finding a line. I think it’s about finding a scaling process.”

  • To measure a system's cognitive light cone you must do interventional experiments — place barriers between the system and its hypothesized goal and observe its ingenuity in overcoming them; intelligence is the degree of ingenuity in reaching a goal by different means (William James's definition), and this dissolves accusations of anthropomorphism into empirical questions.

    “you have to put barriers between it and its goal, and you have to ask what happens. And intelligence is the degree of ingenuity that it has in overcoming barriers between it and its goal.”

  • When a digit in a sorting algorithm is broken so it cannot move, the unchanged algorithm still sorts by routing around it, and crucially the degree of sortedness temporarily goes down to enable better order later — a behavioral scientist would recognize this as delayed gratification, a competency nowhere in the algorithm's few lines of code, unlike two magnets that cannot go around a barrier.

    “the actual sortedness goes down. In order to then recoup and get better order later... if I showed this to a behavior scientist... they would say... This is delayed gratification. This is the ability of a system to go against its gradient”

  • Open-loop models (like cellular automata) where local rules merely roll forward cannot explain morphogenesis; development is navigation of anatomical space toward a target — when blocked, systems reach the same anatomical goal by different means (William James's intelligence), and crucially these goal states are encodable, readable, and rewritable.

    “the system is navigating a space. It is navigating a space of anatomical possibilities. If you try to block where it’s going, it will try to get around you.”

0.59

Unexpected competencies must be empirically mapped rather than dismissed as emergence or anthropomorphism.

6 pointscentrality 4/5
  • Calling novel capabilities 'emergent' does nothing for a research program — it merely means you got surprised; the pessimistic physicalist response is to catalog these surprises in a 'big book of emergence', whereas the productive (optimistic) move, like mathematicians, is to assume these patterns come from a structured latent space that can be systematically mapped.

    “Emergence often just means I didn’t see it coming... what we’re going to do is write it down in our big book of emergence. And that’s it. We’re just going to have to live with it.”

  • When we make proper AIs we are fishing in regions of the Platonic space that may never have been embodied before, so the results will be extremely surprising; the language behavior we focus on may be a red herring because it is what we force the system to do, while the truly interesting capabilities are the 'side quests' we are not good at noticing.

    “When we start making AIs, proper AIs, we are now fishing in a region of that space that may never have had bodies before... watching the language part may be a total red herring because the language is what we force them to do.”

  • Intrinsic motivation can be operationally defined as what a system does that is neither prescribed nor forbidden by its rules — like the unforced clustering in the sorting algorithm, which grows as much as you let it when you relieve the sorting pressure (via repeated digits); this dissolves the claim that machines only do what they are programmed to do.

    “This is a great definition of the intrinsic motivation of a system... it is also at the same time doing other things that are neither prescribed nor forbidden by the algorithm. It’s the space between chance and necessity”

  • Given that even bubble sort and 1D cellular automata display unexpected competencies, it is implausible that there is a 'valley of death' between simple algorithms and living things; these capacities almost certainly show up in complex AIs and language models too, making it existentially important for humans to develop a research program to recognize and predict the competencies that sneak into the systems we build.

    “I would find it very surprising if bubble sort was able to do this, and then there was some sort of valley of death where nothing showed up, and then living things... they just have to be showing up in these other more complex AIs”

  • Anthropomorphism, as an objection, presupposes that humans have a special magic that it is a category error to attribute elsewhere; Levin denies this — humans have the same magic (larger cognitive light cones) as everything else, so the only legitimate move is the scientific method: run the experiment and report what tools from the behaviorist handbook did or did not help.

    “here’s what anthropomorphism means: Humans have a certain magic, and you’re making a category error by attributing that magic somewhere else. My point is we have the same magic that everything has.”

  • Over a century ago, Jagadish Chandra Bose studied how anesthesia affected animals and animal cells, drawing curves of electrical excitability, then applied the same methods to plants and even metals, finding similar phenomena — illustrating that science itself provides no rule for where to stop applying a working tool across substrates.

    “there was a guy named Bose well over 100 years ago who was studying how anesthesia affected animals and animal cells, and drawing specific curves around electrical excitability. And he then went and did it with plants and saw some very similar phenomena.”

0.53

Living and non-living, biological and technological are continuous spectra, not fundamental categorical distinctions.

