Watch Episode Here
Listen to Episode Here
Show Notes
This is a ~35 minute talk about diverse intelligence and our efforts to establish formalisms and methods for communicating with unconventional biological intelligences.
CHAPTERS:
(00:00) Expanding Mind Blindness
(02:50) Recognizing Diverse Minds
(05:00) Embodied Agent Origins
(08:40) Anatomical Intelligence Spaces
(12:40) Bioelectricity: Body's Glue
(15:50) Communicating Cell Intelligence
(22:30) Patterns Beyond Physics
(25:00) Anthrobots: Novel Life Forms
(27:30) Future Beings, Pattern Space
(32:30) Conclusion and Thanks
CONNECT WITH ME:
Transcript
What I am going to talk about today is some efforts that we and others have made to try to expand what I think of as significant mind blindness that affects us because of our evolutionary history and our focus on particular kinds of minds. I'm going to spend the middle part of the talk discussing results that we have using the collective intelligence of body cells as a model system to understand a diverse intelligence that lives in a weird space that's hard for us to imagine, and the progress that we've actually made in communicating with that intelligence. So it's kind of a case study for us.
But I'm also going to talk about some much, much weirder things. If anybody's interested in the primary papers, the data sets, the software, everything is here, and then here's where I keep my own kind of personal thoughts about what it all means.
I want to start off by just reminding ourselves of something that happened over history. Prior to having a solid theory of electromagnetism, we had static electricity, and we had lightning, and we had magnets, and we had light, and we all thought those were different things. We thought these were completely different kinds of phenomena. And we were also unaware that we were only sensitive to a tiny part of this spectrum.
What a good theory of electromagnetism did for us is several things. First of all, it revealed that all of these different things are in fact manifestations of the same thing, and that it is a spectrum, and that the labels we put on this spectrum are convenient, but they're for our purposes. The underlying thing is a continuum. And it allowed us to develop a technology driven by the theory that now lets us interact with this enormous range of things in the universe around us to which we were previously basically completely blind, and that allows us to make applications that improve quality of life.
So I want to argue that this idea, the idea of a unification that reveals a spectrum and an underlying symmetry to seemingly disparate phenomena, is in fact the exact way we should start thinking about diverse kinds of minds. In my group, what I would like to do is to develop a framework that helps us to recognize and ethically relate to a very wide range of beings. And so not just the primates and birds and maybe an octopus or a whale that conventionally we think about, but also some really weird biologicals such as colonial organisms and swarms, synthetic new life forms, AIs, whether robotic or purely software, and someday maybe exobiological alien entities, and some even very weird things that I probably won't have time to really talk about today, which aren't even conventionally physical objects.
And I think what we have here is exactly this kind of spectrum, where we have names for different types of agents along the spectrum all the way from passive matter – and I'm not sure there is any such thing, but we tend to think of the kinds of things that physics studies as sort of passive, dumb materials – all the way up to whatever humans are with their metacognitive abilities, and everything in between. And so what I'm interested in are theories of transformation and change, not sharp categories, but actually how things scale along this continuum, and all the details are here.
One thing to remember is that as we study these kinds of things, what I think we are working up along is this kind of spectrum where things get progressively weirder as you start to leave this conventional area. The kinds of agents that we think about get stranger and stranger.
And one other precedent that we have is in the discovery of numbers. So originally we have the counting numbers, and then somebody invented zero, and then negative numbers, and then eventually people figured out that even the rationals weren't the limit of things, and so on. And what's important here is that each one of these expansions required a conceptual leap. It required us to break prior categories that we thought defined what is essential about numbers. It actually was quite disturbing, so lots of people died, for example, when irrationals were discovered, like this poor guy, Hippasus, who was either killed or drowned on his own from the horror of this kind of thing. And we just need to remember that expanding these categories is not trivial. It's often painful. And the ability to recognize other kinds of things that didn't fit your previous category is in order so that you can do something useful with them, and in the context of what we're talking about now, which is other types of beings, actually have some sort of active, compassionate relationship with them, is not easy.
