[00:00:03] Speaker A: If you have a theory, you can change your theory to, you know, accommodate everything. That's not a good theory, right? So I think that that's what neuroscience needs. We need to go to a new level. Basically our theories are actually, you can put it to test and refute it and reject it and just, you know, move forward. We are hopefully entering that era in neuroscience.
Recurrence is a beast, right? And I think that's the holy grail of neuroscience. Like, can we come up with theories of how these recurrent systems are working, how they like doing the operations, how they change their operations without like constantly changing the synaptic bait and at that fast rate? Because biology cannot afford that.
To me, that's agency. The agency comes from this aspect of having this sort of computational closure built into this system that now you can have an understanding of the environment and then from there you can act on it and you can control yourself. And now you have internal model of the world that's, you know, now I'm going to use the bad word, the forbidden word. That's where intelligence, consciousness, agency, they
[00:01:16] Speaker B: all like are different. Like, you know,
[00:01:24] Speaker C: This is brain inspired, powered by the transmitter. Nima Dagani is a computational and theoretical neuroscientist at MIT.
His work is framed by what he calls an N4, as in Neuro4 framework, and that would be neurophysics, neurocomputation, neurodynamics, and encompassing them all, neuro AI.
So we have a wide ranging discussion on this episode, mostly revolving around his work and views on how brains and physical systems in general perform computations to carry out their functions. Most neuroscientists, I believe, more or less casually use the term computation to describe whatever it is the brain might be doing. But Nima is focused on a principled account of how the various parts of our brains work, how they compose together, and what those compositions can achieve across the many scales of the brain and within the massive recurrence of the brain?
In the show notes, I link to the papers that we discuss, to his website where he blogs about his work, and to some of the nice recent primer videos that he created to help us ease into his work. Those links
[email protected] podcast250. Look at that. 250. Thanks for being here. Hope you enjoy our conversation. Here's Nima.
First of all, welcome to Brain Inspired. Nice to have you on. Finally, like I was just saying, it's been. I invited you years ago but. But your email spammed me or something. So
[00:02:58] Speaker A: yeah, I Think it's because I'm very bad at like, emails.
It just. I thought. Yeah, emails. Anyways, it is a slack, but I'm sorry. It's nice to. Yeah, that's fine. Before.
Yeah,
[00:03:09] Speaker C: well, it's. It's actually a good time to have you on because this past, just this past Friday, I had organized a workshop which eventually got the sexier than my original title to help it out, which was requested. The title of the workshop was Brains aren't Computational Machines, so what are they?
And we had a few invited speakers. And so this is great timing for you to be on because you're all about neural computation, right?
[00:03:35] Speaker A: Fantastic, fantastic. One of the things that I look forward to hear about it also.
[00:03:42] Speaker C: Well, I mean, one of the things that came up over and over, I mean, you know, there were people arguing there's no way that brains compute. They aren't computers, you know, and whether the computer metaphor is, you know, just a metaphor, like, because it's the most recent technology, like throughout history, you know, the brain has been seen as hydraulic fluids, as a information telephone board, etc. Etc. And now we're up to computers, and so we have this computer metaphor. But you would have argued strongly that brains do compute. Am I correct?
[00:04:14] Speaker A: Yes, I do argue that brains do compute, but however, I put serious constraints on how would you find this sort of thing in a way. Actually, let me get back to what just you said. Like you said, people have talked about it as a switchboard of telephone, or people talk about, oh, this is like a steam engine, so on and so forth. So it's as if every new invention that we have, we try to patch a brain label on it.
This is the brain because it does something about communication and it does something about the mechanics of what we think is the mechanics of thought.
Right? And that actually is easy to put something so mysterious and, you know, amazing into this sort of like, box and then just like refer to it as, like, now we have something to talk about, right?
I would say in terms of, like, whether brain is a computer or not. Like, you know, we have to go back to the basics of what the meaning of computation is. How do we define it rigorously, right? So that's actually part of the reason I started developing this sort of like theory of physical computing. Because, you know, you're talking about physical objects that are doing some sort of like information transfer, information processing also, and so forth. And then at the end they do something that they use that information to do something with it, right? So that's so that's the sort of, like, fine detail that we have to, like, really be careful about how we define this.
And in a sense, it's not just like we have to pay attention to brain, the one that we share, but also about simpler systems. Simpler systems are doing something with the information like bacteria, amoeba, slime, mold, all sorts of things throughout the hierarchy of life. They are trying to pick up information about environment and utilize that information, make an abstraction of that information in some way with the machinery that they have. And that machinery is not some sort of like a magic dust in there, right? So I believe in the physicality of this life and. And the mind and everything. Right? But. But that's the machinery that is doing this. And at the end, they have to be able to transform this sort of abstraction that they create from this external world to regulate their internal states and also their interaction with the outside world. So I think that's the sort of, like, umbrella of where you have to tease apart what is computation, what is not. Because some of them could be just information relay. Not every information relay is a computer. Right. It's not an actual description of what a computer is.
So. And that's actually what motivated me to, like, move in this direction and develop the theory of physical computation and so on and so forth, which I'm writing, you know, a bunch of papers. Some of them are very much.
Yeah.
And more which are coming out, and they're basically like, you know, I've written a lot of these things and, you know, I'm now trying to polish them and make sure, like, you know, people can read them.
[00:07:18] Speaker C: Yeah, yeah. Well, there's a bunch on archive right now. And of course we're gonna. We're gonna talk about a few of them and I'll link to them. And we've already discussed that.
Depending on how far we get in this conversation, we'll have to have you back. Because, you know, these are. You're almost like Michael Levin. Are you producing a new paper every week? Is that the. Is that the rate right now?
[00:07:35] Speaker A: No, actually. So I work with Mike, too, and we have actually papers together. In fact, actually, I just sent him a new paper. Just. Of course, Mike is an experimentalist, and he's a fantastic experimentalist.
And the way actually I came across Mike is that I used to be in MIT physics. I used to be postdoc over there, and I used to hang out with some postdocs at the Physics of Living System.
And one of them turned out to me. We used to talk about all sorts of things from information computation, all sorts of things.
Dynamics on limits of information processing, living systems. So one of them told me like, hey, I went to this conference was Neurips. So. And this guy gave crazy talk.
You, he's a madman of biology. You're a madman. You should talk to him. So a theory, Mike. And you know, we ended up like working together and then, you know, the pandemic hit and you know, things change, life change and so on and so forth.
So Mike is an experimentalist and he's doing great experiments.
I'm a theorist, I have worked with experimentalists and I continually do so. So I, I have one leg in the experimental world. In fact, actually I do some experiments every once in a while. So I, I have basically done experiments in everything that I'm writing about because I think that keeps me connected with the limitations of the measurements. And rather than just like, you know, going and making some toy model that has nothing to do with reality, I rather like, you know, say something about what's out there, what we can measure, and also where the shortcomings of the measurements because some of these measurements, they have some limits. But can we use theory to kind of push the boundary and find ways to kind of guide experiments and suggest new routes?
