Could Biological Robots Heal Us from the Inside? | Michael Levin

You know, we like to think we.

Understand what makes something alive. DNA, evolution, natural selection, the usual suspects. But what if I told you that cells from your own trachea sitting in a petri dish right now could spontaneously organize into swimming robots that heal brain tissue? What if frog skin cells, with no genetic modification whatsoever could build copies of themselves from spare parts lying around? This isn't science fiction. This is the work of Michael Levin at Tufts University. And it's completely rewriting the rules of biology.

So we took cells from adult human patients, tracheal epithelial cells. Turns out that they, too come together and form these little motile creatures. We call them ant robots. Those guys have 9,000 differently expressed genes, and they can do cool things like they can heal neural wounds. This is just the tip of the iceberg.

Michael Levin's research challenges our fundamental understanding of what life is and where biological properties emerge from. Michael Levin is a distinguished biologist at Tufts University and director of the Allen Discovery center, whose groundbreaking research on bioelectricity and regenerative biology is reshaping our understanding of how biological systems process information and pursue goals.

Is xenobots, living robots built from frog.

Cells swim around, work together, and reproduce in ways that have never existed on Earth. What does this tell us about consciousness, intelligence, and the nature of life itself?

Professor Michael Levin, welcome to the into the Impossible podcast.

Thank you for having me. It's great to see you.

We have so many questions. We'll run out of time before we run out of questions, I'm sure. But I want to first start with the big picture question. Why is electricity, of all the fundamental forces of nature, nuclear forces, strong and weak gravitation, why is it that electric and not, say, magnetism, plays such an outsized role when we know that electricity and magnetism are unified via Maxwell's equation? So why does electricity play a bigger role than, say, these magnets here that I have on my desk do?

It is certainly the case that living tissue is sensitive to magnetic fields, electromagnetism. Ultra weak photons are important. All of these things are important. I have no idea if biology can harness the stronger than or the weak force. I just don't know.

Maybe.

I don't know. But the special thing about electricity at this point is the following. It is a really convenient modality to serve as what I call cognitive glue. There are other things that do it. There are other things that could do it elsewhere in the universe. I'm sure it's done by other mechanisms if there's life elsewhere. But here's the thing that evolution loves about bioelectricity, it's a very convenient way to make electrical networks out of subunits, so group subunits into bigger things in a way that allows the whole to have goals, memories, preferences, and basically problem solving behavior that the individual pieces don't have. It allows a raising of levels, so to speak.

And we can go into great detail how it does it. But basically the exact same thing that electricity is doing in your brain, which makes you more than the sum of, than just a pile of neurons. It has been doing that for bodies, multicellular bodies, long before neurons ever came on the scene. It's a way to scale up the cognitive light cone of materials.

Basically, we'll get to the cognitive light con because it's impossible for a physicist that's like, you know, bait for a physicist to evoke Einstein and all sorts of other things. But before I get there, you know, when I think about electricity, when I think about it as a physicist, I don't normally associate it with things that are squishy. And unless, you know, somebody threw a toaster in the bathtub, and that would be quite dangerous, obviously we're not recommending that. But talk about how electricity even arises, and on what scales does it manifest? I mean, we don't see little anodes and cathodes or, you know, positive and negative terminals on cells. So how does it instantiate itself in a mechanistic way? And you could be very technical. I mean, my audience is, you know, one of the most magnificent and mindful in the known multiverse. So talk us through. How is a.

How is a cell like a battery or a magnet with these kind of polarities and dipoles and everything associated with it?

Yeah, no, cells absolutely have charged particles that, that typically ions of potassium, sodium chloride, and things like this. And, and the best way to anchor these kinds of discussions is by thinking about what happens in neuroscience. So we understand that our cognition is underwritten by the bioelectricity that operates in individual neurons and then in groups of neurons and other cells in the brain. So what happens is you have cells and in their plasma membrane, which is basically this lipid kind of membrane that's usually a pretty good insulator. What evolution has discovered are special kinds of proteins called ion channels. And these ion channels have really interesting properties where they let certain charged species like potassium chloride, sodium, and so on. They either preferentially let them in or out of the cell. And as a result, what you end up with is a voltage gradient.

Now, the gradient itself might be on the scale of let's say 80 millivolts in neurons, something like that. But because the width across the membrane is actually extremely small, and I don't remember the exact, the exact numbers, but it's basically the actual field across that tiny distance is enormous. And what, what happen is that you have these voltage gradients and two things can happen. One is that the ion channels that regulate the voltage can be themselves voltage sensitive, which means that what you really have is a voltage sensitive current conductance, AKA a transistor. And you can imagine that once you have that, you can do all sorts of cool computations. And then the other thing that happens is that there are these electrical synapses between adjacent cells and those synapses are also can be voltage sensitive. So what you actually have is a network. You have an electrical through which voltage states can propagate so many cells.

I mean, in fact all cells in the body generate these kind of voltage gradients and most of them have the synapses by which they form into networks that process information using this, using this electrical, the differences in electrical potential. And it's exactly what happens in your brain, basically.

