Bioelectric Fields: A Paradigm Shift In Biology | Prof. Michael Levin

I had the honor of sitting down with Michael leav who is working on a paradigm shift in medison and biology you can inject human enogen into a tple and they will make tumors and if you do that but you force the cells into proper electrical state with their communicating with their neighbors you prevent tumorogenesis even though the enogen is blazingly strongly expressed cancer is the second leading cause of death globally and we fight the disease primarily with chemotherapy radiation or surgery but what if there would be a way to tell cancer cells to stop growing

through the use of Bio electricity what you're doing is you're forcing that cell to stay connected to their neighbors and when when you have a mind meld with a large Collective that all says we are building a nice kidney we are building a skin it's really hard for that cell to have an independent thought of you know what I I my goal is to proliferate as much as I can and crawl away wherever life is good right metastasis okay so what we want in our bodies is a mind melt of cells that cells forget about

their own little ego and start identifying with a larger Collective of the organ and the way cells do that is through bioelectricity it is what Michael Levan calls a cognitive glue what's actually important about bioelectricity is that it functions as a cognitive glue and what I mean by that is you know think of think of how it works in the brain we are a a giant collection of uh an enormous collection of individual neurons and yet those neurons uh come together in some way to give rise to a new individual that has goals preferences memories

and capacities that none of the individual neurons have so basically we are a collective intelligence of cells held together by bioelectricity now what cancer cells do is they disconnect the first thing that happens when anagen Act first thing they do is they cause electrical uh disconnection from their neighbors and as soon as you disconnected well now you're in amoeba again and as far as you're concerned the rest of the body is just external environment so that that border between self and world can shrink and grow now what's groundbreaking about the work of Michael lvin and

his lab is that they found a way to tell cancer cells in frogs to start cooperating again and we've done this in the in the Frog model we've already done this so this stuff uh is not just uh you know Wishful philosophical thinking and you know no no it has it has very practical implications that if you follow through you discover Therapeutics but things get even more exciting because in bioelectric fields Michael Lev and his lab found blueprints of organisms so we have now tools where you can literally see for example the electric face this

pre pattern that tells you where should the eye go how many should there be where should the mouth be how many heads should a planarian have these are these are states of distributed bi electrical circuits that are now visible you can you can you can see them better than being able to see them you can change them in this planaria so-called voltage patterns were altered telling it to grow two heads we don't touch the hardware we don't we don't change the DNA we don't tell any of the stem cells what to do all of that

is taken care of by the system all we do is reset the goals you get to make complex outcomes that nobody could make before because the micromanagement is too is too hard and this is where we hit on a bar time shift because normally in biology and Medicine we focus on genes and molecules and managing those you could say we focus on managing the hardware of cells whereas Michael Lan focuses on the software all there is is molecular reactions so if you want uh an extra eye or a worm with two heads just manage the

the molecular reactions that's all you get that that's the conventional View and the implications of that conventional view is that you really need to micromanage all of those chemical reactions to get what you want now that's incredibly difficult because because the mapping backwards if I know what I want well what should the what should every Gene be doing what should every cell be doing what should every molecule be doing that's really hard in developmental biology there's the study of morphogenesis the origin of forms the way how cells organize themselves into shapes and tissues now many

aspects of morphogenesis still remain a mystery to us no individual cell knows what a finger is or how many fingers you're supposed to have but the collective absolutely does and that's how you get get to things like a salamander limb where you you cut you you can amputate anywhere along the length and the cells will build exactly the right thing and then the most amazing part is they stop when they're done pondering upon Lavin's work I asked him where he thinks that the memory of forms in nature is stored Plato and Pythagoras had this idea

that uh there is a platonic space of these kinds of things of these forms and I think we need to take this seriously in that I think one thing that evolution is doing is searching through pointers into that space so I think evolution is not really searching the uh the the horrible complex rugged space of micro states of all the hardware phenotypic space I think evolution is actually looking for pointers into this this incredible uh space of forms that already pre-exist even bacterial biofilms you know the very first uh kind of attempts at multicellularity a

kind of weak form of multicellularity we're already using electrical Network even even individual bacteria living in the network form a kind of collective that does better than the individuals do because it uses bi electricity to integrate information across space and time a wonderful practical implication of Michael Levan's work could be that in the future we can also in humans tell cancer cells to stop which would be fantastic but what I find just as important is a deep philosophical lesson his work teaches us namely that we should do away with all categorical thinking around Consciousness and

intelligence because it's all a matter of scaling according to leevin and we ourselves could be part of a larger intelligence without knowing it I I certainly think it's possible um absolutely I think it's possible but I think it's very uh important that we don't fall into the Trap of mean some people say that's ridiculous that's a misuse of the terms and other people say oh absolutely the whole universe is one giant mind you know guy like okay both of those are are hypotheses you cannot decide between those just by sitting in your armchair and deciding

which one feels good to you you need to do experiments and obviously on the higher level you know on the level of the universe the level of the solar system I can think of how to do those experiments but they're not practical I wrot a little you know dialogue at one point between two two neurons in the in a brain and one neuron one neuron is is kind of like the hard-headed materialist and he says look uh we we live in a cold unfeeling Universe there's no pattern no no nobody cares about us it's just

you know it's just each one of us is is alone and the other one's a little bit mystical and he says I don't know I have this feeling that that we're part of some system that actually you know the universe actually learns and wants things and does and and that you're crazy you know no I don't know there's like these patterns of you know of propagate like back propagating error and whatever and sometimes you know the universe punishes me and and whatever but I feel like there's a there's a purpose behind it and he Ah

that's crazy so so right so from the perspective of of of in that system the the mystical neuron is is correct welcome to the asenia foundation's YouTube channel if you're new to our Channel may I ask you to like And subscribe as it really helps us to grow and make more content like this now without further Ado here is my full conversation with Michael Levan you've been able to tell organisms paneria and to grow two hats right uh by cracking the code um and not only did they listen to that by growing two hats but

the next generation of flaria also did that and there just to me doesn't seem to be a very good explanation for that if we are into of this this um old Paradigm of evolution life and biochemistry so my sort of main question to you as as an opening question is what sort of new paradigm are you working on or what do you think will be the next Paradigm to make sense of what you and your colleagues at your lab are doing I think I think it's important to to say that uh any phenomenon including the