[00:09:27] Speaker C: Yeah. Okay, well, I want to come back to Computation, even some more kind of introductory thoughts and remarks before we really get into the weeds of.
Of your. So I will start kind of with your paper about neural computations, but can you just in broad strokes, I guess, discuss kind of your approach and your worldview? I mean, on your website you call it the N4 approach. So maybe discuss just what that is before I then really badger you about computation some more.
[00:10:00] Speaker A: Okay, sounds good. So as I said, my view about computation is that I don't think it's something so specific to the brain. I think simpler systems do it.
Even other cells in our body are always doing it, but it's limited.
So that's the sort of general framework within which I operate.
And then in terms of N4, so I have neurophysics, neurocomputation, neural dynamics. And what I say is that neuro AI or bioinspired intelligence builds on top of this. And the idea is that in the neurophysics part, or you know, biophysics, for people who work with like single systems, single cell systems, so on and so forth, is the part that we try to really link like molecules, actual molecules, like you know, the operations, the dynamics and whatever is happening over there to some processes. Right. And we have been successful in, in that space to a huge degree in terms of like, we know a lot about ion channels, right. And surprisingly the, maybe the last and first big successful theory of science is still at that level.
[00:11:09] Speaker C: Yeah,
[00:11:10] Speaker A: but so that, that's the part that, you know, you want to know what are the properties of the system from physics lens. So, meaning that like now we, we're looking at measurements at a massive scale, right? So we want to know what are the organizing principles of this. So how do the spikes like get produced together? Or how can you find patterns in like spiking or in local field potential? That tells us something about the collective dynamics, right? That, that's the, so of the neurophysics part. And of course the dynamics part comes in as soon as you're thinking about a system that has a state of space, moves around, visits different parts of it depending on the problem that you're trying to solve. Or sort of like modulation that happens in the background. For example, the neuromodulation is a huge aspect of it because it really tunes how system responds to different things. And that kind of creeps into this sort of like setting up this state space where a collection of neurons can operate a new regime without changing anything else. Like, you know, the machinery hasn't changed, like if you think about it in many ways, right. But now some gears and knobs have started twisting because the neuromodulator comes in and then the system now can enter a new state space.
And then the other part of it is the neural computation in some sense for a lot of people, maybe like, you know, maybe especially people who come from like the systems neuroscience aspect, an old school way of, they're all after like, you know, code in the brain and so on and so forth. Right. But that sort of illusion is quickly disappearing in a sense, because think about it like, you know, they're all trying to say, okay, what is the, you know, the algorithm here? And maybe it is not about the specific algorithm, it's about like, you know, how this network of neurons can do.
They're doing computation, they're doing information processing and everything and they're solving a task. But we can look at it from different perspective. And that's the, you know, the rise of neural networks as very successful machines have kind of like shaken some of these people to the core perhaps.
And I think that's, that's one other aspect. You know, what, what is the neural computation here?
And I think that's the part that we want to be very careful about. Like it's, we want to Say it's the computational principle that basically emerges from the neural system. You have information through, you have coarse graining, so you have abstraction of the lower level computation. As I enter newer or higher areas, especially in mammalian cortex, we have a visual hierarchy or auditory system and so on and so forth. So we have this, we have the state dependent competition, we have thalamocortical control and this, these are really, to me the main things that we need to go after because that's where we can actually find how the system is controlling the flow of information throughout. And then the idea is that if you put this all together, the Neur AI is the umbrella around this. How do we go from this aspect and then into sort of designing biologically inspired architecture? How do we understand.
Or the other way around, have we used mathematical and computational tools to understand physical computation? And that physical computation could be a level of a single cell or single neuron, like, you know, how they process information. And of course the single cell is doing something limited, but like E. Coli swims toward some sort of like, you know, nutrients and then regulates itself and so on and so forth. Right. And then how do we use that information about understanding this? Like little, little ones in, in a circuit and an assembly, right. I can have a bacterial biofilm and it's. The bacterial biofilms are doing something amazing about environments. Or you have a tissue that is doing like, you know, and none, basically it has a non excitable membrane, but it's still doing something in terms of maintaining some sort of like memory of its own shape. Right. And that's actually, for example, the work of Michael Levitt and.
Or we can think about like how all these things can jump into the bigger level. Right? Like that's. Now we're talking about cortex, for example, we're talking about mammalian cortex and a much higher, you know, level of abstraction and computation and, and everything that happens in there. Excuse me.
And in that case, there's one key moment is that like, if we want to know this, we have to be able to understand how these components are working together and working individually. What are their individual properties and how they, what happens when we compose them. Right. Like, you know, we know something about ion channels, okay, we have done like tons of work about ion channels and we know a lot about them. But now as soon as you put a lot of them into a cell and then, you know, how they distribute and so on and so forth, that goes to the next level and then from there, now you have few neurons connecting to each other. Now that goes to the next level, and they may have, like, different cell types. Like, you know, this is something that is largely ignored in a lot of, like, you know, theoretical analysis of data. Right? And because the tools didn't exist for it, and now it's only now picking up that we are able to, like, you know, go and tag neurons and understand them or being like, manipulate them. So we are gradually transitioning to this era where basically we can come in and look at the system, look at components, understand something about a component, go to the next level, understand something about, like, you know, the cell at that level, understand something about the network. And. And that's actually where effective theories matter. So we need to be able to build effective theories that can give us the right interpretation at that level without basically detaching us from the reality of the lower levels. And I think that's key point here, right? Because somehow neuroscience started the other way around. It started like, you know, with philosophers, you know, everybody thinking about, like, what the brain is or the soul is, what the consciousness is, and so on and so forth. And then we had all this sort of vocality capability of explaining it. And then we started like, teasing parts. But, you know, there's a massive disconnect. Right. And for that reason, I think somehow having this sort of like, umbrella that you don't let go of these other levels that we can have a solid description, but yet we can do something more is very important. It's almost like, you know, Philip Anderson's more is different. Sort of like paper. Right?
So that's the sort of like, umbrella within which I work.
[00:18:03] Speaker C: Yeah. Okay, you said a lot there, and I wanted to interrupt you over and over. Okay, so one of the things that you said was, you know, we now have the, the tools, and I thought you were going to say, because you're developing those theoretical tools, but that's. So that's what I thought you were pointing at. But then you pointed to the experimental tools, which is what we also need.
[00:18:21] Speaker A: Yeah, absolutely.
[00:18:23] Speaker C: So, okay, but, and also, I'll just, just note that you said that neuroscience has not historically gone across these different scales across these different levels. And that's true, because I think that that's difficult and people are really comfortable studying one level. And you got to figure out the ion channel by itself, let alone this distributed, integrated, interconnected system of ion channels.
You can think of it as a big mess. So a lot of the work that you do is tying together those levels and going across level. So it's really cool.
Okay, but so set Me straight here about computation, because you know what came up, One of the things that came up in the workshop and frequently comes up is, well, of course we don't mean the brain is a Turing computer.