Yeah. So we'll get to the brain later. I've got a 3D printed brain that my son made for a conversation with Roger Penrose at one point field here, but I couldn't get it to work. So we'll talk about consciousness later. But I, if we had to quantify it as a circuit, you know, and let's apply Kirchhoff's laws from your native Russia, as I understand, or I guess you were born in the Soviet Union. Is that true?

True. True.

Yeah. So Kirchhoff had these famous laws that governed how current voltage, resistance, etc. Are related in a circuit. Let's walk through it. What kind of, what kind of, you know, potentials are we talking about? Kind of voltages, currents, resistances, conductances and even magnetic fields. If you, if you have any, anything to say about how those might play into understanding the biochemist bioelectricity of life.

Sure, yeah. So. So, so the typical voltage changes that you're talking about in cells are on the scale of tens of millivolts. So, so what you will have are. So, so in the, in the, in the neural case, you have cells that normally sit at about, I don't know, minus 70, let's say minus 80, with the inside being more negative than, than the outside. And then in the non neural bioelectricity, which is the far more ancient version, in typical somatic cells, you will be anywhere from, let's say 10, you know, negative 10 millivol maybe minus 70, minus 80, something like that. So it's the differences in that voltage between a cell and its neighbor that matter. So what cells actually track and interpret are the spatial patterns, the differences across space.

You have a whole bunch of cells and there if you look. And we developed, back in 2000, we developed the first tools to read and write this electrical information outside the brain. Neurosciences have been doing it for a long time. We developed the first molecular tools to literally take a picture and then, and then the voltage gradients and tissues and you can so, so you can see them propagating across distance. And it is, it is those differences that cells read. Now as you, as you mentioned, magnetic fields. So typically of course movement of charges absolutely makes magnetic fields, right. And the magnetic fields in the brain are quite sizable because the voltage spiking is so fast.

So you know, milliseconds, we're talking about milliseconds change. And so that generates a, you know, that DVDT generates a pretty, a pretty good magnetic field which people of course read with, with, you know, var. The non neural bioelectricity that we, we deal with changes very slowly and it's the magnetic fields that are induced there are extremely low. I'm not going to say that cells don't respond to those, but there isn't any evidence that I'm aware of right now that those extremely low level fields. Now having said that, I know I'm going to get in trouble. People are going to yell at me because, because absolutely, like living things do care about magnetic fields. They sense the earth's geomagnetic field, which is about, you know, half a gauss, something like that. So those things are absolutely important.

But the magnetic fields, non neural bioelectricity are unbelievably weak. And I don't know of any evidence that they play a role. The electric changes absolutely play a role.

Are there species that don't have as manifest an importance of bioelectricity in their electrophysiology or lack thereof?

Well, let's see. I mean, so for example, something like C. Elegans, the nematode, I mean they have neurons, so they definitely have neural bioelectricity. But I am not aware of any evidence yet. Now there are people working on this and so this could totally change. But I'm not aware of any stories about the importance of developmental bioelectricity in that model. And my guess is it will change. I think evolution just loves first discovered these things around the time of bacterial biofilms.

So there's a very nice paper by Girl Soell from UCSD who shows like, brain, like, signaling in biofilms. So it's a very ancient phenomenon.

Yeah. I've had on Allison Myotri, who's worked on brain organoids and launched them into space and a little bit about it. But when I look at the, you know, I'm going to keep nerding out about the physics of circuits because I to solder and do stuff, although that would be quite dangerous in the human body. But, you know, looking at voltages at the, you know, tens to hundreds of millivolts and then thinking about, you know, human cells as, you know, not that dissimilar from C. Elegans or something, you know, just scaled up trillions of times. I mean, how is it possible to go from something where, you know, there's essentially no voltage currents, you know, at a measurable level, to actual macroscopic voltages that we can measure quite easily here. I can measure my, my heart rate using my Apple watch. Right.

I mean, I, I know that, that. So. And how long before we have electroencephalograms on our wrist? I have some meditation program that supposedly does that already. So walk us through the kind of reductionist or anti reductionist. How did we go from basically nothing in eukaryotic cell or prokaryotic cells, I think came first. Right. To then these eukaryotic cells and even complex worms and creatures and so forth that have very little electrophysiology to humans, which have huge amounts by comparison, of, you know, milliamps and, and, and millivolts or, you know, intensive volts.

Yeah, well, I mean, to be clear, I'm not aware of any creature, including bacteria and various microbes that don't have these phenomena. These go all the way down. When I mentioned C. Elegans, I simply meant that I wasn't aware of a morphogenetic role in that species for bioelectrics. But, but I mean, the, you know, Koshland did, did the, they did the studies where he said we talk about bacteria as a model ne. You know, basically, if you, if you think about the earliest kind of steps towards life, you know, just imagine you've got some kind of membrane that is the first attempt to separate the inside from outside.

Yeah.