two-head thing and anything else if a after it's done after somebody chose how to do it you can always tell a standard uh molecular story uh that is after the fact you can always Zoom down to the molecular events and say that oh oh well there's nothing special here it's just molecules doing what they do it's normal laws of chemistry of course everything everything works and so we have to think about the difference between uh telling causal stories backwards trying to explain what happened versus the kinds of paradigms that suggest new experiments and take science

forward so this this this gets into this issue of of for example reductionism and the determinism and things like that a very simple example is um if if there was a chess game that was played and you can look backwards and a A reductionist can say look I tracked all the electrons and all the protons and everything worked exactly to the laws of physics it was completely determined there's there's nothing surprising here and that's true but how much is that story going to help you play the next chess game right not not at all these

are you know looking backwards you can always tell a conventional story and sometimes especially with our work you would have to put in some I Cycles um so to speak you would have to you know put in some extra bells and whistles to make it fit the convention Paradigm but but you always can and so if you know when when somebody says look you know this is this is this is what they've shown somebody else says oh I'm sure there's a molecular explanation for it of course there's a molecular explanation for it there's always a

a a a reductive explanation for it afterwards the real question is what was the Paradigm that got you to do the experiment in the first place and why wasn't it done before and so you know the first two-headed worms I mean there are many ways to make two-headed worms and the first two-headed worms were seen at the turn of the last century so 190 three or so the question though is since between then and uh 2009 when we did it why did nobody recut them that's to my knowledge nobody ever recut them and the reason

you wouldn't recut them is because if you have a standard uh belief that your genome is what determines your morphology and the genome hadn't been touched well then why would you do it of course it's good you have an expectation of course they will go back to being one-headed once you cut off the ectopic head and our our Paradigm predicted the opposite because we were thinking of the the uh the bi the bioelectric circuits that that store this information as a kind of memory and the most obvious thing about memory is that it keeps State

either for either temporarily or or permanently but it keeps it keeps state so so so I think that's so I think that's important right it's it's not just the phenotype I mean that anybody can get a double head there's many ways to get a double head um the the the new so just to answer your your your specific question the new the new I I think what we have here as a new paradigm is this idea that uh everything exists on a spectrum of cognition and that uh you cannot say where something is on that

Spectrum from philosophical commitments you have to do experiments so so when we ask what is the correct way to uh communicate or to understand uh cells to communicate with or to understand cells you don't know that that the standard machine Paradigm meaning with the Paradigm for simple machines you know this this kind of bottomup emergence complexity dynamical systems theory you don't know that those are your best tools until you try it could well be and I think it is the case that that a lot of these systems uh uh subcellular molecular Network cells tissues organs

are actually uh much further along the Spectrum and that means that what you're really doing if you want to uh predict and control and build new things is you have to uh use tools from behavior from the from the behavioral sciences and so now now this is this is interesting thing I mean this is this is this in in drawing this spectrum and and and in being really really open to what kind of systems might be anywhere on the Spectrum this this goes against two the two both of the two dominant paradigms because you've got

the kind of the molecular biology Paradigm which says look it's crazy to to to um kind of uh to to to it's a category mistake to think that these kind of Behavioral slal um kind of Primitives you know go and and memories and things like that it's it's it's crazy to attribute those to to to these simple systems and that what you need to do is everything needs to be cached out in the in the laws of chemistry because these are simple machines okay then then then of course there's the there's the organicist propos which

has been fighting this this view for you know for for probably centuries which says that no look living systems are special they cannot be treated as the systs Mach simple machines they are essentially different there's something categorically different about living systems and maybe maybe living systems are cognitive systems but but the point is these are these are not machines okay this is this and so and so the perspective that I bring is is uh goes counter both of those because I'm saying no actually there is no categorical difference what there is is a spectrum of

scaling and that even the simplest things have certain properties that might be usefully exploited by a cognitive approach and that all what we need are not stories of categorical differences and and uh you know an emergence what we need are stories of scaling we need to understand how the very simple competencies of even you know even simple molecular networks are capable of learning and things like that and and so what we what we really need to understand is how cognition scales up and so that's that's the idea the idea that we need to be free

of these old categories right that that's another thing you know when people say it's a category mistake well categories aren't you know fixed by philosophy and then we hold on to them I mean it used to be a categorical mistake to a category error to think that um the same laws could apply to the motion of physical objects here on Earth and motions up in the heavens right those that that was cons at one point that was considered a terrible category error so categories need to categories need to to to shift the science the science

needs to tell us what the categories are not the other way around and I think that yeah both of these stories are now crumbling the you know the these these categorical stories are crumbling even point to sort of the Mind Body story sort of that dichotomy is also sort of not very useful right um and a question I have Michael but this is again speaking in dichotomies and in and in sort of categories but just to to help viewers understand we usually have sort of this Hardware software sort of few one things in our era

of computing just the metaphors we we like to use right and then sort of your DNA would be the somehow the software and other proteins within the cell would be the hardware right and you also sort of crumble that whole distinction right and you is it right that and I was thinking would a way one way of looking at it be that that we have this analog uh way of of thinking in the sense that I mean in the old days the analog computers the sort of the algorithm was in the hardware the configuration of

the hardware right so in this compared to the cell um it's sort of the the the how the proteins on configured it's a hardware story and and you sort of have discovered no there is software but it's not sort of on that Hardware level so it's a different Paradigm of computing within the cell is that a a way of looking at it is that a useful way of thinking yeah I mean I I I think the first thing to to nail down is the this idea that uh all of these things that we talk about

Hardware software uh instructions representations touring machines all of this stuff the these are not claims about the real system you know people always say well some people say that well you know living things are not computers they're not machines right because nothing is anything all of these things are frames that we take you know the when when when when you adopt a computationalist perspective or some other perspective you're not really saying that's what this thing really is okay nothing really is anything what you're saying is this is the frame that I'm going to take in

interacting with that system let's see how far it gets us so so that that's all we're really arguing about and and uh the great thing about that is that multiple observers can have and so this is something that Josh bong and I developed which is which is a perspector called poly Computing which is this idea that the same physical event can be seen as as wildly different computations by different observers who interpret the same the the same exact physical event in different ways and so uh multiple observers can be looking at something and say oh