But then you start to get into the triviality of computing like you write about. We don't want to say that the rock is computing just because it has like transitions. So there has to be something special about it. But is computation not well defined or is the definition changing over time like all other terms that we use in neuroscience, consciousness, memory, etc. All these ill defined terms that we use.
I mean, part of what you're doing is you're formalizing a different notion of computation. And of course you point to all these different notions of computation throughout history.
So is it just my, you know, I brought up the worry that when. And you've written about this also in your papers that and on your blog, that like someone will say, yeah, okay, well this circuit computes, for example. Yeah, but they don't really. It's like a filler term because they're not formalizing what computation
[00:20:18] Speaker A: is. Exactly.
They go to this sort of like hand waving sort of arguments. Right. But that, that is. And they become a true believer that it is or it isn't.
Right,
[00:20:29] Speaker C: Right.
[00:20:30] Speaker A: So
[00:20:31] Speaker C: my concern is like, when you say like the brain computes, that the word computation just gets applied to whatever the brain is doing. You call it computation. And that kind of trivializes the notion of computation also. Exactly. You avoid that.
[00:20:43] Speaker A: I think that's a big issue. Right. Like that we don't want to open. You know, if computation is everything computes, then it's a useless word, then why are we even saying it?
If I throw a rock and then say, oh, you know, there are papers like, you know, the rock computer's trajectory or does the solar, you know, know what compute its, its future state? Like, I mean, if that's computation, then like, okay, so everything that moves, everything that changes is computation. So not every state changes computation, not every information transfer is a computation. And that's actually the key point here because in that formalization and you know, I don't want to throw out jargons here because, you know, I don't want to bore like all the audience and you know, if they're more interested, they can go and read. Yeah, paper. Papers are deep with the bad. Yeah, yeah. And I try to bring different levels of descri description that, you know, people can the video.
[00:21:35] Speaker C: So, so I'll just, I'll tell people like you're Making these videos that are kind of explainers of the concepts. And they're, they're fantastic videos. They, your papers are very dense. The videos are also very dense. Like, even though you're putting, you're simplifying things, there's so many things that you've. That you've included that, you know, I need to watch it a few more times. But they were very helpful and they're well done.
[00:21:57] Speaker A: Yeah. You know, it's funny because I think that's. That's true. Like, my papers are very long in general. Like, I don't know, generally I don't write short papers. I like very long papers because I want to, like, really get into the details, right? So it's not that I cannot write short papers, but I think I don't like short papers. A lot of them are just like, you know, they hide all the real stuff.
But, you know, back to that question, back to that question about what is computation? Here is in, in this case, I think the way I have created this sort of framework actually, initially, again, it came from thinking about a brain. But then I decided I need to go to something to a simpler scale first. Describe at the fundamental level of what it is that the physical competition is.
And I want to point out to the arrow of description, because the definition of the competition, we have like the Turing church, that sort of definitions that we have stuff, and they emerge from thinking about Hilbert's problem, right? Like, Hilbert came up with this set of 23 sets of problems, famous sets of problems. And I think it was 1912, but, you know, I may be wrong about it.
So. And he said, basically, okay, so these set of problems, can we basically, can we come up. So 1930 to 9 12. So we said, can we come up with a way of, of understanding if we can like, you know, if you have a set of mathematical problems or basically, can we figure it out, like automatically can we find out, like, you know, have this way of completely answering this sort of thing, right? And you know, I don't want to go to all details of it.
And then, you know, the response to that was that Elizabeth Cherish and Alan Turing came up with two different ideas and they both point out, say, like, no, we can't, right?
And then from there, this sort of like, operations here is basically the problem that you have some sort of procedure. And this procedure, of course, for mathematicians is just a set of like, you know, you chain a set of functions and so on and so forth. You can think of it that way, right? And can we basically Know, like how to get to the end of this.
Because if you know, then you can actually solve everything, right?
So that's the, you know, background idea to, in a sense, right? And in that case you have this sort of trajectory from basically going from this description to the sort of like the operations to the sort of machine that we designed to do those operations. Even if we know, you know, Turing complete machine doesn't necessarily solve everything, right? So that's the idea in a theory of physical computing, the arrows are reversed, which means that now you are actually handed a physical system, right? You handed the physical system. And you want to look at this physical system and you want to go back from it. You want to understand this physical system, right?
And, and figure out what are the degrees of freedom in the system and how does this system go from the sort of state transitions that it can carry to the sort of operations that interests us.
And that's basically the idea of this theory of physical computation. And that basically you start, instead of going computable function to model of computation to physical device, you go from a natural system to physical operations that it can carry and a mathematical abstraction or the abstractions that build on top of it. And that actually brings this whole idea of the whole Pandora box opens basically at this point. Because now that you have a physical system that you can find its trajectory, the evolution of its states, then you can link other systems to it.
Now you can have linked dynamical systems that have physical underlying nature. And then each of them has certain degrees of freedom. And of course, you can build it in a hierarchy. You can build as a nested sort of thing. You can have multiple realizability.
All kinds of things actually emerge out of this, right? And that's actually when you look at it from the other way around. So instead of like, you know, because a lot of people, when they talk about like, is brain a computer is E. Coli a computer is such. And so is a computer. They go the other way around. They are thinking about, I have this sort of like, function. How do I shove it in there? Like, you know, like, let's build a DNA computing system. Like, but that's, that's not what the physical theory of computation is. This physical theory of computation is that like, let's have a clear boundary between physical processes and abstract processes that emerge from it. But these abstract processes are embedded in those devices themselves, right? And here it can be that, you know, they're done in a way that another layer, for example, you have a chain of reactions, regulatory systems and a bacteria Right. And then another chain comes, combines with this and then has some abstraction of the underlying level. So let's say the lower level, the lowest level, right, Has N degrees of freedom.
But by the virtue of the operation that it does, it does some. And then the way that is linked with another set of molecules is chain reaction in like, you know, in that machinery of the cell. Now you have an abstraction of those large degrees of freedom to fewer state. So now that by itself is that a sort of information processing that is more than just relay of information, right? So now you have, what we know is in the mathematical formalism is the functor that takes the transition of the evolution of the system from the physical state to the abstract space. But here the abstract space is itself embedded in some physical machinery itself, right? And of course, at some point, one observer can come from the outside, poke something in and connect it to some external device and it starts having its own abstraction of that system. And that's what we do as neuroscientists. For example, we put an electrode in there, we measure something and then that measurement could be a measurement at the level that is relevant to the competition we're looking at, or it could be actually wrong. It's also a possibility, as I was mentioning earlier. It's only now that we are basically being enabled by the advancements in technology and tools to look at all the other things. Like you look at, you know, the activity of industry cells, look at the activity at the level of dendrites. I think this is just like, you know, it's going to open up this whole field wide open to a completely new frontier.