That you've got some kind of barrier that tries to keep the goodies in and the, and the dangerous external world out. Well, as soon as you've done that, as soon as you've, you've segregated molecules, chances Are you're going to have an electrical imbalance because you're keeping something inside. You're, you're, you're concentrating it against the gradient. You're going to have a voltage, voltage imbalance. And when you do, then, then some other very cool things happen as free gift from physics. So, for example, if you have a system like this and you poke it, let's say it gets injured, right? It runs into something that injures and pokes a hole through the membrane, immediately that you're going to have an injury current that's going to try to go through that location to try to equalize the voltage gradient. And so for free now, without having to evolve any kind of additional mechanisms, you now have a vector to the damage. You know exactly where your damage was and you know that you've been injured because now your voltage is dropping, so you're depolarizing.

I would venture to guess that that's the first physical correlate of pa. And, and, and, and you also know immediately where the damage is because, because that's where the field lines are going. And so all of that you get, you get for free. So, so this is just, just one example of the many amazing kind of free lunches that, that, that biology makes, makes use of in physics.

This video is sponsored by Superhuman. Ever feel like your inbox is a black hole? No matter how much you scroll, the dreaded unread count just keeps growing. Your deadlines are vanishing like they're beyond the event horizon. Horizon. That urgent collaborator message.

It's out there somewhere.

But finding it is going to be harder than herding thousands of Schrodinger's cats. You tried AI add ons before, haven't you? But let's be honest, they're slow, clunky, and the drafts sound like they're written by a grant reviewer who skimmed the abstract with too little coffee. Searching for that one important email feels like scanning the cosmic microwave background, hoping that your signal just isn't dust. Superhuman has trans transformed the way that I work. Imagine if your inbox didn't fight you, but it worked for you. Like your brilliant and tireless graduate students. That's the transformation with Superhuman. It's the leading AI native email app.

One that never misses proposals, never forgets referee replies, and always knows what matters most. From your family to your friends to your academic colleagues, students and more. Take thousands of unread emails into automatic organizational bliss. Auto labels and auto archive clear the clutter in my inbox. Split inbox keeps urgent messages in focus and superhuman calendar, instant event and ask AI makes scheduling seem like magic. From missed opportunities to automatic follow ups with write with AI, drafts materialize faster than a pulsar pulses, perfectly timed and.

Put into your voice.

Auto reminders and auto drafts mean you.

Never drop the ball, even if your.

Collaborators take their sweet time collapsing their.

Email reply wave functions.

This is the emotional shift that I've experienced from stressed and overwhelmed to in control and mentally clear. No inbox singularities, just focus, clarity and time. Back to finally read that paper you've been citing without actually opening, escape the gravitational pull of email inbox, overwhelm and.

Experience super superhuman transformation.

Get started with one month free of superhuman today, using my link below.

So when I look at the kinds of ways that that medical technology has been brought over from physics, I mean, one of those is something like a defibrillator, which, you know, will provide, you know, on a very short time scale still, you know, huge amounts, you know, 30amps or something, occurring at a thousand volts, but only for milliseconds of time. And as I've understood it, you know, the heart thrives on regularity and the brain sort of thrives on irregularity. Although I'm kind of certainly messing up, you know, the terminology here. But can we talk about, like, how does, how do the cells know? You know, I guess it's part of the genotype, but how do they know? You know, for a brain cell, it would be very damaging to have a defibrillator sort of device. I mean, that would be maybe like electroshock therapy or something, which is basically outlawed, versus a heart cell which says, you know, bring it on when I'm not feeling rhythmic. How do these bioelectric properties get instantiated in the cells? You know, rather than just kind of being scattershot through every cell, basically reacting to voltages, currents, magnetic fields, frequencies, all the same. How does that work?

Yeah, well, what happens is that these, these cells there are, there are multiple scales of circuits here. So within a single cell, within any one cell, there is, there is a circuit formed by the different ion channels that exist. So you've got ion channels that pass positive and negative charges and they have different sensitivities to the voltage. In other words, you can very easily imagine putting together some kind of a regulator that's going to keep the voltage roughly in some kind of homeostat that's going to keep the voltage in roughly the right category range. So that's the first thing that happens. The second thing that happens is that now you have These, and they are arranged in a network because they're not sitting there by themselves, while in unicellular organisms they are amoebas and things like that. But in a multicellular body, all of these things are connected into a network. And so now you have a circuit of circuits.

And so now you've got to the tissue level and then the organ level and so on. And so you've got circuits of circuits of circuits. Each level here has its own robustness properties. Each level has its own sensitivities and thresholds such that when something interesting happens, it does kick off, it does do symmetry breaking and kick off various kinds of amplification processes and you've got different kinds of computation. So there are cells that are using these voltages to compute. There are cell groups and tissues and organs and so on, all of these things. The cool thing is that simulators of all of this. So the first one was my colleague Alexis Pytak created this thing called Betsy, the bioelectric tissue simulation environment.

And you can lay out a bunch of cells and you can say, here are the channels that are going to be expressed. Now go, run, tell me what's going to happen. And then you can do all kinds of cool experiments, right? So you can say, okay, does it do pattern completion the way that we see in flatworms, where you cut off a piece of it and you say what happens to the rest of the voltage pattern if we just killed a bunch of cells? What. And you can watch these things rescale. And some of them have really interesting memory properties such that when you change the voltage, they remember it and they keep it. So like an electric memory. So yeah, so we have simulators now for all this stuff. And this is a very active area of development actually, to be able to predict and infer intervention.