We as humans are really primed to recognize this kind of thing. So here's this Oscar-winning performance where this little squirrel-like creature sets up a little accident scene here. He just puts it right on the neck. He knows what his owner is... he's got a very good theory of mind. He knows what the owners are looking for in terms of paying attention to him. He's going to check in a minute to see whether anybody's watching. So these kinds of things, it's very easy for us to understand that there's a mind here and so on. And that's because they operate at the same spatio-temporal scale. They operate within the same space as we do. They have roughly similar biological goals. And so things that are driven by brains, especially mammals, are not really problematic usually.
Well, there are some weird situations. For example, we have theories of mapping of functions onto brain structure and so on, and yet there are clinical cases, and Karina Kaufman and I reviewed it in this paper, there are cases where individuals have radically reduced amounts of brain tissue despite normal or even above-normal performance. So even that's not simple, the mapping between brain and performance, even that's not easy.
In order to really understand what we are and how we come to be as embodied agents, we need to look at our origins both evolutionarily and developmentally. So all of us start life as a little blob of chemistry and physics. Here's an unfertilized oocyte. Through the incredible process of embryogenesis, we become fodder for either behavioral science or maybe psychoanalysis and other deeper kinds of things.
One thing developmental biology shows us is that there is no magic, kind of bright line at which physics takes over to psychology. There is no special place where that happens. It's a continuous gradual process and it's really important to be able to understand that process. And by the way, this is not the end of the road because there are other processes that can happen afterwards, including a dissociative disorder known as cancer, I'll talk about that in a minute, and really other very strange things such as your cells, for example, after the death of the body, your cells continuing their life in a totally novel embodiment such as these anthrobots that I'm going to show you in a minute.
And it's also important to note that even the notion of an embryo isn't simple because if you look at an early blastoderm, so let's say there's a couple of hundred thousand cells here and you say, "Well, there's an embryo," what are you actually counting when you say there's one embryo? What is there one of? Well, what there's one of is alignment. What there's one of is a story that all of the cells have bought into, a self model that guides their movement in anatomical space, in the space of possible anatomical configurations. And the reason there's one embryo is because all the cells agree that is where they're going.
If you were to, and I used to do this in duck embryos as a graduate student, if you take a little needle and you put some scratches in this blastoderm, each one of these islands doesn't feel the presence of the others and so they align among themselves, they decide they're the embryo, but they all do it and this is how you get twins and triplets and things like this. That means that the number of individuals inside an embryonic medium is not determined. It's certainly not set by genetics. It could be anywhere from zero to probably half a dozen or more in a standard blastoderm. And so how many actual individuals or cells can emerge from that medium is not obvious and it's a physiological outcome that shares actually a lot of both mechanistic and functional similarities with issues of individuation in the cognitive sense, so things like split brain patients, dissociative identity disorders. You know, how many individuals are within a given set of real estate in the brain is... and how many can actually fit? What's the maximum carrying capacity? We know this for standard computation. We do not know how to calculate this for neural substrates.
And so it's an incredibly dynamic process, self-making, literally self-making, is a very dynamic process and it's driven by this sort of thing. This is what we are all made of. This is a single cell. Now this is a free living organism, but you get the idea. It's a single cell. There's no brain. There's no nervous system. There is however, great competency in managing its own tiny little cognitive light cone. The spatio-temporal region of goals that it can pursue is small but it's very good at doing what it needs to do.
And even this, I mean, a lot of people are now catching on that cells have a kind of intelligence and competency and whatever, but it's not just that this goes down to the cell level. No. It goes far beyond that because even the chemical material inside of that cell, not the cell itself, not the nucleus, not any of that, but just the chemical pathways, for example, gene regulatory networks or other chemical pathways, just from the way the math works, they are themselves capable of at least six different kinds of learning, including Pavlovian conditioning. You don't need anything else. Just a set of interacting chemicals is already potentially capable of several different kinds of learning.