So that's this, that's why the angle, I think this sort of like the lens that you can use to look at the computation at physical system matters. And, and you know, this story also, it didn't come from actually neuroscience anyway, right? Like some people started like building this sort of, like initially philosophers have built this sort of like these ideas about mapping. Like, you know, Hilary Putnam, right? Like he was a mathematician and a philosopher and he came up with the idea of this sort of mapping between different states. But then the problem is that like in a lot of these, these descriptions in a definition is too loose, right? Or is so bounded by the objective of linking everything to brain and minds, right? Like that goes into the category of semantic mapping and so on and so forth, that it's, you know, load bearing, right? So it doesn't enable you to build real theories because you jam in all this extra stuff or it's so Naked and bare that everything is computational, it becomes useless. Right. So I think this is the key point here. How do we approach this in a way that we have that rigor?
[00:30:27] Speaker C: And do you.
Well, before I ask you about category theory, because this is sort of category theory is a necessary ingredient to this. Actually, I'll go ahead and ask you about that. I mean, you use the word functor is category theory without it, it. Could you do.
Could you explain the computational aspects?
[00:30:48] Speaker A: There is something about category theory is that category theory is a language that shows you how you can basically transfer a set of basically morphism and objects and relationships between them. So you have objects and the relation between those objects. How can you transfer to another state?
Right. So think that way.
I don't want to throw in wrong words, but basically you can have a category, you can have another category and then you have a functor basically that maps every object to object, every morphism to morphism or relationship between those objects. You can have basically a junction and so on and so forth.
That's what used to show the multiple realizability.
Or you can have natural transformation, this is a technical term and capillary theory, that you can basically lift something from one level to another level. Right. So it's a set of tools, basically.
It's not that.
It's not an ornamental tool in a way. Right.
It's not a decoration, it's something that you can use it to map from where you are and you have a description of it to another space where you want to have a new description of that system. Them. Right? That's the idea. And in fact, actually the recent paper that I released about wave computation, much of it I go without using category theory. And that's intentional because I want to show that you can actually look at the systems. So just to give a background to the audience, there was a very interesting paper that was released by a French group and it was published in prl.
And the idea was that they had a sort of like a bath of liquid and basically you can vibrate this bath, some sort of like oil, right? So they can vibrate this bath and below some threshold and then you could put drop a droplet on it. But the droplet doesn't like, you know, diffuse and go away, rather actually bounces on this. And each time it bounces on the surface, it leaves an imprint. So it leaves an imprint. But that wave that is being created now starts pushing the wave and the waves goes back on it. So you have this. It's very interesting because you have feedback, you have memory, you have writing, all sorts of things. And they show that even you can actually push the system.
And then it can have a trajectory, a completely chaotic trajectory.
And then if you shift the phase of your wave now, you can actually go back and erase the path which follows exactly the same thing. So what is this system? And they call it. This is a Turing complete system, these wave walkers. Right. And in that paper that I released later about, about this, I basically go in, I build a model about like the sort of like physical waves and how this sort of like wave is registering the movements of the walker, how the walker does this. So they do this as experimental test, right. So and I now build a theory around it. And then I from there I am able to build again link to that sort of theory of physical computation through like what it is that this system is doing that and how the coarse graining looks like here because now you have like basins and these basins of where your space is now, how you can have abstraction of it, so on and so forth. And at the end I end up in something which I call the computational closure.
[00:34:22] Speaker C: Yeah. And
[00:34:22] Speaker A: that's actually what separates basically systems that have a sort of abstraction that is built in, in there in their own like, in their own like, you know, structure. Like, you know, that's like our brain, right? Like our brain does that. Like our brain has computation closure, but not. Well, that's. Every computing system has it. That's also another thing, right? Like not every computing system has it. So even that wave walker is doing something. It has all. It has a Turing like primitives, but it's not an autonomous physical computing device. And I think that's a huge difference. Right. And that's also what can, you know, clarify and get us rid of all these sort of like crazy notions that come up. And you know, everybody gets confused by, oh, hey, LLMs are like, you know, conscious and all these things that come out, right. Like, and people argue about it or they write paper about it and you know, we need to be much more precise about these things. And that's the sort of effort that.
[00:35:23] Speaker C: Yeah. So you think that the computational closure would be essential for something to be conscious, I would presume. Right.
[00:35:29] Speaker A: So I have a little bit of allergy about using the word like consciousness, because I know, okay, you can jump into this, like, you know, all right, what is it? And of course some theories out there and you know, from integrated information theory to stand beyonds like global workspace and so on and so forth.
The problem is Again, they are layers of explanation that are easy to better.
And I think that's not a good way of approaching scientific problems, because if you have a theory and you can change your theory to accommodate everything, that's not a good theory. Right. So I think that that's what neuroscience needs. We need to go to a new level. Basically, our theories are actually. You can put it to test and refute it and reject it and just, you know, move forward.
Physics have been really, you know, in physics, scientists have been really successful about this, and we are hopefully entering that era in neuroscience.
So,
[00:36:38] Speaker C: yeah, okay, so we'll steer clear of consciousness, but going back to the closure idea. So correct me with what I'm about to say. So what you were saying with the wave walker experiments is that the experimenters can flip the phase, can change the phase, right. And then new properties will emerge. But the point about the computational closure is that phase flipping mechanism, we'll say, can't come from the outside. And it's the same with the way all neuroscience decodes neural activity. Exactly. You can decode anything you want out of that neural activity. Almost.
[00:37:13] Speaker A: Right?
Yeah. And I think that that's exactly the point. And, you know, that's why I say that that wave walker has turned primitives. It has memory, it has erasure, it has writing, Right. It has all those components, but it's not an autonomous component. So I build a com, you know, a run, a ladder run, basically saying, like, where you are in this, you can. You can be at a place at the lower rung of this ladder, and then you have memory, right? And if I knock on this table, it has a memory. If I bend it, it has a memory, right? Like if I bend something, like, you know, spoon magic, like bending a spoon and so on. And so it has a memory, right? But then you go to the next level, and then now you have. You can have computation in some sense, but it's not necessarily an autonomous computation because it's the experimentalist who goes there and switches the action over there that closes the loop, right? Yeah. And that also is the case about when we are looking at it as an observer. This is actually very interesting because now it really creates this sort of thing. Like, how do you deal with a system with a physical computational device that is doing something and you want to observe it?
Obviously you're going to interrupt something because you're pushing things in there and manipulating stuff. But at the same time, that observation, you have to be careful about what, what operations that are using or what deductions that you have is yours versus the computer. Like, you know, if again, back to that sort of like a neural coding sort of problem that I pointed out to earlier, we can, you know, build all sorts of decoders. Right.
With all sorts of decoders, that means an information is there in the system.
It doesn't mean that that's what the system is doing. Right. Like that's the machinery that we have built on the outside of the system that is giving us this illusion of that's what the system is doing. So to understand that what the system is doing, we have to look at it from a very different lens. And I think that's why suddenly a lot of people have lost their identity, in a sense, and they are facing an identity crisis because they don't. So if we were looking for all these neural codes for this long time and now you look at this, I don't know, some neural network trained on a bunch of videos and suddenly does sort of like image recognition and all sorts of things, then where's the code in that one?
Right. Well, I mean,
[00:39:47] Speaker C: in a sense, these big neural networks are just big decoders. So neuroscience is in the business.