I I I see this is this is clearly a touring machine and somebody else says well actually I see it differently and because of that I'm able to do something else that you're not able to do this is this is fundamentally important also in medicine you know you you you you want your orthopedic surgeon to think that you're a mechanical machine right if you have an orthopedic surgeon that doesn't think that your body is a mechanical machine he's not going to get very far but you don't want your psychotherapist to think that you're a mechanical

machine and so depending on what you're trying to do different frames uh are useful so the molecular biologists absolutely have useful um have gotten useful mileage out of this idea that the DNA is the software yes you can do certain things that way but that's certainly not the only perspective and there's there's another very useful perspective in which what the DNA does is simply Define the hardware of the cell so the DNA tells you which proteins Which ion channels which promoters you're going to have which determines that those determine the connection between the biochemical reactions

that happen in the cell and so on um that's that that Parts the the hardware and now you can ask okay so what is the physiological software that runs on it what is the uh what's the plasticity of it how do you communicate with it what what is it paying attention to you know what are the Cog of competencies of this of the software and I also want to be really clear that by saying hardware and software I absolutely don't mean that living things are well captured by the current architectures that we have you know

this is I'm I I I I do not hold to a computation lless perspective where I think everything is a computation and that's all that's all that there is it's I do think it's a useful perspective in some cases but it certainly doesn't capture everything that that that we need to do and also I think that people are um too hyperfocused on the kind of the modern um you know whatever the the current embodiments of these things so people are thinking about you know current language models and and they say AI can or can't do

this and then what they have in mind are these language models well that's just I mean that's that's a tiny sliver of of the space of possibilities in computer science and the same thing with with with the idea of software and these architectures you know of of course living things are not a vono and computer uh but there are interesting aspects of um of the software Paradigm that are really I think useful in in in biology for example uh people who who program have a have a have an interesting um dualism that they buy into

which means that they they have this Paradigm that your algorithm your algorithm determines what the computer does it makes the electrons dance now you can imagine um and I I wrote this this this um you know this little little just for a joke I wrote this little Dialogue on my blog where this this this candidate comes to a software company for a job you know he's interviewing for a job and and the guy says so uh you know you want to you're going to code up some cool stuff for us and he says look uh

the computer is a physical device it obeys the laws of of of physics what what's this what's this magical algorithm that you think is going to is going to push around the atoms no they do what they do you know and and so right so so so so it's not that that's wrong but if you have that perspective are you going to code anything useful you're not right and so and so you need to really believe that these these these higher levels have a have a causal functionality to them otherwise nothing happens and I think

you know they very naturally people very naturally do believe that software actually does in some in some important way make the electrons dance right so that is really critical also in the kinds of things that that that we work on you know this this idea that you can't just be stuck at the lowest level and just to this is all just very fascinating stuff but to reflect on sort of the Mind Body and also on this this Hardware software though it is sort of a limiting um um um metaphor I know that but um how

important you think is sort of are these bioelectric fields within for instance our own body to what degree sort of is my makeup done by just jeans and sort of totally explainable from sort of molecular way of looking at it and to what degree do we have to sort of invoke the stuff you are working on is there sort of something you can say about that for people who are new to this sure yeah um again keep in mind that the idea isn't just to explain looking backwards you can always explain things at the lowest

level in fact when when when I've sometimes when I discuss this with people and I say look reductionism you know you like you want to reduce it to the lower level and then and then you could say ah so you want to be talking about quantum foam right I mean reduce everything to Quantum for they say no no no that's crazy you know chemistry not not okay so so that tells you right there you're not really trying to reduce you've picked a level and and and you sort of stay stay with it right so so

to me the idea isn't to explain because you can often explain things at level after the fact the idea is looking forward so the idea is okay what Insight have we gained to enable us to do the next grade thing which means new experiments new research program new new biomedical treatments new synthetic bioengineering to improve you know everyone's life um which which of these paradigms helped you make the next great discovery right and it's and so so I think that uh what's what's really important here is that if you if you do have this Paradigm

that um it's it's the genetics that determine everything and and I mean there are many people besides me I'm not the certainly not the only person to be saying that that you know the genetics is not a a hard determinist kind of thing that many people saying that but what's what I think is interesting and different about our work is that if you if if if you think about Evolution as not producing solutions to specific problems which is the typical way of thinking about um you know genetic search is that is that you're going to

find the genome that produces something that has the highest Fitness in the environment and that's it you're looking for solution to a specific Fitness function what I what I think evolution is actually doing is is uh developing problem solving agents and I don't just mean clever animals that do things in three-dimensional space I mean molecular networks that can learn cells and tissues that solve problems in physiological space and the anatomical space and so on and so if if you if you have this idea that uh that that what what comes out of of of evolution

are these uh problem solving agents with competencies with certain degrees of intelligence that is up to us to discover that opens up a whole new frontier of research programs around how do we detect unconventional intelligence in operating in different problem spaces how do we communicate with it in ways that are effective to change their behavior right so so for biomedical purposes and so on and uh of course of course it will also have ethical implications in terms of being making us much more open-minded about uh what what kinds of things deserve care and compassion and

so on um and uh that that's that's the idea that that that through the you know bioelectricity bioelectricity isn't Magic right there there's probably my my guess is that throughout the Universe there's lots of different physical modalities that can serve as this kind of cognitive glue that I think what's important about bi electricity here on Earth but but here what I think Evolution has done from that's how you call it Michael cognitive glue sort of cognitive glue is how you bu bio electric bio electricity is sort of cognitive glue it's a kind of cognitive glue

and what I mean by that is that I think it would be a real mistake to see the bioelectricity as just another piece of uh chemistry and physics that you need to add to okay we already know about you know methylation and all you know all these other chemical gradients and all this stuff okay fine so here are some voltage gradients I think I I you could do that and some people do that but but I think that's a big mistake because because what's actually important about bioelectricity is that it functions as a cognitive glue

and what I mean by that is you know think think of think of how it works in the brain we are a a giant collection of uh an enormous collection of individual neurons and yet those neurons come together in some way to give rise to a new individual that has goals preferences memories and capacities that none of the individual neurons have so we exist because the neuron we we as as human Minds exist because the individual neurons and other cells in our brain are tied together via some uh U interaction policies and some mechanisms that

underly those policies that enable them to function as a as a whole as a collective that is more than the sum of its parts that's the cognitive glue now in the brain it's not controversial we know what it is in the brain the cognitive glue is the electrophysiology that runs the electrical networks in our brain so so my point is that Evolution did not uh wait until brains evolve to make use of this this is extremely ancient and even even bacterial biofilms you know the very first uh kind of attempt at multicellularity at a kind