And so already not only are we collective intelligences made of cells, but actually the cells themselves are a collective intelligence made up of components that do all kinds of interesting things that I don't have time to talk about today. But we are, for example, in our lab, taking advantage of the ability of these molecular networks to learn, to make applications such as drug conditioning and things like that.
So our body is a multi-scale competency architecture. Every level from the molecular networks on up is a kind of agent that solves problems in various weird spaces. And because of our evolutionary firmware, we are pretty good at recognizing intelligence in three-dimensional space. We're obsessed with the 3D world because of our vision and medium-sized objects moving at medium speeds like birds and things like this. We can sort of pick that up. But biology exploits intelligence and goal-directed navigation in lots of different spaces. So the space of possible gene expressions, the space of possible physiological states, and the one we're going to talk about most is the anatomical morphospace. This thing that's essential for intelligence, this perception-action loop, happens in all kinds of other spaces that are very difficult for us to recognize.
This has lots of implications for things like AIs and various hybrids and organoids and things where people look at them and say, "Well, this is not embodied. It just sits there. It doesn't move in the physical world." But the physical world is not the only world where you can do this critical perception-action loop. And so what I mean by anatomical intelligence, I don't just mean complexity, and I don't even just mean reliability, I mean problem-solving.
And so here's a very simple example of a goal-directed system. So you have this axolotl. They are incredibly regenerative. They can restore most of their body parts. If it loses a limb anywhere along this axis, the cells immediately detect that they've been deviated from their target state. They will work really hard to get back there, and then the most amazing thing of all, they stop when they're done. When do they stop? They stop when a correct axolotl limb has formed. So this is an example of a very simple homeostatic process. The system has a goal, you deviate it from that goal and it expends lots of energy to get there, and then it stops. But this ability to... and in fact, it will work no matter where you cut. It makes exactly the right steps from here to there, so it can regain its position in that anatomical space it navigates.
But this is not just about damage. That incredible plasticity leads to phenomena like this. If we make this tadpole with no primary eyes... This is a tadpole of the frog. Here's the mouth, here are the nostrils, the brain, the tail. So if we make a tadpole with no primary eyes but we put an eye on its tail, and I'll show you in a minute how we can do that, what happens is that that eye makes a single optic nerve. The optic nerve does not go to the brain. Sometimes it synapses on the spinal cord here, sometimes to the gut, sometimes nowhere at all. And these animals can see perfectly well. How do we know? Because we built a device that automatically trains them on visual cues, and we can show that they learn for visual assays.
And this is amazing. Why do you not need additional rounds of evolutionary adaptation here? You've got an animal with a radically different sensory motor architecture. The eye's connected to the spinal cord or some other organ. Why does this just work out of the box? And that's because every instance of embryogenesis or regeneration, or remodeling or metamorphosis, is not a hardwired process driven by the chemistry of genetics. What it is, is a problem-solving process that takes place when an active agent interprets the information that it has, including its genetic affordances and everything else. So it gives rise to this incredible plasticity that we're going to look at the implications of momentarily.
So what we'd like to know is, how does all this work? So what's the cognitive glue that binds individual competent subunits like cells, and in fact, molecular networks into cells, into tissues, organs and so on? What are the mechanisms and policies that allow these things to work together and embody a higher order individual that knows things that its parts don't know?
Well, we know how this works in our bodies, as far as cognition is concerned. The cognitive glue that binds individual neurons together into us is electrophysiology or bioelectricity. It's the electrical networking of the cells in your brain that allow us to know things that our neurons don't know and so on. Well, it turns out that that in particular is a scheme that evolution had discovered long before there were brains and muscles and things like that. Using electricity to bind things together into a larger scale collective intelligence goes back to at least bacterial biofilms. It's very ancient.