So you record some neural activity, or a large swath of neuroscience these days and has been for a while. You record neural activity. You decode some information using a neural network, your favorite bespoke model or whatever it is, Right? Yeah. You decode the information and then you say, ah, that's how the brain must be doing it. And that's, that's the big error. Right, that you point
[00:40:15] Speaker A: out. Exactly. And, and you know, we can learn a lot from, from those systems, how they operate. Because now we have, we can look at under the hood of those systems in a way that it's not easy to do in neural systems, in real biological systems. Right.
And we can build methods that can give us some way of like, you know, decomposing the actions and properties of those networks. But it doesn't mean what we learn from there is necessarily, you know, transplantable into actual biological systems. Because the degrees of freedom of biological system is like way more. It's just beyond those systems. Right. Like, as I was saying thing, you know, look at the dendritic computation.
It's something that, or like there are astrocytes, there's inhibition, there's excitation. Oh, it's not even like, you know, one type of inhibitory cell in there. They're like different types of. In every cell. Yeah. So the main point is that we don't mistake.
We don't mistake Make a mistake of basically thinking that this object that can do something that behaviorally.
Right. Is similar to this object of interest is the same thing.
And this is very crucial because a lot of people, they take, they, you know, they're training the network and they say, hey, look. Yeah, this is what the brain is. No, that's not what the brain is. You're doing something is. You can. You're capturing something or what the brain does.
You're tapping into some aspects of it. But that's not. And there's no equivalence here. So. Yeah, and, you know, I think the value here is to think about if we had access to all the information in a given system, which is in this case, this artificial neural networks, what do we learn from there?
How do we build models, how do we build methods that can give us a way of describing these systems? Then that is going to be useful to look at the actual biological system.
But of course, then we have to come up and, you know, expand those methods, expand those models that they would match the realities, the physical reality of the system that we want to study.
So that's a long road that, you know, we have to basically build methods, combine it with experiments. Right. And I think this is why experiments are super crucial, because it's. It keeps us bounded in reality.
I can sit down here and make like zillion, you know, mathematical and theoretical models about like a point neuron model that has nothing to do with what the actual neural system does. It's a good abstraction in some level, but at
[00:43:03] Speaker C: some level
[00:43:03] Speaker A: it misses a lot. In fact, actually, like, I just put out a paper not long ago, but I think I sent it to you too.
[00:43:08] Speaker C: The coastline. How long is
[00:43:10] Speaker A: the coastline of a neuron? So that's kind of like points to this issue.
[00:43:15] Speaker C: Yeah, well, that reminds us that we can easily slip between scales.
It reminds us that everything is at a different scale. But it's easy for us to slip in talking across scales as if we're talking about the same scale.
[00:43:31] Speaker A: Exactly.
[00:43:32] Speaker C: Yeah. Going back to the closure idea, I think, that you've written about.
So AI, quote unquote, AI has sort of driven, as a neuroscientist, has sort of driven me crazy in terms of calling units, neurons and sort of usurping a lot of the language and going back to the closure idea. So now we have AI, quote unquote agents. And I think this drives you a little crazy too, because there's nothing actually agential about those because they don't have the computational
[00:44:04] Speaker A: closure that's such a Misnomer. It's terrible.
Yeah. And I think this also, it almost bleeds into how we treat them because they say, oh, these agents, they got out of control and they went. Hacked somebody. No, they didn't go out of control. You didn't put the right boundaries.
Just again, it's not that they had like this sat down and decided, all right, let's start a revolution of hacking. I don't know what was it? Hugging face. They hacked something. I don't remember what it was, but like, you know, and every day something comes. It's as if they're giving them agency. And I think that's a. That's a wrong issue.
So agency goes back to the sort of computation closure. Right. And that's actually, that's why I avoid using labels like consciousness or something like that. Because I think what we need to deal with is that we need to deal with the systems that they are biological systems that interact with the environment. They have a boundary. They can sample information from the environment through the machinery that they have at their disposal. And the machinery is very different from system to system. In simpler system, it's just some sort of ligands and receptors. And as you go through, like, more complex systems, then you have the evolution of brains and eyes and so on and so forth, like, you know, all the whole machinery that was built around it. But then at the end, they use that to make an abstraction of the environment, do some prediction about the environment.
Basically, that is their internal model of the world, right? So they are building this internal models of the world that they can act on it, right? And the action on it goes in two ways. One is through the interaction with environment, and two is regulating internal system. Right. We have homeostatic systems. All you have set points, right?
And that's the sort of knobs that the systems do use all the time to kind of like maintain their life, right? And that's, to me, that's agency.
The agency comes from this aspect of having this sort of computational closure built into this system that now you can have an understanding of the environment and then from there you can act on it and you can control yourself. And now you have the internal model of the world that's, you know, now I'm going to use the bad word, the forbidden word. That's where intelligence, consciousness, agency, they're all like, are different. Like, you know, you can. You can put labels on it, right? But
[00:46:38] Speaker C: all the bad words
[00:46:38] Speaker A: I want to be precise about, like, this is what we measure, this is what we can do.
[00:46:42] Speaker C: Yeah.
It Kind of reminds me of.
So Henry Yin comes at this from a control theory perspective and he says, he writes about the crisis in neuroscience that essentially it's the same sort of thing, but from a control theory perspective where the reference signal or the set point is set externally by an external observer. Right? And so you need a control system that sets its own. Own reference signal internally somehow. And that's where the closure of computation comes in.
[00:47:11] Speaker A: Yeah, yeah, exactly. And you know, we shouldn't forget that at the end the whole nervous system is there for the regulation of our body, right? That's what the nervous system does basically. At the end it tries like, you know, it's a vast machinery, right? You have all sorts of systems interacting with each other and they have to be regulated, otherwise it's going to fall apart.
So, you know, we build this like crazy complex model of the external world and in our minds, but at the end it's just for us to be able to interact with it and maintain the internal state in a way that we stay alive. And that's why like, you know, we get hungry, you know, we get tired, et cetera, et cetera. So we need to. So you can still see that the operations that we need to do at the end is not that different from like, you know, bacteria, right? Okay. They're reproduced by splitting, you know, it's different, but they eat by, you know, the mode of eating is different from ours, but at the end it's the same thing. So we just a collection of cells.
So that's the sort of like multicellular system that now you have specialized organs that do something better than the, you know, single cell counterparts.
[00:48:26] Speaker C: Well, does a single cell have computational closure or must it have computational
[00:48:31] Speaker A: closure? In fact, actually I'm working on something on E. Coli to show whole chemo taxis falls into this sort of physical computation direction. So I'm building theory around that one. And that's actually where basically I link the systems biology descriptions that exist. Right. Because there are a lot of systems biology also being carried out. But the problem there is that then again, there's a disconnect from describing the dynamics and so on and so forth to what the system does as a whole as the sort of log information processing unit. And that's actually where we need a theory like this, basically.
[00:49:14] Speaker C: Well, so I'll ask again, I guess. I mean, did. So does a bacterium or E. Coli, does it.