of weak form of multicellularity we're already using electrical networks and there's this very nice data on this from um goral and other other people showing that uh even even individual bacteria living in a network form a kind of collective that does better than the individuals do because it uses bioelectricity to integrate information across space and time so they use information ac they they use bi electricity to share nutrient information across the across the The Colony and they use it to organize feeding cycles as a as a kind of biorhythm inside the that's amazing yeah we

knew already of course about molecular signaling but that's sort of the the the the hard way way of secreting molecules that sort of inform the the sort of a colony of of bacteria right but this is these fields I'm just so um and this is really a l lay question but I me what is needed what's the substrate needed for these fields sort of ions or what is needed sort of create to create these fields yeah yeah so again I'll point out that that the cognitive glue function could in theory be be um implemented by

many other mechanisms you know maybe out there in the wide universe there are other ways to do it probably there is but but but but what I like about bi electricity is that it shows us one example one clear example of how it happens and I'll and I'll and I'll I'll tell you what uh how how I see this happening so so a single uh let let's just talk about single cells um a single cell and and by the way I don't believe that single cells are the origin of of cognition I think it goes

well below that but let's just serious okay let's go into that as well yeah yeah we can yeah we we can get into yeah no I think by the I think cells are already you know kind of high on that on that Spectrum so so so here's how it works if you have a single cell within the cell membrane you will have these little proteins called ion channels and these proteins have the ability to move uh charged molecules you know so pottassium sodium things like that in and out of the cell that gives rise to

a voltage gradient between the inside and the outside of the cell um some of those channels this is this is quite clever Evolution uh discovered that some of those channels can themselves be voltage sensitive which which means that what you really have is a voltage gated current conductance that's an that's a transistor that's another way of saying transistor and so that that gives the cell something very interesting which is a kind of memory or historicity because if something happens and there's a particular voltage state that changes the channel which then changes the voltage state which

changes the ch right and so and so you have this feedback loop that can hold memories and it can make decisions and so on it's already even right there alone with that with that just with that one ey on channel protein you have some very interesting uh um you know atoms so to speak of of cognition already right there but when you put cells into networks what happens is that the electrical properties of one cell can influence the electrical properties of another cell just like in the brain and and and the whole thing propagates uh

you know you have information flow throughout the throughout the network now the cognitive glue portion comes in in the following way uh one way that cells can connect uh can pass information is to send each other chemical signals now if I secr so so you cells I secrete a chemical signal it goes floats over to you you receive it with a receptor it's very easy for you to recognize that it came from outside and so you know it's a signal that came from outside and you can uh ignore it you can pay attention to it

you can interpret it in various ways you can believe it not believe it whatever but uh Evolution found something very cool which is called a gap Junction and the Gap Junction is a set of proteins on either side of two adjacent cells that have they make a little like a little conduit you can imagine two two submarines docking together via a little hatch and so what happens when the Gap Junctions um open between cells is that small molecules freely pass from one to the other okay now this this makes for a very uh very interesting

thing because let's say you let's say you have two cells uh one cell gets poked there's a there's a there's a damage event one cell gets poked with something there's calcium Spike you know that that serves as a a memory of the damage that calcium Spike now propagates to the next cell now the next cell can't tell that that memory of damage belongs to somebody else because a calcium spike is a calcium spike it it does not have metadata on it that say who does who who whose memory this is so if you have a

bunch of cells that are connected with gap Junctions what you find is that they have a really hard time separating which memories are yours and which memories are mine it means that we are now in a mind meld that makes it really difficult not impossible but really difficult for us to have our own independent thoughts because because we are now sharing them it's literally literally this kind of crazy mind melt so so so so the way the cognitive glue works a few a few things happen first of all it's a it's a it's a mind

meld with respect to a lot of different memories that's the first thing so now it's not me and you now it's just we as far as memories that Define a cognitive agent so you can already see how the scale up you know the scaling is starting to happen the second thing is that because lots of cells are joining together they their um physical area of concern becomes bigger so that now things that happen over here matter to me over here right and and because signals take time to pass through the Gap Jes now even the

temporal aspect is thicker because now now it's now now now there's a you know it takes there's a now moment which is for the collective which is which has a minimum width to it right which is certainly the case in neural systems as well also there's the notion of stress propagation so if if uh if if we think about stress as just the distance from the target of some homeostatic this is still very fast right Michael if you talk like you talk about this temporal window of process sort of this this this field among cells

that's still very fast right just just Electronics uh minutes minutes oh minutes it can take a while all right minutes because well I mean so so so one thing you can do is you can take any Neuroscience paper and just do a simple find replace you take anytime it says neuron you just say cell and anytime it says millisecond you say minutes and then you got yourself a nice developmental biology because what evolution did is just turn the same it just pivoted the exact same mechanisms from from navigating uh anatomical space to navigating three-dimensional space

and then it had to crank up time because because now you have to you have muscles and nerves you have to run around and not get eaten so so so you know developmental processes are slow you have you have weeks and or or months to form a certain embryo or to regenerate or whatever you you have time so all of that all of that stuff is slower uh so so so what's happening here is that uh you have the you have your cell cell sheet connected by Gap Junctions now now imagine if all this all

of the um all of the cells are trying to maintain some kind of homeostatic State let's say pH or hunger level or something right there's a there's a Target set point and they're trying to maintain it in a cycle well stress is basically the the measure of error how far away from your goal you are right if all you do is keep your own stress then everybody has their own little battle going on and everybody's doing their best but they're all it's just a collection of Parts but now imagine that you're starting to leak stress

you're starting to leak stress through Gap Junctions through other mechanisms externally whatever if if I'm sitting next to you and you are super stressed out and you're leaking your stress now I'm a little bit stressed and what happens is that in order for me to reduce my stress one of the things I might want to do is help you not because I'm altruistic but because you're stressing me out and and if I reduce yours so so what's happening here is by but this kind of stress sharing means that your problems become my problems and things

that are happening over there that are making a you know some sort of a mess over here are now suddenly my concern so what it does so so what what these kind of mechanisms do is they enlarge the cognitive like the cognitive lone is the the the size of the space and time bubble that you care about what are you what are you actually actively working to to fix and as it grows your your area of concern become bigger than a single cell it's now the whole thing now now having having done all of that