And more generally, I don't have time to talk about this, but more generally, what evolution, I think, has done is to pivot some of the same navigational tricks across all kinds of spaces as life got more complex. So from metabolic spaces, physiological spaces, gene expression spaces, eventually multicellularity, and you get anatomical morphospace. Then muscles and brains appear, and you've got 3D behavioral space and then eventually linguistic spaces, and who knows what else? But the interesting thing here is that what our brains are using to embody conventional cognition is exactly the same molecular mechanisms, meaning ion channels, electrical synapses and neurotransmitters, exactly the same machinery as the collective intelligence of the body uses to make anatomical decisions.
And so this is interesting. I don't plan to talk about consciousness too much in this talk, but just when I give these kinds of talks, I remind people that when you attribute consciousness to other humans that you're talking to or other beings, you typically use one of three criteria. Either you're looking at mechanisms. You say, "Hey, they use the same mechanisms that I use and I'm conscious, so therefore they must be too." Or you're looking at behavior, so some sort of creative problem-solving or something, and you say, "Okay, this is a sign of cognition and consciousness." Or you're looking at evolutionary continuity. You say, "Well, they're on the same evolutionary stream, and therefore they must have had it too." If you like any of these criteria, then you have to take very seriously the idea that every structure in your body potentially has some sort of likely non-verbal, and so you don't hear from it, but some level of conscious experience, because it fulfills all of the criteria that we normally use to attribute consciousness to each other.
So what we're interested in is understanding, okay, if we're going to use the collective intelligence of morphogenesis as a model system, one of the things we'd like to do is to communicate with it. And in order to do that, the first thing we developed were some tools to basically read the information content of its mind, and so basically to read out the electrical activities the way that neuroscientists read out the states of the brain. And we developed a fluorescent dye technology that allows us... So here are a bunch of cells in a dish deciding whether they should be a collective or not. Here are some cells in an early frog embryo. So these are not models. This is real data in time lapse. And you can see all the electrophysiology that goes on as these cells figure out who's going to be left, who's going to be right, anterior, posterior and so on.
So we have all kinds of computational methods that we've developed to model this phenomena and to understand how it can underlie memories and large-scale circuits that do pattern completion, which is basically regeneration. You lose a part of the body and then you can sort of complete that pattern later. So we have all kinds of simulations.
I want to show you some of these patterns. Not only do these electrical patterns serve as a kind of cognitive glue to bind cells together into a single body. And so here we have, this is a frog embryo putting its face together. And so this is one frame from that video. We call this the electric face, because what you see here is that before the gene expression and the cell behaviors that lead to forming a face, this is the memory that you can read out of what it's planning to build. So here's where the eye is going to go, here's where the mouth is going to go, here are some placode structures. So the bioelectric stores a memory of what the thing is going to build, but it's also a cognitive glue across individuals. So here you can see when we poke this one, these two find out about this in short order. And the same thing here, you poke this embryo, there's a wave that goes... Each one of these is a whole embryo. These are not cells. These are whole embryos. So you can see that there's communication within and communication between individuals. They're sitting in an aqueous medium, which is how the wave can propagate. And then here are just a bunch of cells where you can see the incredibly complex electrical activity that they're able to drive.
Okay, so we're able to read these things. Now, if you're going to communicate, you can't just listen. You have to be able to convey messages too. And so this is just one example. I could show you hours of this kind of stuff where what we can do is say to some cells that are going to normally make a gut, we can give them the message that says, "Make an eye." And that message is encoded electrically. We introduce it using ion channel, potassium channels, encoded by RNA that make a little voltage gradient that looks just like that eye spot that I showed you a minute ago. And sure enough, the cells get the message and they make an eye.
Now, there's something very important here, which is that when I'm talking to you all, I don't need to worry about reaching in and making sure that all your synaptic connections are doing the right biochemical things so that you remember what I'm saying. You are going to take care of all of that because you're a multi-scale cognitive system. All I need to do is give you prompts with a very thin linguistic interface, and then you will do all the hard work of arranging the internals of your brain to act accordingly.