[00:49:18] Speaker A: Yes, it does to me.
[00:49:20] Speaker C: Does need it? Does it have. Does it have it and does it need it? Is it necessary? Are two different Questions, I guess.
[00:49:25] Speaker A: Well, you know, it has evolved to do so and I think, yeah, and it does need it because the bacteria does need to process and say, okay, where's the sugar? Where's the gradient of the sugar that I need to swim to? Right? Because if it doesn't, it can die. You're going to swim in the wrong direction or it's going to tumble around too many times and it's going to run out of fuel.
So that's a major issue for bacteria. That is the issue for bacteria, right?
[00:49:55] Speaker C: Yeah. I guess the question is related to the idea. We originally were talking about brains and computing. And that's the sort of tired question now. It's like, is the brain a computer? Right.
But if the bacteria is also implementing computational closure, then you might conclude, well, there's really nothing special about brains, then if the computational regime can just be imported, transferred across to the single cell.
[00:50:24] Speaker A: But there are different levels of this, right? Like, know, the different levels you, you can have the state.
It, it means that you can put these systems into this sort of like space as like they are equivalent in some ways, but some are doing like far richer sort of abstraction. For example, just remember what I said, like, you know, in terms of using the sort of refinement and real multiple realizability and so on and so forth that we build into this like theory is that you can explain the abstraction and the abstraction levels that you can create.
So, and that is different in a bacteria from a collection of bacteria, right? So even there you can have a bacterial biofilm that it does something in a single bacteria cannot do. And this is very interesting, for example, because they're very cool experiments. They show you can have a bacteria biofuel and then in it you start like having pulsation of calcium, potassium waves and calcium waves that go round. And you know, because the bacteria at the boundary are the ones that are actually sensing the external world in a way, right? The ones at the center don't. So this is very interesting right there. Even in such simple system now you have emergence of a new property, right? Not all of them aren't going to do the same exact thing, but through this sort of communication, through waves of ion ions that they create, this pulsation is now carrying a sort of information for them about the environment. And this is, you know, again, that's sort of like you build abstraction on top of abstraction, but all in physical world. So there is no spooky like, you know, demon sitting there and doing these things, right? It's, you still you know, you're bounded to molecules. That, that's the reality of physical life. But that's fine.
Now you have construction on top of it and then you re. And that's actually another aspect. Like we reuse these components to do something, then they do some sort of like, you know, computational or information processing sort of thing. And then now you can use them. And that's what evolution does. Like, you know, reuse and use and adapt and modify them to build better and more robust systems. Right. That's the test. The, the objective of evolution is to do that. So it just creates more.
[00:52:44] Speaker C: Yeah. So I want to get to. Because you write about this specifically in terms of some cortical canonical motifs, Right. That the cortex performs and linking them together. Before we do that, I had the thought while you were talking earlier about your approach.
So there's David Marr's famous three levels approach to understanding things. Right. And so you start with, with the computational level, which is the goal of the organism or the bacteria or something. And then below that is the algorithmic representational level, where you work out what algorithmic steps it would take or what algorithm you should implement to accomplish the goal or the computation computational level. And then finally you go down to the brain and say, well, it must be happening in there, so we have to figure out how it's happening. And I'm.
And classically, that's the way it's like top down, that's the way you're supposed to do it. There are people like Yuri Bujaki, who wrote the Brain from Inside out, is his most recent book. But he says, no, no, let's go from the implementation level to the computational level. Let's see what the physical tissue is capable of. And that can constrain what we can say about the algorithmic representational level, which would constrain what we can say about computational level. How does your, your approach relate to the Mars levels approach?
[00:54:11] Speaker A: So, yeah, it's interesting because you're mentioning two things which are, I can see the relationship over there between this sort of mindset of looking at different levels of the process carrying forward something from past to the future, and how you can kind of,
[00:54:32] Speaker C: you're, you're interested in like the, what is available physically, like what the physical tissue can do, you know. Exactly. But you're also interested in the computation. So that, that's why. That's
[00:54:42] Speaker A: Right. So that. And that's. Yeah, exactly. I think that's what we need to be able to kind of like move away from like putting random and loose labels, right?
And in terms of the general principle of like, you know, Mars principle, like, you know, you have three levels and so on and so forth. And people write a lot about it and then they, you know, they go back and forth or whether they like it, they think it's useful or not, and so on and so forth. Like I had things like Jonathan Pilio and Martelling for example, they had this sort of back and forth papers on it that they agreed to disagree but then they ended up like agreeing.
I would say so. So I think that's a very interesting approach.
It might be more useful in terms of how we anchor looking at behavior and wanted to link that to some description that we find at the lower level. So it's very useful from that perspective. In fact, actually with a colleague of mine now there we talked about it, as you know, we won't go into details of it, about how to build knowledge graphs around this idea and so on and so forth.
But I would say I'm more on the side of saying we need to observe the actions of the system and different states and that like really that's just the bottom line. Biophysical measurements and then physiological measurements, right? So we need to do this at different levels and we need to understand what are the degrees of freezing on the system, what are the effective theories that we can build on top of this. And then by doing this, this then you can have a description of the abstraction that is built in the system, right? So if you want to think about it like the relationship to Boujaca's idea or Mars idea, or for instance free energy principle, it has, it touches all those things in some ways, right?
But all bounded by the actual physics. I said like, okay, so now we can actually talk about it.
We can have a measurement here. In this measurement we have this, this system of interest and it can visit these sort of estates and it has these degrees of freedom. From these degrees of freedom, now you can build another layer on top of it that now you can have an abstraction of that physical layer to a new one. And now you have the evolution of the physical state in this system.
There you have the emergent layer of abstraction and then so on and so forth. So basically you can see it's almost like layers of an onion, right? It's like you peel more and more and more. So, so. But all is about that scale, right? Like about how you have properties that emerge at a scale that are not a spooky emergences because that's a key point, right? Like a lot of people who talk about complex systems, they may end up with like, you know, being a little bit like goofy. Right. Because then they talk about like, oh, this thing is just like emerges. I said, okay, what does emerge? Like, yeah, but on the other side, you also have the pure reductionist view, which has been a play of neuroscience, I would say, because it's, it's a big issue. Right. Like everybody, like, you know, you put it, I mean, let's think about it. People used to put one electrode in the brain and then start like telling us the stories about like how the whole memory and attention and everything work, but just like observing. Yes, that's what we did. Right.
But it is changing and it is changing because both experimental technology has enabled us to do more. And then now it opens up new questions, questions that we didn't need to do before.
And also we have now this equivalent sort of advancements in AI that is giving us a clue of, okay, well, we can build other systems by just purely connecting bunch of nodes and then they do operation, emphasis on the brain, then what the brain is for. Right. So this is the sort of the two sides that are helping us to sharpen our questions in a way, if we want to understand it properly.
[00:59:00] Speaker C: Okay, very good. Okay. More is different.
So this idea has come up multiple times. I mean, you just used the term emergence a minute ago as well.