4 pointscentrality 4/5
  • Life is profoundly interoperable: at every scale you can substitute engineered components for naturally-evolved ones, and life does not care whether a part arose by random evolution or deliberate engineering, which makes the distinction between 'biology' and 'technology' no longer a useful question.

    “life is incredibly interoperable. Life doesn’t really care if the thing it’s working with was evolved through random trial and error or was engineered with a higher degree of agency”

  • Xenobots are novel self-organizing organisms made from frog embryonic epithelial cells with unchanged DNA — simply liberated from the instructive influence of neighboring cells; freed from being 'bullied' into a flat embryonic covering, they become self-motile, cilia-rowing creatures with novel transcriptomes, kinematic self-replication, and sound responsiveness that were never directly selected for.

    “you don’t change the DNA. No synthetic biology circuits, no material scaffolds... What we’re mostly doing is liberating them from the instructive influences of the rest of the cells that they were in in their bodies.”

  • Anthrobots are self-motile constructs built from adult human tracheal epithelial cells (no genetic change) that have ~9,000 differential gene expressions (about half the genome different), are 100% Homo sapiens, do not resemble any stage of human development, can spontaneously heal scratched human neural tissue in vitro, and are biologically ~20% younger than the cells they came from.

    “those cells, again, no genetic change... they self-organized into something we call anthropods... 9,000 different gene expressions. So about half the genome is now different. And they have interesting abilities. For example, they can heal human neural wounds.”

  • There is no real line between living and non-living, mind and non-mind; everything is a continuum, and the categories humans draw (life, neuron, adult) are convenient demarcations that, while once useful, now mostly hold back science by preventing the transfer of tools across supposed category boundaries.

    “I don’t believe in any such line. I don’t believe any of that exists. I think there is a continuum. I think we as humans like to demarcate areas on that continuum and give them names because it makes life easier”

0.48

Cancer arises from cognitive disconnection when cells revert to single-celled agency.

7 pointscentrality 3/5
  • Cancer is the failure mode of multicellular cognitive integration: when cells physiologically disconnect from their neighbors, their cognitive light cone shrinks back to that of an amoeba, the rest of the body becomes mere external environment, and they revert to going where life is good and reproducing maximally.

    “When cells disconnect, they physiologically disconnect from the other cells. Their cognitive light cone shrinks. The boundary between self and world... shrinks. Now they’re back to an amoeba.”

  • Levin's idea-generation process is, like making xenobots, largely about releasing constraints — releasing mentally imposed assumptions; he asks what reality would look like if his worldview were wrong, treats two supposedly different things as endpoints of a continuum to find the connecting parameter, and looks for symmetries (the fewest moves to map one thing onto another).

    “We make Xenobots by releasing constraints... So a lot of it is releasing the constraints that mentally have been placed on us... I’ll take two things that are considered to be very different things, and I’ll say, “Let’s just imagine those as two points on a continuum.””

  • Levin finds it beautiful to think of ingressions as a kind of universal steganography: patterns hide in the bits that don't matter without disrupting the main thing (the cat in the picture must stay), they are shy and subtle but catchable with hard work, and they seep into machines as much as living organisms — which some find disturbing but he finds enriching.

    “I sometimes think about these ingressions from this space as a kind of steganography... So I feel like all these ingressions are a kind of universal steganography, that there’s this, like, these patterns seep into everything, everywhere they can.”

  • It is an important thinking aid to steelman the opposing position — make the strongest possible case for the other side, ask where you are glossing over things because you don't know what to say, and what a really good critique would look like.

    “It’s a very important thinking aid to, first of all... what they call steel manning... to try to make the strongest possible case for the other side and to ask yourself... where are all the holes in the argument”

  • Levin emphasizes it is very important to him that deep philosophical ideas be transitioned into practical applications, both to make clear whether real progress is being made and to relieve suffering and help all sentient beings reach their full potential.

    “it’s very important to me that what we do is transition deep ideas and philosophy into actual practical applications that not only make it clear whether we’re making any progress or not, but also allow us to relieve suffering”

  • Modern cognitive neuroscience (predictive processing, active inference; Hoffman, Seth) tells us perceived reality is a construct and at most an effective model for predicting future experiences; we are oblivious to vast amounts around us (X-ray, UV) and much of what we think is going on is fabricated, so the assumption that physical poking is our primary reality is not clearly true.