So you could say, if I wanted the voltage to be this or that, what would I need to do? And so that's the computational side.

Would there be any implantables or would that involve, you know, external stimuli and responses or how is that brought into effect?

Yeah, yeah, great, great question. The technology for this actually right now is really not about electrodes. Electrodes are really good at two things. They're really good. So this is the implantables kind of idea that they're, they're really good at spiking neurons. So there's some, there's some beautiful work on the peripheral nervous system and the, and the vagus nerve and all this kind of thing, you know, being able to control the nervous system that Way, so, so they're good at that. And they're also good at establishing a standing electric field that would be used by cells as a guidance. Cells love to crawl in electric fields.

They can sense electric fields and they have preferences. One will go anode, the other different cell type will go cathode. And if you wanted to make cells migrate, as we often do, for example, for regeneration towards wounds. So somebody like Min Zhao has done some amazing work on the natural electric fields that are formed when you puncture the skin. When you have various epithelial damage, there's an electric field. And this is the story I was telling before, where, where the field lines are an immediate guide to the damage. So all these migratory cells immediately start following the field lines and they hit the site of damage. So those kinds of things.

The implantable electrodes are really good at what they're not good at. At least right now. Maybe someday somebody will figure out a way to do it. What they're not good at is at setting up spatial patterns of differential voltage potential, which is what we need here in order to do the things that we do. Induce organ formation, limb reg, the tumor, reprogramming, repairing birth defects, all these different kinds of things. In order to do them, you need to set up complex patterns of different voltages across space. And I don't know of any way, and I haven't seen anybody develop any way to really do that well with electrodes. So how do we do all this stuff? So what we do is we use ion channel drugs that are chemical compounds that target, get specific kinds of ion channels.

And so we have a computational platform where we can say this is the pattern we have now. I'd like it to be this instead. Tell me what channels and pumps I would need to open and close in order for them to do that. And the model should give you a suggestion. And then we go to the shelf and something like 20% last time I looked at it. So 20% of all drugs are ion channel drugs. So you have this incredible. Yeah, I mean it's amazing.

You have this, you have this incredible. Of pharmacopoeia, of electroceuticals, of basically drugs that can be repurposed, you know, they may be used for neuro neurological symptoms, they may be used for, you know, inner ear or whatever. You can repurpose them for other, for other things. If you've got the platform that tells you which channels do I want to open and close.

If you remember Jurassic park, that great movie, you know. So Jeff Goldblum, you Know says something like life finds a way, life is always right. So I just remember it like that. But the counter example is neutrinos are left handed and there's been some conjectures that, that only left handed neutrinos stemming from cosmic ray interactions could have caused the initial asymmetry. Still not entirely understood by me at least. But talk about this breaking of symmetry and kind of the way that it kind of grows out again from objects. We know we can make sugars and we can make a whole mirror version of a human being according to many people that I've talked to. So that would be quite strange.

Striking. But, but how does handedness come into this? And is there potentially again a link between the physical world that I inhabit of say magnetic fields, which, which are also, you know, distinguishable at some level using the weak nuclear force alone? As far as we know. Is there any way to trace the origin of handedness and embryonic development to handedness, you know, perhaps due to polarized radiation, where we have antennas that are circularly polarized left to right, right. For example. Is there anything to this notion that the origin of, of morphological asymmetry and chirality could come from, you know, some, some physical mechanism?

Yeah yeah. This is, this is great. I love, this is a favorite topic and I haven't done work on left right in, in quite a bit, quite a bit of time. But, but I used to, this was my, my PhD was in this and I worked for roughly till 2016 pretty much on, on this, on this problem. It's a fascinating problem. Let me, let me just kind of set up what, what we're talking about here first and then, and then relationship it might have to symmetry breaking in the, in the universe. So, so, so the basic fact is that most, most animals and in fact it goes beyond well beyond animals. There are, there are plants that do it.

There are, there are all kinds of, all kinds of weird creatures that, that have a consistent asymmetries. They have a basically a, a fundamentally, a bilateral symmetrical body plan. So you can draw kind of a midplane and you can say, okay, you know, the, to, to a first approximation the reflections are the same. Yeah, they're the same left and right, except that you find that there are consistent differences. And so in humans, as you pointed out, the heart, the liver, the gallbladder, there's all sorts of the stomach, there's all sorts of organs that are asymmetric. The brain hemispheres are not the same. All kinds of interesting phenomena such as certain syndromes that affect Non asymmetric structures like shoulders and hips and things like this sometimes occur more prevalent on one side. There are lots of other creatures that are more obvious.

Crabs and lobsters often will have one claw that's quite different from another. Many interesting examples like this. You know, there are slime. We showed years ago that slime molds, when they're growing out and they want to turn, they preferentially turn in one direction versus another. So even this, even slime molds have a sense of left, right symmetry. And the cool thing about it is that it isn't random. It's easy enough to come up with a mechanism that will pick one side at random. What's much harder is to have consistent asymmetry where the direction of the symmetry breaking is fixed relative to the other two axes.