The same thing is true here. We provide a very simple signal. It's a prompt. It says, "Make an eye." We don't tell the cells what to do. We don't tell which genes to turn on and off. We don't talk to the stem cells. We, in fact, we have no idea how to actually make an eye any more than I have any idea of how to manipulate your synapses so that you remember what I'm saying. The system takes care of all of that because it is a multi-scale cognitive agent that takes high level information and it handles all the stuff downstream, including if we only get a few cells here, they will do the job of telling their neighbors. So the blue ones are the ones that we injected. All of this other stuff, this lens that's sitting out in the flank of a tadpole, we didn't touch any of those cells. These cells told those cells what to do. They said, "You need to participate with us to make this thing because there's not enough of us to..." And by the way, these other cells resist. They have a different vision of the future that they were following before. And we can actually track the communication that goes back and forth and to see who wins because it's a cancer suppression mechanism. Cells don't just pick up whatever message you happen to give them. They resist. They have their own idea of what they should be doing. So it's really all this morphogenesis is a constant battle of world views. It's a battle of models, and the cell collectives have to decide which model of the future they're going to go with.
Another thing you can do with this interface is to actually change the scale, the size of the self. In other words, the scale of the cognitive light cone or the size of the goals that it can pursue. So again, the individual cells have little tiny cognitive light cones and they remember very simple goals like hunger level or pH or things like that and then that's what they pursue. Their homeostatic set points are very simple things like that. But these kinds of, for example, morphogenetic systems have huge grandiose goals like building a limb. And no individual cell knows what a finger is or how many fingers you're supposed to have, but the collective absolutely does. And you know that because if you try to deviate it from it, you'll see that it knows what the set point is.
So there's this question, can we show how connecting cells into a larger network allows them to have larger goals? How do you expand the size of the cognitive light cone? And this kind of weird way of thinking about it has therapeutic benefits, of which we try to show some kind of practical implications for all this strange stuff that we say so that other people can see why it's useful to think about it this way.
Here's a human oncogene injected into a tadpole that normally makes a tumor. You can track why. It's because oncogenes make the cells disconnect electrically from the rest of the network. When they do that, as far as they're concerned, the rest of the body is just external environment. So it's not that they're more selfish, it's that they have smaller selves. The border between self and world shrinks. And whereas before it was part of this big collective that could entertain thoughts about building a proper skin and muscle and whatever, now it's just an amoeba doing its thing, so metastasis.
But what you can do is forcibly reconnect these cells to the collective and we've done that here. And if you do that, the oncoprotein is still there. It's labeled in red so you can see it's blazingly expressed. This is the same animal. But there's no tumor. And we didn't kill the cells. We didn't fix the genetic lesion. We didn't do any of that. What we did do is force the cells to be part of this electrical network where it is now part of a larger self with larger goals, that working on anatomical things instead of being an amoeba in an environment.
So just to kind of recap what I've showed you is that this bioelectrical interface allows you to detect, reset, and communicate with a very unconventional collective intelligence. And what it does is think about how to move an anatomical configuration through the space of anatomical possibilities. And if we had more time, I could show you a whole bunch of different problem-solving competencies that it has along those lines.
So all of this kind of way of thinking about it suggests to me a different kind of biomedicine. Everything that we do today is mostly bottom up. It's focused on the hardware here. If we actually focus not only on the software but actually on the intelligence inside the agential material of life, you end up with all kinds of applications, most of this I didn't get to show you, as far as training cells and tissues and top-down communication via electroceuticals and things like that. And so I think the future medicine is not going to look like chemistry. It's going to look a lot more like a kind of somatic psychiatry. Bioelectrics is one interface layer. No doubt there are others. There are biophotons and other things that probably will be developed. But I think the most well-developed at this point is bioelectricity.