you now realize that uh these Gap Junctions are not just uh empty pores so they actually have a have state to the meaning they can open and close based on things that happen and now you have a fairly sophisticated set of connections and the collective the the the the whole tissue Collective has more computational capacity than any of the individual cells meaning they can remember more things so now so now what you've got is you've got a collective whereas before it was every cell working for itself to optimize metabolic State you know so you're working

in metabolic space Maybe in transcriptional space things like that all of a sudden now you've got a collection of cells that has a much larger memory and computational capacity and it cares about large scale things like maybe curvature right the overall curvature stress or maybe um you know other aspect the size of the collective or now you've got now you've got a larger Collective that can operate in anatomical space and so this is how intelligence levels up no no individual cell knows what a finger is or how many fingers you're supposed to have but the

collective absolutely does and that's how you get to things like a salamander limb where you you cut you you can amputate anywhere along the length and the cells will build exactly the right thing and then the most amazing part is they stop when they're done because that error that Delta to the final pattern has been reduced to acceptable levels but that the pattern that they're the set point that they're trying to recreate was something that no individual cell could do and it's not in the in the genome none of that stuff is in the genome

what's in the genome is protein where is it because one way to look at when you say that that's the amazing part when it knows where to stop right when my cells know the body is sort of complete and you also talk about you even sometimes in your writing use the word story right there's this story that sort of collective story of of of what the the um correct outcome look like looks like where is that story stored or written because it seems to be accessed a new when sort of we have an and and

a new organism I mean this was this is what truly puzzles me yeah uh okay well the first thing I'll say is that if if if we really knew the answers to all those questions we would be you know the world would be unrecognizable so I'm not I'm not going to I'm not going to claim to give you the final answer but but I can tell you what we know now uh there are two questions here that you're really asking where is it where and how is it stored and where does it come from and

what happens in new organisms okay those are two those are two very related but quite different questions on the first one we we kind of actually know we can we've developed tools uh to to to see it so we can read and write and and and to some extent people have you know in the biochemic in the biochemical world people have talked about chemical pre patterns and and morphen gradients I mean that that's that those are all conventional Stories We we've added to this this idea that we can actually see these bi electrical patterns so

so we should we have um tools much like neuroscientists do in the brain to read memories out of the brain right so people are are working on that so so we have now tools where you can literally see for example the electric face this this this P this pre pattern that tells you where should the eye go how many should there be where should the mouth be how many heads should a planarian have these are these are states of distributed bi electrical circuits that are now visible you can you can you can see them and

better than being able to see them you can change them that's that's that's what really that's what makes it um clear that they really exist because another way of saying it would be just to go back to your your original question another way of saying it would be forget all this all this kind of cognitive talk all there is is molecular reactions so if you want an extra eye or a worm with two heads just manage the electrical reactions that's all you get the the molecular reactions that's all you get that that's the conventional View

and the implications of that conventional view is that you really need to micromanage all of those chemical reactions to get what you want now that's incredibly difficult because because the mapping backwards if I know what I want well what should the what should every Gene be doing what should every cell be doing what should every molecule be doing that's really hard but but my claim is the reason the reason that I think uh my my claim about gold directedness and intelligence in these tissues is true is not just because we see these patterns because you

can also explain them at the molecular level the reason is is that by targeting those those patterns and interventions by using the techniques of of of changing memories you get to make complex outcomes that nobody could make before because the micromanagement is too is too hard right so so so so think about you know the the a simple Neuroscience example of this is you you have a rat and you want the rat to do a circus trick you know sit on a little bicycle or something and somebody says okay I can see this thing is

a is a is a deterministic uh molecular system so here's what we're going to do we're going to find out all the different muscle motions that need to happen to make the body do it then we're going to track back to all the nerves that need to be activated then we're going to track it back all the way up into the brain and then we're going to go through and then I'm going to tell you exactly which pixels on the retina you need to activate with stimuli to make it do the thing now in in

theory that's possible in reality the sun will burn out before you figure out any of that but someone who understands what the rat really is can say no no no you don't need to do any of that you train the rat what do you mean you train the rat top down control you co you collaborate with the rat you incentivize the rat by um and and and and what's Magic about that is that it then pushes uh that that that view of it that top down control pushes all the complexity onto the system itself you

don't need to figure out which neurons need to fire the rats already very good at knowing what what needs to be done given certain top level controls it's a communication problem it's not a micromanagement problem so so so so the same thing the same thing is going on here these um these patterns we we know we know this is a this is a a goal directed phenomenon because we change the goals we don't touch the hardware we don't we don't change the DNA we don't tell any of the stem cells what to do all of

that is taken care of by the system all we do is reset the goals that's the beauty of goal driven systems is that you can communicate with them you can reset the goals maybe they learn some some cells and tissues Lear so so so so we know where the information is stored okay at least partially it's stored it's it's represented by a distribution of voltage gradients across the tissue and those gradients serve as the reference point for what to build and when to stop okay that Michael how how robust is these uh are these states

and how much of the organism has to stay intact for this information to sort of just uh be accessible again yeah in case of P areas it seems that sort of just cutting off one hat sort of didn't sort of the information was still there to grow the to know know again it should grow two heads right yeah oh oh planarian can be cut into 275 pieces so so in in plaria it is incredibly uh holographic it's it's the circuit rescales so the interesting thing about that circuit is when you cut it into pieces the

the circuit rescales and so um you know in other organisms it's not quite that robust but still you know even even even mamalian embryos including humans in the early stages you can cut them into pieces and you get monozygotic twins and triplets and quadruplets you know you don't get half bodies at least at at early stages so so there is a lot of rescaling of this kind of thing going on um the second question you asked though is is much harder and much deeper which is where do they come from in the first place and

so so let's just let's just look at that for a second where do they come from and what happens in New in in new ones um overall the question of where things come from is uh is is needs to be treated differently than than a lot of the other questions we deal with because uh and and mathematicians have been on top of this for for a long time starting with Plato and Pythagoras and probably long before that which is that there are many things in the world that are kind of free lunches in in the