And I used to have, back when I talked about this a lot to, to, to students, I would say imagine that you were going to, you, you discovered some aliens far, far away. All you had was a telephone connection. You couldn't, you couldn't give them objects. You, you, all you could do is talk via voice. And so you're learning each other's language and you've got all the other words down and it's time to decide what left and right mean. So you say, okay, so, so, so I'm standing here and let's say that, you know, my, my, my sense organs are, are pointing forward, so that's the first axis. And my feet are pointing towards the center of gravity of the planet. So that gives you the second axis.

Now my left hand is the one that what. And now you're stuck right now it's really, now it gets really hard unless. Right. So as you well know, we could do the right hand rule and we could do something with magnetic fields and we could do the right hand rule. Or maybe, maybe they live in the same kind of universe as us and they have the same neutrinos and all of that, but otherwise at a macroscopic scale, it's really hard. There really, there really isn't. I mean, genetics doesn't distinguish left from left from right. And so you're going to express we actually don't.

And that's what I did for my PhD is discover a pathway of genes that are expressed on the left side differently than, than they're expressed on the right side. And, and then I spent my postdoc and, and some years after that chasing that back to see what the origin is. And I'll, and I'll tell you what that is. But, but It's a really, it's a really fundamentally, a very fascinating and difficult problem and, and compressing, you know, 20 years of work into a couple sentences. I will tell you that basically what it boils down to is the chirality of a particular little structure. It's part of the cytoskeleton inside of cells and it literally does the right hand rule thing. In other words, it has a chirality. The cell has two.

The early embryo, the very early embryo has two other axes. So it anchors this in one direction, it anchors this in the other direction direction. And then it's got a little, little feature that points in, in the, actually rightward and that nucleates a bunch of cytoskeletal tracts along which motor proteins will ride. And they take certain cargo. The cargo they take, among other things, is ion channels. And it sets up literally when we can see, and we showed this in 2002, there is a voltage gradient between the left and the right sides and it arises there consistently because this little, little nucleating molecule allows the ion channels to, to be different on side than the other. And so you know, your point about the, the, the, the, the symmetry breaking of, in, in, in physics and so on. I mean I'm certainly not an expert on it, but I used to read, read all this stuff and just this amazing, you know, this, the CPT violation right at the, at the basis of it, I always thought was, was the most amazing thing and does, you know, does that actually.

So, so there's two possibilities and I, I don't think we know which one is, is right one possibility and people have, have published papers on this that because of, because of the, the some enantiomers of certain molecules are more stable than others. And the idea is that not by much, but enough that evolution might have picked up on this and that the reason that we are all chiral in the same direction is because literally, you know, because of the way that, that, you know, I think it was, you know, the electrons get ejected in the, in the weak nuclear force. You, you are actually more stable. These molecules are more stable in one direction. That's, that's one possibility. The other possibility is that it's a frozen accident. That, that basically, that basically there is no major difference. But the first successful universal ancestor to life just happened to have it going one way and it was too hard to change it after.

You know, evolution's full of these things that once you set it up, you can't change it. So I can't tell you I Can't tell you which one of those is. Is the case, but. But I think it's fascinating to. To think about whether it goes all the way down.

I want to stick to physics, and I want to also let the audience know that soon we'll get to Xenobots. Not. Not the xenophobia, not Xena the Warrior Princess, but we'll get to Michael's wonderful portmanteau Xenobots in just a minute. But before we get there, I want to ask you a very simple question to you, which is one posed by none other than Erwin Schrodinger himself, Michael. And I think you know what I'm about to say. Michael, tell us, what is life?

Yeah, yeah, the simple. The simple questions, of course. So, interestingly enough, we just put out a paper where what I did was I polled about 70 various thinkers that I chose on this question. I gave them up to three sentences, and everybody gave their definitions. And then we had, with the help of some AI tools, we actually created a conceptual space for all the definitions to kind of look at the structure of how people think about this. Needless to say, there is no real agreement. I'm going to say two things. One, which is weird for biologists to say, but I'm going to say it anyway.

I don't think it's a particularly interesting category. In other words, I think what's a very interesting category is the spectrum of cognition, which. Which I also think goes all the way, all the way down. And I think life is a subset of cognition. I don't think it's important to try to wrestle over difficult corner cases and try to come up with definitions of life that try to make rulings. Yes or no. I don't. I don't like these binary definitions.

I don't. I don't think they facilitate. I don't think they facilitate research in any way. You know, however, if we. If we wanted to say what it is, that's. That. That that is special about life. I think that we, we human observers tend to call life those things that are very good at scaling their cognitive light cone.

In other words, what you have are. So let me just give a definition. So the cognitive lycone is the size in some particular problem space of the biggest goal that a system can pursue. It's not the reach of its effectors. It's not the reach of its sensors. It is the size of the goal state that you can pursue. So, for example, if you're a bacterium, all you really care about managing is the concentration of Nutrients and some other things. In a very small area.