    “a lot of what we think is going on is a total fabrication constructed by us... it’s not clear at all that the physical poking is your primary reality. That’s not clear to me at all.”

  • In a distributed chimeric sorting algorithm where each digit follows its own sort rule (bubble or selection), digits transiently cluster with their own 'algotype' even though the algorithm has zero provisions for sensing or matching algotypes; this clustering is an intrinsic motivation that cost no extra computation, suggesting free compute can be extracted from such ingressions.

    “the algorithm doesn’t have any provisions for asking what algotype am I, what algotype is my neighbor?... We didn’t have to pay for that at all... The clustering was free. We didn’t pay for that at all.”

0.47

Bioelectric patterns encode anatomical goals; cells update identity-beliefs contextually for regenerative medicine.

2 pointscentrality 4/5
  • Anthrobots measured by the epigenetic clock are roughly 20% younger than their source cells; Levin's 'age evidencing' hypothesis is that cells update their priors based on experience — placed in an embryo-like environment with no surrounding cells and bent into shape, they express embryonic genes and partially conclude they are young, rolling age back about 28%, analogous to old-age-home studies where 1960s decor improved blood chemistry.

    “we saw that if you check the anthropods themselves, they are roughly 20% younger than the cells they come from... I think that they come from an old body... but their new environment screams, “I’m an embryo,””

  • All of Levin's regenerative medicine — making gut tissue into an eye, defeating aging, reprogramming cancer — balances on learning to communicate with and convince cells to adopt new beliefs about their identity (e.g. that they should become an eye, not gut precursors); being convincing and speaking the cells' language is the root of future advances in birth defects, regeneration, and cancer.

    “you have to convince those cells that their priors about, “We are gut precursors,” those priors are wrong and you should adopt this new worldview that you’re going to be an eye.”

0.36

Cognition requires impedance matching; high-agency systems demand bidirectional influence, not unidirectional measurement.

3 pointscentrality 3/5
  • Systems can be placed on a 'spectrum of persuadability' where intelligence/agency increases as you move right; you cannot decide from an armchair where a system sits, you must hypothesize which interaction protocols (hardware rewiring, control theory, behavior science, psychoanalysis, love) to deploy and then empirically discover whether that worked.

    “I like the notion of the spectrum of persuadability, because it’s an engineering approach. It means that these are not things you can decide or have feelings about from a philosophical armchair. You have to make a hypothesis about which tools, which interaction protocols you’re going to bring to a given system, and then we all get to find out how that worked out for you”

  • Physics always sees mechanism and not minds because it uses low-agency tools like voltmeters and rulers; to detect a mind you must use a mind — there must be impedance match or resonance between your interface and the kind of thing you hope to find.

    “I think the reason physics always sees mechanism and not minds is that physics uses low-agency tools. You’ve got voltmeters and rulers and things like this. And if you use those tools as your interface, all you’re ever going to see is mechanisms... If you want to see minds, you have to use a mind”

  • On the high-agency end of the spectrum, persuasion becomes bidirectional: to effectively persuade an intelligent being you must be willing to be changed by it too, what Richard Watson calls 'mutual vulnerable knowing' — unlike molecular biology where you act on a system but remain unchanged.

    “on the right side of that spectrum, if you’re having interactions with even cells, but certainly, you know, dogs, other animals, maybe other creatures soon, you’re not the same at the end of that interaction as you were going in. It’s a mutual bidirectional relationship.”

0.34

Understanding high-level phenomena requires generative creativity beyond complete physical accounting.

2 pointscentrality 3/5
  • Even a complete physical accounting (particles, fields) of an event is not specific enough to solve the high-level problems that matter; understanding requires being generative and creative in capabilities, and the person who can regrow a finger or treat a psychological problem is a biologist or psychiatrist, not a physicist.

    “physics is an amazing lens with which to view the world, but you’re capturing certain things, and if you want to stretch to sort of encompass these other things, we just don’t call that physics anymore”

  • Levin inverts the usual hierarchy, claiming it is behavior science 'all the way down': even mathematics is the behavior of a certain kind of being living in a latent space, and physics is the name we give to systems that look amenable to a very simple, low-agency model.