physics sense that uh Evolution takes advantage of for example just is a stupid example let's say um let's say in a certain world it's very simple there's an evolution going on and the most fit thing in this world is a certain kind of triangle certain certain shape of triangle so you know Evolution goes and and lots of it does lots of generations and it gets the first angle and then lots more Generations go by and it gets the second Angle now here's now here's the magic part you don't need to do anymore you know what

the third angle is yeah you get it for free isn't that amazing isn't it this is not philosophical fluff this is incredibly practical Evolution gets to save onethird of its time here because of this incredible free gift from geometry the fact that once you know two angles in flat space you know the third so so now there's tons of this stuff tons of it you know when when you make a when you make a voltage sensitive Gap Junction um ION channel rather you have a transistor when you have a couple of transistors you have a

logic gate when you have logic gates you have truth tables and and you can build you know whatever functions you want do you need to evolve the elements of that truth table no you get it for free do you need to evolve the fact that the nand the not and gate is is universal it's special no you get that for free where does it come from right where do these things live the you know the rules of comp the laws of computation the laws of geometry the the laws of mathematics where where do they live

right so so so you know so so Plato and Pythagoras had this idea that uh there is a platonic space of these kinds of things of these forms and I think we need to take this seriously in that I think one thing that evolution is doing is searching through pointers into that space so I think evolution is not really searching the uh the the horrible complex rugged space of micro states of all the hardware phenotypic space I think evolution is actually looking for pointers into uh it's it's it's looking for physical embodiments that that basically

uh take advantage of this this incredible uh space of forms that already pre-exist I mean in a certain sense it's the physics and the chemistry of the physical world in another sense these things you know the laws of number Theory and all this other stuff they would be exactly the same if the universe was different right if all the laws of physics were changed at the Big Bang this I I I think and I mean I'm certainly not not an expert on the mathematics of it but but to my understanding all of those things would

still be exactly as true so so there's something else here that isn't captured by our our typical accounting of what happen s in the in the physical you know kind of the physical world and for this reason and and so I think that is what evolution is doing it's it's it's um finding Hardware that is exploiting some of these things and therefore and so now now we come to the to the to the final step is what happens in novel organisms specifically I the reason we make xenobots and anthr robots and and various other things

all all these synthetic beings is to explore the latent space around the one point that normal development gives us so normal development gives us here's a here's a frog embryo or here's um uh here's a an oak tree leaf and this is what it looks like here's the shape here's that and we say okay we know we know what the oak genome does it makes these nice flat green leaves fine Well turns out that there's a there's a there's a there's a space of of there's a latent space of possibilities which if if uh which

we wouldn't know about if we hadn't seen that there's a bioengineer a non-human bioengineer which is a wasp and these wasps will will lay an egg on these leaves and the signals from the from the eggs will uh basically hack this the leaf cells to build this incredible like red spiky yellow looking Brown thing this these Galls these amazingly shaped Galls we would have no idea that the normal cells the genetically normal cells of the plant are able to build this because all we ever see is one point we see the normal default behavior in

fact it's so consistent it's robust that we we really think that's that's all it does but with a little bit of with a little bit of prompting and I use that word advisedly with a little bit of of prompting you find out that oh no actually there are these other shapes around and so who knows what else is there so so this is this is what I think of these synthetic um when when I call you know people say why do you call them xenobots they're they're living organisms or they're you know they animal Caps

or they something else the reason I call them uh I call them xenobots is because I view this as a vehicle for exploring the latent space I want to see what else what other forms and functions are around in the space around this uh this this this this you know what what the genomic Hardware can do and so now so now this is I think what is what is the just the beest beginnings of the answer to your question when we make anthr robots you know these are these are um little little biobots that are

made from from tracheal cells of of adult human donors they have all kinds of cool properties they not only do they form interesting shape and they they swim around on their own they have they're self motile but they have new capabilities like for example they um uh they they repair neural wounds and uh in in other they they take neural neural neural scratches and and they heal the sides together and they do some other stuff that um that nobody's seen yet it isn't published yet but hopefully it will be later this year we'll have some

some cool stories about them and so so where does all this come from or in the case of zenbot kinematic self-replication the fact that they can make copies of themselves by coalescing material from that they find outside in the environment I I I think that what we're seeing here is uh the ability of biological Hardware to explore and take advantage of other uh other regions of this this platonic space of of possibilities um that um that that allow them to solve problems and so that that's that's what I think is is is happening here that

when you say where is it encoded where does it come from we have a couple of uh ingredients to that to the answer you have the genomic Hardware which is well this is it's it's this way because that's where it was encoded so some of it is certainly genomically encoded some of it is emergent in the standard version of a complexity and emergence you have you have Simple Rules like cellular aom they do stuff and then boom there's in so some some of them are some of them are are Hardware encoded some of them are

emergent uh others are really drawing from this uh from the space of wherever the laws of mathematics and computation come from right and others are generated on the Fly by the cognitive processes of the system so these are patterns much like you know where do where do the patterns in your in your brain come from well they come from you trying to organize your experience they come on the fly from you as sense make as a sense making agent so th those are the ingredients and where do they come from yeah amazing and so the

um the standard few because that's sort of also um the amazing implication for for instance cancer could be that we now have this Paradigm of micromanaging that disease right but just targeting it on a hardware level whereas if we could do the goal setting on the the cellular level saying to these cells okay the new the goal is again to cooperate then that could be a a future C cure of cancer right which could be an implication of Your World totally and and we've done this in the in the Frog model we've already done this

so so we've we've shown that um you can after you can inject human enogen into a tple and they will make tumors and if you do that but you force the cells into proper electrical state with the communicating with their neighbors you prevent tumorogenesis even though the enogen is blazingly strongly expressed and that's not because you don't you don't destroy the enco protein you're not managing the the the cell you know the the the micr level details of of transformation no what you're doing is you're forcing that cell to stay connected to their neighbors and

when when you have a mind meld with a large Collective that all says we are building a nice kidney we are building a skin it's really hard for that cell to have an independent thought of you know what I I my goal is to proliferate as much as I can and crawl away wherever life is but right metastasis so so this is uh we we we can actually see this because the first thing that happens when anagen ACT first thing they do is they cause electrical uh disconnection from their neighbors and as soon as you

disconnected well now you're an amoeba again and as far as your concerned the rest of the body is just external environment so that that border between self and world can shrink and grow right so so we started we started like both our each of our lives but also evolutionarily we started our life as a as a microb as a single cell with a tiny little cognitive lonee we only care about what happens in the cell everything else is is environment and with Evolution and development that cognitive L cone grows and then it could sometimes shrink