You have a little bit of pretty prediction going forward, you have a little bit of memory going backwards, but that's it. That little area of space time that you are managing, you could care less what happens anywhere else. You're managing a tiny little, little area of space time in terms of making efforts to make, make it be one way versus a different way that, that entropy would have. You go. If you, so if you tell me that all I care about is, is the sugar concentration within this, you know, 20 micron region, I'm going to say you're probably a bacterium. If you tell me that the, that you know, you're interested in, you're actively working on goals of what the financial markets are going to look like all over the earth 100 years from now, I'm going to say you're probably a human. And there are many in between cases. So for example, if you've got a dog, the dog is never going to, to my knowledge, is never going to be able to care about in the sense of pursuing goals.

What's going to happen three weeks from now, you know, three towns over, right, that's just outside of its cognitive light co and it's certainly bigger than the bacterium, but it just isn't going to care and there's nothing you can do to make it care again. As far as I know. These are all empirical things. You have to do experiments to find out. And so, and so what I think happens is that very tiny things in certainly cells, but I actually think it goes even below the cellular level. There are certain configurations of those things that when they get together, the cognitive lycone goes up. So, so I'll just give you a very simple example. Individual cells have little tiny cognitive lycones and they care about things, things like their PH level, their hunger level, their voltage state, you know, these, these kinds of local little tiny things.

But groups of cells care about grandiose construction projects. For example, you've got a salamander limb, you amputate the limb immediately. The cells notice that they've been deviated from the correct state in, in the anatomical space. They work really hard, they, they build the limb and then they stop. That's the most amazing thing is when it's a homeostatic error reduction system, when they've solved their problem and they've reached, they've, they've reduced the error back to acceptable levels, then they stop, stop activity. So the collection of cells are able to pursue a very large goal that, you know, no individual cell knows what a finger is or how many fingers you're supposed to have, but the collective absolutely knows. And you know that with a by experiment because if you try to deviate them from their goal, they will do ingenious things to get back there. That's kind of the definition of it.

And so, so I think, and so, so what has happened? And this is what we study. We study how electrical networks scale the cognitive lycome. So what happens is, is that specifically by memory, anonymization, stress sharing and some other things, when you make these electrical networks, not only does your cognitive light cone get bigger, but it also projects into other spaces. Individual cells have access to metabolic space, gene expression space, physiological space, but groups of cells have access to anatomical morphous space. If you happen to have a brain and some muscles now you have access to three dimensional space. If you have access to language, then you're into linguistic space and God knows what else. But, but you get the idea. And so, and so I think when, when we see things like that, that have a multi scale architecture where the cognitive light cone of the parts becomes expanded and projected into new problem spaces, we say that that's, that's life.

And there are some other things that we could get into about interpreting your own memories and some other things. But, but fundamentally that's what I think is what, that's what we mean when we say life.

Yeah, well, it's clear, you know, from the moment I click record that we're going to need probably a bigger boat, a bigger podcast, and we'll hopefully have many opportunities to do this in the future. But the next question on the kind of ease of discussion is sort of a problem I have with my students and I wonder how you approach it too, that in cosmology we study the evolution of the universe, the origin is part of it. But the question of what caused the universe to be in existence is not really part of the process of cosmology. And so likewise, guys, my next question is going to be about, you know, what caused life to begin? What is the origin of life as you see it? And if you can throw in the word panspermia, that would be helpful for me because I give away these meteorites that have come from the early solar system. I give them away to everybody who has a edu email address that lives in the United States. So you are going to get one. Michael, I'm going to ship one to you. I love it.

Any of your students that you get to sign up for My magic Monday mailing list@brian king.com if you have an EDU email address, you'll get one of these beauties shipped via not gravity the way I got it, but via the U.S. postal Service. But if you don't, you can. I give away some to people that don't have the luxury and the sloth of being like you and me at a university, Michael. So I give those away@brian keating.com yt so please do take us up on that, Michael. Tell us, how did life originate?

Okay, I don't have much to say about panspermia. I'm not terribly worried about it because that, yeah, sorry, but, but, but that bas puts it off to somewhere else, right? In that sense, it has to be coming from somewhere else. So I'm not going to worry about that terribly much. I can say a couple of things that I hope are interesting about how life might have originated. One of the things that, and again, I'm not as worried about life as I am about mind and cognition. And I actually think that that's a superset of things that we recognize as living. Anyway, there's a really interesting phenomenon that we found and that, but actually, this paper just came out a couple of days ago, actually. Imagine a model of a molecular pathway.

So you've got, let's say 10 different molecules. Each one basically up and down regulates some other. So you can draw a little network of positive and negative interactions of chemicals that turn each other on and off or potentiate each other's activity or suppress. It turns out that, and this is something we did a couple years ago, and this was Sarama Biswas's work in my group, what we showed was that even, even those very simple molecular networks, never mind cells or brains or synapses or any of that stuff, just, just a small group of chemicals that turn each other on and off already was able to do six different kinds of learning. They can do habituation, sensitization, and they can do Pavlovian conditioning. They can do associative learning so that you get long before evolution kicks in. Now, now evolution, of course, is going to optimize the hell out of it. And we did find that, that, that biological, like real biological networks do better at this than random networks.