So let's talk about the work that you've done. Maybe you can introduce the two canonical motifs. Right, so there are these. I will broadly introduce it. I'll let you get down a little bit more in the weeds. Right. So there are two of various canonical motifs, one's called divisive normalization, that have been shown to carry out sort of a normalization procedure among populations, small populations of neurons. And another motif is called a winner take all motif, where essentially a decision is made because of the inhibitory and excitatory neurons and the way that they interact.
Some population of neurons eventually wins and takes all. And these are two canonical motifs that are thought to just be repeated over and over and over, like you said, which evolution does in the cortex of the brain.
And I don't know if have these two. So what you've done is asked, when you link these things, these two procedures up and make them functions of each other and put them in a series in various orders, what happens? And has no one asked that before?
[01:00:28] Speaker A: Perhaps not.
[01:00:31] Speaker C: That's one of those very. What sounds like such a obvious and simple thing to do, but
[01:00:36] Speaker A: it's not simple to understand,
[01:00:39] Speaker C: implement. Right that's
[01:00:40] Speaker A: why we're here. I'm sure people have thought about it and I'm sure, like, you know, I'm not the only person who thought about, oh, what happened.
[01:00:46] Speaker C: Seems like a Steven Grossberg would have thought about this or something, you know.
[01:00:50] Speaker A: Well, some people, I'm sure also some people will come and say, oh, we have done this before. Like.
Well, people who always. Well,
[01:00:57] Speaker C: maybe, maybe not. Okay, but, but so tell us more what happens and if, if I misspoke about what the two.
[01:01:04] Speaker A: I think actually maybe this is actually a very good point that you raised. How have people thought about it? As I said, like, I am 100% sure people thought about like, you know, what happens when you combine these things, right.
The main thing is like, do we have the right way of looking at it? Because, you know, you can try to look at a circuit and describe it in different ways, right? Like you can look at the circuit and describe just purely by the connections, right? That's the connectome, the old, like, we are the connectome, I am the connectome sort of approach. Right. Or, but you know, we know for works of Eve martyr, for example, he can have few, you know, very unsophisticated neurons in lobsters gut and then they can same wiring can give you zillion types of dynamics.
[01:01:56] Speaker B: Right,
[01:01:56] Speaker A: okay, okay, so we know that. Then we can think about, about. All right, you can have a description of a system, the dynamical perspective of a system, and you can say, all right, so here's a bunch of neurons, I connect them, you know, put some boundary conditions on and so forth. And here is going to go. I push it. It's going to go here, right? So you draw a trajectory. If it push, if it gets pushed, it goes there. So that's another aspect.
But then comes the question that I was asking in that paper, and that is more about actually.
Well, we know evolution reuses components, reuses modules, right. It's not that it just makes it bigger and bigger by adding things randomly. Rather actually reuses different bits and pieces, almost like Lego pieces. Right, right.
And then from that sort of composition, it builds something more sophisticated.
Right? And that's why, for example, we have columns of the cortex, right? Like what's happening in columns of cortex. You have a bunch of columns. Each of them is doing some operation and of course comes the whole connectivity of how they get feedback from the other parts of the cortex, from thalamus, you know, they send back to the reeling nodes of talents and so on and so forth. So you have all these Sort of things. But then deep down you can tease apart and say, all right, well now we also have this sort of like repeated motifs, right?
What do these motifs do? You can describe those motifs from the perspective as mentioned. You can describe the cell types, you can describe the sort of like wiring, and you can describe the dynamics, but none of those lend themselves to describing the operations at a newer scale. And that's again back to what we were discussing before, is that we need to figure out what are degrees of freedom at a given level and what happens when you go to the next level. Then the next level in this case is that you compose two circuits with each other, right? So you have. There's a normalization, basically, the population changes the game of that neuron's output, basically. So that's the kings of neuron. And then neuron's output changes. So that's something that was shown by David Hicker and Matteo Carrandini and so on and so forth especially. And this helps
[01:04:18] Speaker C: systems adapt to changing stimuli. Like when we walk outside and it's bright, this helps us adapt our visual experience of
[01:04:26] Speaker A: that. And it's great, right? So. And then you have also the veneer take all, basically. You have two populations competing with each other. And, you know, one becomes active at some point and silence is the other one through some sort of, like, common route to an inhibitory neuron. So we have this.
And then the idea is that, like, can we think about, like, using these things to compose them together and build a new system? So that's actually where the idea came from. Because I was thinking about, like, how do we program networks, right? That's what. How do we program recurrent networks? If you have a faithful network, it's much easier. You know, people do it all the time, especially in like a AI nowadays. They're like, okay, they have this thing and they do the modification and they say, oh, we also add this to it. I mean, also add this to it. And then you have a big thing and then you run the back prop and it works. Recurrence is a beast, right? And I think that's the holy grail of neuroscience. Like, can we come up with theories of how these recurrent systems are working? How are they, like, doing the operations? How they change their operations without constantly changing the synaptic weight and at that fast rate? Because that's just a. Biology cannot afford that, right? Like biology cannot afford back propagation at that scale. It has to do something different.
So that goes in a few different directions from that sort of like, perspective, like Tackling recurrence, right? And that's a major theme of my research nowadays, is like looking at recurrence, but looking at it from different perspective. One of them is how to compose a system, right? How do you program a system?
And to do that, that. That's the idea of this sort of like chronograph theory and holonomy decomposition that come in. And you can use those to basically think about a system that is composable. You can also factor it out, right? It's almost like, you know, you have. You can say, all right, a natural number, you can decompose it, you can factorize it, right? And you get a bit. Eventually you get to some prime number.
There's some power of those prime numbers. Now, the same thing happens to be also true in finite state machines. You can basically describe that system and break it down to bits and parts. Let's call them funnels and gears. And funnels are the ones that are basically, they have an end state reminiscent of the name funnel. And gears are the ones that actually could switch, could switch between different states so they don't end up up in some final state.
Now, when you look at the machinery of win or take off or there is a normalization, then you can actually approach it and try to figure out what are the generator explicit funnels in the systems and is there a general explicit gear and that. The idea here is that you move away from this steady input or like one push to the system and see where the dynamic goes, right? Because that's a dynamical view. Now here, the view is that you want to know, you. You harness that, of course, you harness that sort of knowledge, but you go one step above it and you say, what is the response? What, what is the computational repertoire of this system?
If I can zap it it continuously with all sorts of inputs that it can have, right? So in this case, you can think of it as words, right? Like, you know, the word input comes in and it's as if like a sentence comes in. Like, you know, one input after another, input after another, input after input, and it's changing all the time. And that's really what the systems are experiencing. In real cortex, for example, you can have an input to this one, then the other one could be off or can have input to both of them or, you know, both of them could be off. So you can create a whole circuit of like, dictionary of all possible combinations of these things and just like, you know, push the system with all this and then figure out what ends up at the End. Where does it end up at the end basically, to does it end up in a funnel or does it end up in a gear?
Okay, so that's the sort of like approach you can have here.