again and so so this is so this is this is you know one of the very clear examples of how this stuff uh is not just uh you know Wishful philosophical thinking and you know no it has it has very practical implications that if you follow through you discover Therapeutics and um I heard you say the word mind melt I like I like that a lot because it's a good thing for an organism in a sense that our cells have a mind melt which makes that that my body functions well do you see that possible

on higher levels I mean what are your thoughts on sort of a um a Gaia uh like thoughts on a sort of planetary mind or Universal Minds me do you are you open to ideas like that uh yeah uh so so the fir so two two things I want to say about that one is that it's I I certainly think it's possible um absolutely I think it's possible but I think it's very uh important that we don't fall into the Trap of I mean some people say that's ridiculous that's a misuse of the terms and

other people say oh absolutely the whole universe is one giant mind you know guy like okay both of those are are hypotheses you cannot decide between those just by sitting in your armchair and deciding which one feels good to you you need to do experiments and obviously on the higher level you know on the level of the universe the level of the solar system I can think of how to do those experiments but they're not practical and on the level of ecosystems you know maybe maybe it's practical to to to try some of the stuff

can you train an ecosystem maybe uh but but again all of this is these are empirical questions of how to recognize yeah sorry to interrupt you because you say you can't think of experiments though not practical and this is what a lot of physicists do right they they come up with these but do you have do you have an idea how you could sort of make that testable from your how would that be done sure sure so so so that's one of the things that that our framework is trying to do is is to develop

um General road maps for discovering novel intelligences and one of the simplest ways and this is certainly doesn't uh go all the way but one of the simplest ways to to know when you've communicated with an intelligence is when you can train it right because because simple even Simple Minds have certain kinds of learning that they can do habituation sensitization associative learning prediction and and so on so you can imagine like when somebody says to me um well you know you're your view then then I guess you say even the weather is is is is

you know has a mind well I don't say that it does I say that it might now the question is like could we do an experiment so so for example uh you have a I don't know a tornado or some other kind of weather system do we you know we we could in theory if we had the the the you know the the the tools to do it we could in theory test whether they habituate we could in theory test you know can you train a a dynamical s and you know and that might sound

crazy how how would you train a you know you know a hurricane or something but you know all of us are just uh a kind of dynamical system right now now we are certainly we have some qualities that that hurricanes and things like that don't have but again it's a spectrum it isn't a categorical difference and we are all temporary dynamical systems in in in met in um you know in in metabolic space and things like that and we can we can learn and it's entirely possible that if we had a way to stimulate weather

patterns in a particular way you might find that wow there's there really is some kind of habituation here there some kind of sensitization here so so you could you could now now a couple things to say one is that we've done this with um uh uh Pathways molecular Pathways this is why I was saying earlier that I don't think cells are the bottom level because all it takes is half a dozen genes interacting with each other to show associative learning meaning pavlovian conditioning like that's it you don't need the rest of the cell you don't

need anything else just just from the math alone for ju just a few um interactive subunits can already give you this so could could you know could weather patterns do it could ecosystems do it I mean if I had to bet $10 right now I would say probably they can but we cannot assume that we need to do the experiments and so the experiments are training communication now now now now right and so they the perturbative that that's another thing is you have to do perturbative experiments you cannot just observe because when you observe things

going as you you know somebody did these experiments in the 70s where they made these videos of um triangles and squares and circles moving around right and they would have human volunteers looking at them and they would make up stories they would go oh it's very clear the circle is in love with the triangle and the square is trying to push them apart like it's obvious okay so you you can't know any of that from just observing if you want to test the elligence of your system you have to perturb it you have to put

barriers between it and its goals and you have to see how clever is it to pursue itly you have to interface to the system so so the last and and the final thing about that is um it's all well and good when you're testing these uh kind of primitive systems these protoc ogni of systems but one thing that people often ask is well what if we too are part of a some sort of giant intelligence right so so you know so ga so maybe you know maybe our or our society or or the biosphere or

something is this like giant intelligence po quite quite possibly um I suspect that it isn't possible to get um um conclusive evidence of it I I'm going to guess there's some kind of girdle theorem or something like that around the fact that you really can't know when you are part of a bigger system that's so much smarter than you I think it's probably impossible to know but I do think that we can gain evidence of it and what would that look like right what what what would that evidence look like so I have a feeling

and again this is just you know me going much further than we have rigorous mathematics around I think what it would look like is synchronicity I think that what you would see is events that do not look causally connected at your level but they are collected but they are connected by meaning they are connected by um some kind of links uh in in from the perspective of a cognitive Observer and those kinds of things that don't look like well the the physics don't look like they should they should you know connect together but they do

connect um on a meaning level I think that's what it would look like you know if if we if we were part of you know we would see patterns in the world whether they be you know social patterns Financial patterns uh you know movements of matter and energy in the ecosystem whatever that would have that would not have a neat um uh the explanation at the at the physical level but they would make sense uh from from some from some you know some some higher level Observer point of view and so you know this is

like uh uh I I I wrote a little you know dialogue at one point between two two neurons in the in a brain and one neuron one neuron is is kind of like the hard-headed materialist and he says look uh we we live in a cold uning Universe there's no pattern no no nobody cares about us it's just you know it's just each one of us is is alone and the other one's a little bit mystical and he says I don't know I have this feeling that that we're part of some system that actually you

know the universe actually learns and wants wants things and does and and you're crazy you know all the no I don't know there's like these patterns of you know of propagate like back propagating error and whatever and sometimes you know the universe punishes me and and whatever but I feel like there's a there's a purpose behind it and he's Ah that's crazy so so right so from the perspective of of of in that system the the mystical neuron is is correct because in fact they are part of a larger Sy they are part of a

larger system that has a you know that has a global agenda that that neither of them could possibly comprehend but but that's you know it's an experimental thing I think to gain evidence for it I like your openness and your rigorousness that that of you have to sort of avoid sort of um U going astray with these wild ideas not making them testable of course and there's one sort of deep deep I think a bit philosophical question I had I I read in your AI paper that I like the met you said um and when