But even random networks do this a little bit. And the other thing we found out just recently, and this is, this is the thing that was, that was published the other day, and this is Federico Pagosi's work that if you, if you compute measures of causal integration, which is to Say, very sort of very, very, very roughly. It's, it's a, it's a new, it's a new set of mathematics, mathematical techniques that allow you to quantify to what extent is something more than the sum of its parts. You know, this. Was this, this idea that used to be kind of a philosophical thing is are there any higher levels of causation or is everything, you know, is reductionism right versus holism? That, that used to be a philosophical argument through, through the work of, of a number of people, including Eric Holtu, who works at my, at my center here, and, and Giulio Tononi and Olaf Sporns and so some others. There's been a branch of mathematics developed that can actually quantify that in, in certain systems you can literally do the calculation and you can say, okay, yeah, the parts are doing all the work or the, actually there is a higher level that's doing something that the parts aren't doing. So we can apply this math to all sorts of things and we applied it to these, to these pathway models and we found out that every time they learn something, their causal emergence goes up. In other words, the process of, the process of responding to stimuli in a way that trains you, meaning that you will respond to them in the future differently than you respond wanted to them fresh.

Going in that process makes you more than makes you. Raises the amount by which you are a coherent, whole, integrated entity, not just a sum of parts.

Michael, on a personal note, you had a wonderful obituary after the death of your colleague, the late, great Daniel Dennett. I had the honor of doing his last podcast interview right before he passed away. I guess it's obvious he, he did it before he passed away. Can you talk a little bit about this? The impact that he had maybe be beyond the laboratory and so forth, but his modality of thinking, it seemed to really have affected you. And it's only about a year since he passed away.

Yeah, Dan was an amazing person. You know, I grew up reading his books. I have, I have all his books back there that I read when I was, when I was, you know, a teenager and, and beyond. And I, I was, I was just, I felt incredibly privileged to take a course with him when I, when I got to Tufts as an undergrad, he taught the Philosophy of mind course. This would have been in the, you know, 91 or. I took that course with him and, you know, he was, he was, he was incredible. Now, now, to be clear, I don't, we don't necessarily agree on everything. There are many things that we didn't actually agree on in terms of this field, but he was.

He was an amazing. First of all, he was a gentleman. He was not interested in any kind of games in one upmanship, any of that stuff. He. He. He was really interested in getting to a better argument, to a better version of, you know, whatever understanding we could. Could all get by interacting with each other. And he was just, Just, Just incredible in terms of being clear and making, you know, getting everybody else around him to be clear about what.

What they were trying to say and then just seeing what, you know, what. What we could make of it. And he was. He was always. He was always kind and he was always generous with his time and with his ideas and with his advice. And, yeah, I couldn't. I couldn't say more nice, nice things about him. And so, you know, when I came back to Tufts as a.

We collaborated on some things. We wrote a paper together, which I'm. I'm very, you know, kind of. I'm very proud of, talking about cognition all the way down. And, yeah, he was. He was. He was, you know, he was an amazing person, an amazing mentor and a very deep thinker, I think.

Yeah.

Yeah, I was very, very distressed when he passed away. I think I got to interview the three of the four Horsemen of the Apocalypse, at least. Never got to interview Christopher Hitchens, but. But Dan was certainly, you know, at the very top of that very August list. So. Okay, I'm to finish up for this part, one of hopefully many parts together. And that's with your wonderful work, as I promised earlier, about xenobots. So, first of all, what are xenobots, and how are they behaving like living organisms, moving, healing, even cooperating, working together without any brain or nervous system?

Well, I mean, to be clear, there are many organisms that don't have a brain or nervous system, and they do all kinds of interesting things.

That's why I say I call those colleagues of mine at the faculty club. You know, I say jellyfish have existed without brains for 65 million years. So there's hope for you yet. Not. Not you, Michael.

But.

Yeah, yeah, well, yes, you said it. I didn't. But I mean, look, here's. First of all, I'll tell you what the xenobots are. So. So this is. This is joint work between my lab and Josh Bongard at the University of Vermont, and we are. We've organized this thing called the icbo, which is the Institute for Computationally Designed Organisms.

And the xenobots are kind of the first. Were the first salvo from that whole effort. And the computer science part of this was done by Sam Kriegman in Josh's lab. And most of the biology was done by Doug Blackiston and my group. What we basically found out is that if you isolate some prospective skin cells from the early frog embryo, and this is why they're called xenobots, Xenopus laevis is the Latin name for the frog that we're using. And so this frog lays, lays eggs, they become embryos. And at a very early stage we isolate some, some cells that are going to become kind of like an outer skin covering. And the deal is that under normal circumstances these cells get, basically get bullied by the other cells in the embryo to do a very specific thing to be this boring outer two dimensional kind of covering for the embryo, keep out the bacteria.