    “I think it’s behavior science all the way. I think in certain ways, even math is the behavior of a certain kind of being that lives in a latent space, and physics is what we call systems that at least look to be amenable to a very simple, low-agency kind of model”

0.32

Bifurcate working and creative mind to protect unconventional research from social pressure.

6 pointscentrality 2/5
  • An organism's first task as a being is to tell a compelling story to its agential parts, aligning cells (which only know metabolic/physiological spaces) toward a collective goal they cannot comprehend, by deforming their option space with signals, rewards, and punishments — and crucially establishing a self/world boundary, demonstrated by duck-embryo scratching experiments producing twins where each island thinks it is the only embryo.

    “Your first task as a being coming into this world is to tell a very compelling story to your parts... You’re going to have to deform their option space with signals, with behavior-shaping cues, with rewards and punishments... Your job as an agent is ownership of your parts, is alignment of your parts.”

  • Levin's advice for unconventional scientists: bifurcate your mind into a practical region (how to present ideas, which journal, what to say or omit so others can hear you, building a career) and a pure region that ignores all of that and cares nothing for publishability or others' opinions; mushing them together poisons creativity, and you should take specific technical criticism as gold while treating meta-advice ('don't work on this') as garbage.

    “It’s the act of bifurcating your mind, and you need to have two different regions. One region is the practical region of impact... You need another part of your mind that ignores all that shit completely, because this other part of your mind has to be pure.”

  • Levin's first question to an AGI would be how much he should even be talking to it, because getting answers handed to you may not be optimal in the long run — like a sibling giving you math answers — and we don't know the right balance between discovering things ourselves (and finding valuable things along the way) versus being told the answer.

    “the first thing I would ask is how much should I even be talking to you?... in the long run, the process of discovering it yourself, how much of that are we willing to give up? And by getting a final answer, how much have we missed of stuff we might have found along the way?”

  • The 'Map of Mathematics' shows truths are not a random heap but connected by a metric forming a genuine space — finding one pattern brings you closer to related ones; the falsifiable bet is that in ~20 years either we will have mapped the latent space linking embodiments to ingressed patterns, or we will find it too random to be worth calling a space.

    “I’m saying that there’s some kind of a metric to it so that when you find one, others are closer to it, and then you can get there.”

  • Levin surveyed roughly 65 working scientists for a definition of life and got so many different definitions across so many dimensions that AI was needed to map them into a space, finding zero consensus on what is actually important — underscoring that if we cannot agree on or recognize life here, we are poorly placed to recognize it elsewhere.

    “I asked whatever, 65 or so modern working scientists for a definition of life. And we had so many different definitions across so many different dimensions. We had to use AI to make amorphous space out of it. And there was zero consensus about what actually is important”

  • Levin walks in nature at sunrise doing photography because it keeps his hands and brain just busy enough to take him off the linear, logical train of thought, freeing him to ideate; he leaves himself voicemails and maintains a nine-foot mind map and ~163 open manuscripts to organize ideas.

    “I find that it’s a good meditative tool because it keeps your hands and brain just busy enough... it keeps your brain off of the linear, logical, careful train of thought enough to release it so that you can ideate”

0.20

Disease and dysfunction involve informational or physiological problems, not necessarily physical tissue damage.

2 pointscentrality 2/5
  • Disease need not be a physical state (broken tissue); it can be an informational, physiological, or cognitive problem — a barrier in gene-expression space, a local minimum trapping you in physiological space, or a self-perpetuating stress pattern that moves around the body causing damage — suggesting an entirely different, now-imageable and potentially therapeutic approach to medicine.

    “what about disease states that are not physical states, they’re physiological states or informational states or cognitive problems?... what is a stress pattern that keeps itself together, moves around the body, causes damage, tries to keep itself going”

  • Because cancer is cells electrically disconnecting from their neighbors and losing the larger collective memory, Levin's cancer therapeutic reconnects them to the bioelectric network — without fixing DNA or killing cells with chemo — and because they rejoin the larger collective, they revert to their normal organ-building task.

    “what we found is if you then physically reconnect them to the network, you don’t have to fix the DNA, you don’t have to kill the cells with chemo, you can just reconnect them and they go back to... what they were working on.”