we rais a child uh I was now threeyear old but like two years ago he would like just Babble I I I would know he didn't understand the meaning of those words but he was just trying and then somehow he came to meaning right so you could say it came from syntactical uh playing around with words to meaning and then you in your paper say maybe that could be the same in AI I was just wondering what your thoughts are here because sort of the Chinese room experiment and sort of in philosophy it seemed quite

well established that there's this Great Divide right this this sort of categories of syntax and semantics what are your thoughts here can that be bridged and do you think and and also how would you apply that then to your cosmology in the end because would that mean that uh physics has semantics to it I'm just curious what your ideas here it really triggered something Reading me because you are on to something here but I mean the well the the first thing to say is that uh what I don't mean is that today's AIS are like

human children so after that after that piece came out I got a whole bunch of emails saying you don't know what children are you don't know what AI is okay let's let's be clear I I I know children are I I have to the the modern modern AIS are absolutely not like human children okay let's F fine but but um that that doesn't mean that um uh it isn't incredibly difficult to specify what is it that that that magic oomph that we're supposed to have that that synthetic or hybrid kinds of Agents don't have I

mean a lot of people feel very strongly that there are machines and then there are organisms and okay so so what is that right what what what exactly is that and people have tried to specify this for a long time there I I don't I've never seen a good story of it I think that what we actually have is is is a kind of scaling and um you know in terms of syntax and semantics I think that that is a that is a computationalist framing that is useful for some things but it is not useful

for many other things and I don't believe these are these are real crisp categories at all it starts it starts uh right at the level of remember the Gap Junctions right the the connections between the cells it's already broke the distinction between between data and machine are already broken there because what happens is when ions flow through the Gap Junction they they change the voltage of the other cell which in turn controls whether the Gap Junction is open or closed so what you have here because the Gap Junction is also voltage sensitive it can be

so what you have here is the data the information that comes through the machine changes the physical state of the machine that's it the distinction between between pure data and the hardware that this this unchanging touring machine that just operates on Hardware gone biology doesn't respect that at all so so um I I think I think you know uh semantics is when when we say semantics all we all we are saying is that there is some Observer that is able to squeeze adaptive saliency from whatever it's looking at so there is a physical set of

events let's say my own memories I am an agent that knows how to interpret my own memories to some extent making up a lot of it I mean you know confabulation is not just for language models like we all do it all the time because our memories are very compressed uh uh our our um our engrams are very compressed representations of the past we have to constantly reinterpret them and we add a lot of a lot of stuff to it uh real real agents and real observers are ones that uh can interpret and embellish their

own memories in real time and and are constantly in this process of um sense making storytelling about what they are what the outside world is and so on and so there are multiple Observers that can look at the same physical events and see nothing see something or see something completely different and so so no I don't think there's any any sharp distinction here and we spoke we spoke about intelligence and uh I I've listened to a lot of interviews with you and I know that you're not very keen on sort of questions philosophical questions like

what is consciousness Etc but if I'd ask you what is consciousness for and I'm talking about phenomenological phenomenal phenomenal Consciousness sorry what it is for in what are your thought thoughts here or your latest ideas around Consciousness okay um uh in the last couple of minutes I guess I guess the first the first thing I will say is that I'm not not keen on it I think about this stuff all the time it's just that I I try to spend most of my time uh most of my public dis you know work and like my

official work on things that I think can make a very practical difference right now and I don't I don't yet have a theory of Consciousness a new Theory Of Consciousness to offer I mean I'm working on something but but it's not it's not useful yet so so I'm not I haven't been talking about it but uh I will I will just say this uh Ju Just just two things one is that I think that uh for exactly the same reason that we associate Consciousness with brains we should be very open to the idea of Consciousness

in various other body parts that's the first thing and the fact that you are not conscious of it means nothing because you are also not conscious of my Consciousness you're not conscious of your right you know right hemispheres Consciousness that doesn't mean anything uh but but but for those same reasons in terms of intelligent behavior in terms of all the same mechanisms uh all of that stuff is is throughout the body you so so that's that's the first thing the second thing is that um what is it for well that goes back to our platonic

space concept which is well well what's the you know the the fact that nand is universal or the fact that um uh you know the the angles of a triangle will add up to what's that for so I think we should be really careful in saying that everything is for something from the standard evolutionary point of view something things are not necessarily for anything in that selectionist sense some things are just features of the universe and I think that Consciousness uh roughly tracks um agency which we could you know we'll need another another hour to

kind of talk about what I think that is but but but basically the question of uh what what I think Consciousness is is so so Mark s says that Consciousness is palpated uncertainty about the outside world and I what it feels like to have to make decisions and what do I do next and and the uncertainty I think that's that's that's I like that a lot and I think that's very close I would say that Consciousness is the palpated uncertainty about what your own memories mean so it's a constant real time we we it doesn't

feel like that it's it's it's not how we describe it but I think what it really is under the hood is our constant uh sense making this idea that as an agent it's not just about looking backwards and explaining what happened it's about looking forwards and saying what is the best story I can tell about myself in the outside world that helps me know what what do I do next in this uncertain dangerous world and so that that that that is what I think Consciousness is and I think uh you know we we have there

there are many open questions about why it feels certain ways and whatever but fundamentally uh that that's that is what it's for because because without that process I I don't think anything could survive but also I think it's uh it's an important fundamental property of the universe that you you cannot tell a simple selection of story about it Michael thanks so much and um I just have one and maybe you can answer short and I think you've already given an answer with this last answer you gave me but I thought just the final question for

you how has your work change your your personal life I mean your outlook on life I just heard you say that you're a father of two because I can imagine working on stuff like this it it does something to how you just uh live your life yeah yeah I think I think it does um and I think uh I think uh uh there it has it has ser and I I think we need a lot more time to talk about this but I think it has it has re it has real implications for um for

ethics and for um how you relate to other other beings and how you think about um enlarging your your um your sort of light cone of of of compassion and things like that um yeah and and so yes I definitely have some thoughts about it but I I think we'll need a different for next time thanks so much for your time and I wish you a very good week thank you Michael you so much yeah thank you it was a great discussion thank you thanks bye thanks unfortunately we had to wrap our conversation a bit

abrupt as I only had one hour with Michael Len who has a very full schedule but if you have any further questions um to Michael Evan please leave them in the comment section and I'll do my best to see if I can arrange another conversation with this great mind who is working on a true Paradigm Shift thanks so much for watching and please like and sub subscribe see you in one of our next videos