So now we can ask, okay, so if the future of all of this is really the study of patterns, it's the study of patterns of thought and behavior that occur in these weird media such as the electrical networks of your body, where do these patterns come from? And the standard answer in biology is, well, they come from two places. Biologists love two kinds of explanations for things. One is history, in other words, a history of selection. The pattern is what it is because everything else died out, and this is what we have now. So in other words, there's a historical explanation for why this pattern, not some other pattern. And the other thing that biologists like is physics. In other words, patterns are constrained by physical effects, and the properties of the physical world. And so some of those patterns just come from the physics.
So what's interesting though is that there are some patterns, and we can start with the ones studied by mathematicians, that have neither a genetic nor an environmental explanation. In other words, this thing that you're looking at here, and then here are some videos of what the pattern actually looks like when you start tweaking the formula, this is a very simple function in complex number Z, so Z cubed plus seven, for example. If you do something called a Halley plot, where it's a very simple algorithm for plotting the roots of this, and how the roots behave of this function, you get this amazing complex pattern. Doesn't hurt that it's kind of very organic looking, which is why I use this example.
But what's salient here is that there's a tiny seed that encodes a very complex, a very rich kind of structure. And if you wanted to explain why it looks like this, there are no facts of the physical world that explain it. There is nothing in the history of the physics of the universe that explains why this is the way it is. The explanation does not come from the physical world. You cannot tweak any of the constants of the universe at the beginning of the Big Bang and get a different pattern out of this. It is an extra thing that is distinct from information you get from your genetic lineage and information you get from your physical environment. And there are plenty of patterns like this and like this and like this in that space as you play with these kinds of things.
So now we can start to ask, okay, so what does this have to do with biology? Could we find any living forms without a large-scale history? In other words, if biologists will say that whatever form you have is driven by your evolutionary history, are there any beings that are picking up forms that cannot be explained in this way?
So I'm going to show you just one. This is something called an anthrobot. If I showed it to you for the first time, you may have already seen this. If you saw this for the first time, you might be tempted to guess that this is some sort of primitive organism that we picked up at the bottom of a pond somewhere. You might take a guess about its genome. You might think that this will have an early genome of an early metazoan, for example. If you were to actually sequence this genome, you would find out that it's 100% homo sapiens. These anthrobots are self-assembled from adult human cells. They come from tracheal epithelia that human patients donate. Some of these patients are alive and well when they donate this, some are not. So for those that are not, this is a weird kind of life after death, where these cells self-assemble into a motile little organism. It matches no stage of human, of normal human development. You would not be able to guess from the human genome that this is what you would get. Actually, you wouldn't even guess from the genome that the human body is what you would get. But you certainly wouldn't know this.
So it's interesting. What capacities does this creature have? It's never existed on Earth before. There have never been any anthrobots. There's never been any selection to be a good anthrobot. What can it do and where do those patterns come from?
Well, one thing it can do, if you plate a bunch of human neurons on a dish and make a big scratch through them like this, one thing you'll find is that these anthrobots cluster in what we call a super bot, so they make this like pileup. And then what they do is over the next four days, they start healing the wound. And so if you lift it up, this is what you see. They're sitting there, knitting across the gap. Okay?
Now, who would have thought that your tracheal cells, which sit there in your airway for long periods of time dealing with dust particles and whatnot, that given the opportunity to reboot their cell hood, to reboot their multi-cellularity, would go on and make this kind of self-motile thing that, by the way, has the ability to heal your body? So obviously this is a technology that we're developing for in the body repair. It's made of your own cells. You won't need immunosuppression once we figure out how to ask these cells to do the things we want them to do. This is all kinds of useful technology.
But the other important thing about this is that, first of all, they have four different discrete behaviors, and you can draw an ethogram of those transition states between those behaviors like you would in any animal. They have a drastically remodeled transcriptome. What I mean by that is this volcano plot shows different genes that they express differently than the tissue they came from. Each one of these red genes, there are about 9,000 of them, so almost half the genome is differently expressed than it is in that tissue. So when we say, why do humans look the way they look and have certain properties? You say, "Well, it's the human genome." Yeah, the same genome can also make this. And if you enable them to make this, and we don't edit the genome. There's no weird nano-materials. There are no scaffolds. No synthetic biology circuits. The cells do all the work. When they do this, they will pick and choose what they express from the genome, about half of them, to be in a completely different kind of form and function.