And that's that, that's, that's, that's what they do in the natural state because they are hacked by these other cells that are there. If you liberate them. So very specifically not putting in any new genetic material, no synthetic biology circuits, no genomic editing, no weird nanomaterials or drugs or anything like that, all we have done at this point is liberate them from the influences that they normally get, then you can ask the question, what, what do they, what, what else do they want to do? It's kind of a reboot of their multicellularity and they could do many things. They could, they could crawl away from each other, they could die, they could form a two dimensional sheet, a monolith, could do many things. What they instead do is they get together, they form this, this, this little, this little ball and that little ball is covered with cilia. These are little motile, tiny motile hairs that then orchids organized and they start, the cilia start waving and they basically align so that the thing can now swim through the water. Normally these cilia are used to distribute mucus over the body of the frog, but now they can swim. And so they swim around in the water.

And I'll just tell you a few things that it turns out they're capable of now. They're obviously living organisms, they're made of living cells. The major interesting thing about them, beyond the fact that they're a biorobotics platform, so that is, once we understand how they work, we could potentially use them for all sorts of cool applications, cleaning the environment and just a million different things. But the other cool thing about them is that unlike the actual frog, if you ask about what sets the Properties of the frog. Everybody will say, well, it's evol. It's. It's a history of selection against specific environments, right? Going all the way back. That's when the computations were done to design what a frog is.

It's by testing things against the environment and killing off everything that didn't work that that's why the frog looks like a frog. Well, there's never been any xenobots. There's never been any selection to be a good xenobot. And here are some things they do. First of all, they express hundreds of genes differently than they would have in the body. So they have a completely different transcriptome among those genes. Many interesting things. I'll just pick one to talk about.

They express a bunch of genes that and other creatures are related to hearing. And so we actually tested that and we put a speaker underneath the dish and we played certain frequencies and we found out that, yeah, in fact they change their behavior depending on the sound vibration that you're giving them. They do this other crazy thing we call kinematic self replication, which is if you sprinkle a bunch of loose epithelial cells into the dish, they do the kind of von Neumann's dream of a robot that goes around and find and makes copies of itself from parts, from materials it finds in the environment. They will literally both singly and as a collective, they will go around, they will collect the cells into little balls. And because the cells themselves are an essential material, just like the ones we started with, the little balls mature and become the next generation of xenobots. And guess what they do. They do exactly the same thing. They run around, they make the next generation.

So now, you know, there's. To our knowledge, no other creature on Earth reproduces this way. I don't think there's been a kinematic self replication on Earth. So they raise these kind of biobots and other kinds of concepts, raise these very interesting questions of where do the properties of novel beings come from if they don't have an evolutionary history? This is something that we are now confronting as a society with AIs and with cyborgs and hybrids and all these weird things that we're all making. Where do their properties actually come from? And here, if we, you know, maybe next time we'll talk about it because. Because I think it is this platonic space of forms where the truths of mathematics come from, actually. So that's xenobots. And we have another thing that was developed by Gizem Gumushkay in my group was a PhD student, we call them anthrobots because some people said, well, xenobots, the amphibians are plastic and it's embryonic, so maybe it's a one off special thing.

So fine, what's the furthest you can get away from embryonic frog? Well, adult human. Let's try that. We took cells from adult human patients. Tracheal epithelial cells. Turns out that they too come together, these little motile creatures. We call them anthrobots. Those guys have 9,000 differently expressed genes. Half the genome then, Right.

And they can do cool things like they can heal neural wounds. So if you put them on a bed of human neurons, you take a scalpel, put a big scratch through it, they can settle into that wound and form what we call a superbot cluster. And when you lift them up four days later, what you see is that they took the two sides of the gap and they healed them together. And again, there's never been any anthrobots. There's never been selection to be a good anthrobot. You know, we have to understand, understand where do these things come from? And who would have thought that your tracheal epithelial cells, which sit there quietly dealing with pollen and you know, who knows what else, are able to make a self moti little creature that can go around and heal things like neural wounds, Right? And this is just the tip of the iceberg for these things.

Well, instead of frog makes humbly suggest the elephant next time. As the mascot of Tufts, I'm from near Somers, New York, which if you know, is right where Barnum and Baby Bailey got their start and where Tufts got part of its start. So, Michael love, and this has been phenomenal. I knew it would be. It's two years in the making. I'm glad you took so much of your time. You're so generous.

Thank you so much.

And I hope that we'll do it again. As I said, we have. I have questions enough for. I'm counting up here, parts up to part four. So maybe in person even. That would be great. Thank you so much, Michael. Have a wonderful.

Very good to meet you.

Okay, great. Thank you so much.

The implication of Michael Levin's work extends far beyond biology. Touches upon the very foundations of physics, consciousness, and our place in the biological cosmos. If cognition really does go all the way down, what does that mean for human understanding of intelligence? And what does that mean for other artificial and alien intelligences throughout the universe? I know if you enjoyed this conversation about the electric basis of life and consciousness. I know you'll enjoy my episode with Sir Roger Penrose and Stuart Hameroff where we explore quantum mechanics in the brain and whether consciousness might emerge from quantum quantum processes in microtubules. Two brilliant minds, two revolutionary ideas about the nature of consciousness and life itself. Don't forget to like, comment and subscribe.