So just to kind of end here, I want to say a couple of quick things. The first thing, and so this is kind of the most exploratory aspect of all this work, is that I see all of this stuff, the xenobots that we build, the anthrobots, the chimeras, all the things that we make, I see them as exploration vehicles.
I see them as enabling a systematic mapping of the relationship between the interfaces that we make, which are the physical objects. I think all physical objects, including cells and embryos and bodies and robots, everything, all of these things are physical, are interfaces or pointers into a space of pattern. Mathematicians have known this for a long time. They study one region of that space, and there are all kinds of truths and patterns of mathematical objects that don't come from the physical world. But I actually think it's a much richer space, and it contains other very high agency patterns that we recognize as kinds of minds.
Everything that we do is basically poking holes in this kind of separation between the interfaces and the patterns that then ingress through these interfaces. The very last thing I'll say is lest we get too smug about our ability to access this space. Among us are certain geniuses that can perceive incredible truths and works of art and literature and so on. They're picking out things out of that space and letting them come into the world. Lest we think that this is some kind of a human capability, I will just point out that, first of all, there are no humans as a sharp category. We are at the center of all kinds of continua, both developmentally, evolutionarily. Whatever you think that humans are or can do, you have to have a story for where it comes from and how it scales.
In fact, both with biological changes and technological changes, this standard human category becomes extremely diffuse because of the plasticity of life that allows us to be interoperable with all kinds of modifications. If you in the audience are a therapist or interested in crisis meeting and helping people through this, you have a set of constituents that count on you for help. You really have to start thinking about your patients of the future. The standard human patient is not enough. There are going to be all kinds of interesting new beings with all kinds of needs that we are going to have to address in some way. Some of these will have the same existential concerns as us, some will have very different concerns.
I think that because of this interoperability, all natural life is a tiny corner of this option space. Pretty much every combination of evolved material, design material, software, and these patterns that you can't forget about as an ingredient, is some kind of... there will be hybrids and cyborgs and every possible kind of being that we are going to have to enter some sort of ethical symbiosis with.
The last thing I'll say is this: even... there's some serious humility warranted in all of this because a lot of people think that A, it's complexity that makes all this work, or that B, there's something special about the biological material and this kind of trial and error process of evolution. Our work, and again, I don't have time to get into the details, but even looking at deterministic, extremely simple things that computer scientists study like sorting algorithms, already you see that, yes, it does the thing the algorithm tells it to do, but if you know how to look, you also find it doing all kinds of interesting side quests that the algorithm does not ask them to do that are in fact kinds of... that are things that you would recognize as aspects of cognition. Very minimal, but yet there they are.
I really think this goes all the way down, this ability to have the freedom to express patterns from that space that are not fully determined by the physics of your world or the algorithms that we think you are following. It is always richer than that, even in extremely simple systems. Just to remind us again, Magritte was telling us that our formal models are not the thing itself, and I would say the same thing. There are a lot of organicist thinkers who are down on computationalism and they say, "Living things aren't machines." And I say, "Right, nothing is a machine because a machine is a formal model that humans have defined to handle certain aspects of what physical objects do in the world, and it doesn't fully describe even the simplest kinds of systems, never mind, of course, the biological systems." So nothing is a machine, but some things look like a machine to lazy observers, and I think for exactly the same reason that the rules of biochemistry don't encompass the story of the human mind. Formal models of algorithms and materials do not encompass the story of minimal beings.
So really, I think we have to get much wider in our perception of what we can relate with. I'm going to end here and just thank the people who did the work that I showed you here. Some post-docs and students and collaborators in our lab, and disclosures are three companies that have spun off from the work that we do. I of course thank our funders and the model systems because they've done the most hard work in teaching us about this stuff. Thank you so much and I'll stop here.