Then you can ask a different question. So, okay, now I know the properties of this individual motifs or canonical circuits, let's call them canonical circuits because it's a little bit above the motif.
Motif would be like more, hey, you have a connection between two neurons back and forth. That's more like a level of motif. Like here is like one layer above it, right? Now you can say, all right, now I know what the DN or DU normalization or WTA for we know take all does.
So I can figure out that. And I can see where you have. If you have a set of inputs of this sequence, and that's like the witness word, like the shortest witness, you can end up in this sort of state. So you can push the system. And you know, and this is like computationally very expensive, by the way.
It's insane like how you have to like rerun and run like zillions of configurations, but, you know, it's doable for at least small canonical circuits.
And then at the end now you can say, all right, now let's connect them with each other. Now if you connect them with each other, you can end up with a different state. So for the divisive normalization, let's say you have, you know, four states, right? For input states, which ones you push. And then for the winner take call, you have eight. So now you put them together, you have this like 4 by 8 sort of grid rid of states within which you can move around, right? And the question is now that you start like saying, all right, so now let's say it's. There is a normalization that pushes the winner take call, but then again from where to what, right? Like that matters, like who is connected to whom is what defines the boundary or the constraints of how the system as a whole is composed and operates.
So now the dynamic actually of these two systems, like expands enormously, right? It's just like a completely different beast.
But you want to look for transitions that are happening within the basically directions of only one system. So let's say if you know, it goes, if you can only see changes in the winner take all sort of like states, that is a different sort of composition than if you saw changes in both winner take all and divisive normalization. And that's the point of this paper that is showing how you have emergence of this sort of Dynamics that if on based on the sort of composition that you have and who pushes whom with a certain set of inputs, certain train of inputs, those like, you know, words that I was mentioning, you can actually push the system that now you have generation of gears out of purely funnel systems. This is quite actually interesting, right? Like, how can you create this sort of thing? It's doing something that none of them could do before, but it's also able to do this sort of like emergent computation only and only if is going through this specific route. And you know, this is part of what evolution does. Like, you know, it connects them in different ways and so on and so forth. But we can use that, we can use that actually to think about if we want to design a new system, a recurrent system or a system that has recurrent components into it, then we should be able to mathematically computationally predict the properties of the system. And that means it will dramatically change our way of looking at this system. Because instead of like relying on, all right, we need to just collect all sorts of data and you know, have gazillion labels or come up with like crazy solutions to, to do the back prop. No, we can actually go in a different way. We can program these systems just the same way that evolution did. Did, right? Evolution did it by trial and error. It didn't create a sort of like mathematical theory of how to, how to do it, but we can come up with that mathematical theory to be able to now go there, understand the systems in biology. And for example, we can, you know, experimentally can look at a certain circuit and then tease apart the connections and wiring and different cell types.
As we talked about it, these emerging technologies use optogenetics, for example, to manipulate things or like, you know, from pharmacological modulations and push these sort of operations in these estates that I was talking about in this sort of like configuration board of the compositional board, right?
That's one way is that we understand those systems better. And then the other way around is that now we want to build an artificial system so we can actually go in and play around and depending on the task of ahrust, we design the system to do that best by some minimal configuration. Or if we want to have like more general systems, they perhaps need to do something like what evolution did. Like, you know, we build some sort of like composable units that we know what they do and then we compose them and then, you know, becomes a more general sort of like a, a computing machine so that that's the Sort of approach behind this, sort of like a more is different. And that's exactly the idea of more is different, like Philip Anderson's more is different, right? Like that it says you can have symmetry at the symmetry of laws, like a very simple things, right. But then from that, from the composite, from the different layers that built on top of it, then you can end up with new rich phenomena that are not explained animal directly by the lower level. And I think that's, that's the key point here, that why, you know, pure reductionism is bound to fail. It doesn't mean that we don't have to go and measure like you know, the transport of ions through a single ion channel. We should do that, we should do that, we should measure that. We should like, you know, vary the temperature also. You know, of course these are done, but like, you know, that's needed to understand the components. But to go from those components to a different layer we need to have this sort of machinery and abstractions.
[01:15:14] Speaker C: But why would we want to use that knowledge to build better AI? If I can just make my feed forward network bigger and essentially get the same results. Is it just about efficiency or is there something, you know, is the emergence principle, does it go beyond, Is there something special about building, you know, an AI system in this way that you could get outside of efficiency?
[01:15:43] Speaker A: So the two answers I have that one, of course efficiency matters dramatically, right? Efficiency matters dramatically because nowadays you can see that they are building data centers as big as, I don't know, cities and so on and so forth. So know there's research bound, like, you know, that's something and it has negative effect on environment and so on and so forth. And maybe like someone comes in, well, you know, let's just build them in the future we'll come up with some green technology solution and we don't need to worry about it.
[01:16:14] Speaker C: But it's almost disappointing to just, if that's it, it's almost, I don't know why, but it's kind of disappointing to say like it just seems like there's, it's less special if it's just a matter of efficiency.
[01:16:25] Speaker A: So I think this is the, the idea of scaling, right? Like, you know, people are saying like, you know what, this is it, like we just do this, we just make it bigger and bigger and we have more data and eventually it's going to solve everything.
Right.
And in a sense actually it touches us personally as humans. Like okay, this is it, you know. Right, yeah, that's just, that's what I Am, but like first I don't have any problem with that. Like, because fundamentally I think, okay, yeah, we're just a bag of atoms. Sure, there's an organization and we're doing something like, you know, and it's, it's very cool. Like, you know, I get hungry, it doesn't matter, like if there's some neurotransmitters going up, I get hungry and I, you know, I have a joy in like drinking glass of wine because yes, there's taste, but I can explain all of that, but it doesn't, you know, separate me from that experience of, of being human and having experiences and so on and so forth. So I have no problem with that. And if there's like some we can make alternative things that are as intelligent or more intelligent than us, well then be it. That's, you know, that's like what does my dog think about me? Like, he's intelligent, he lives his own life. But yeah, he's happy. If I go with him like, you know, for a run, he goes for a run. He has his own experiences.
So that, that's one thing. But I think energy efficiency matters. So that really matters because we can't push it to some level and it's bounding problem. Another one is the data.
At some point, yes, we can measure more data, we can measure more data and so on and so forth. But again, we can always think about it that making abstractions of the external world is a different thing rather than just explaining everything. If you build models that are maps, which are as big as the territory that we live in, they are useless maps in a way. Right? Like you want to Google Maps because it tells you, all right, well there's buildings and whatnot and you want to go from here to point there. You just go through the road and it shows you a traffic on it and then you know how to get there. You don't need to know what's the color of the door of that room to get there. Right? So that already tells you the abstraction can get you there much faster and without cluttering your mind. So that's one, one thing, but more importantly, and that's actually a paper that I have to release it, I'm just very obsessed about every aspect of it. It's, I don't knowing version 16 of edition or something. And in that paper I basically showed that there is a fundamental limit, fundamental physical limit of what feedforward systems do.
And that is, is something that you need recurrence for it. So without recurrence you cannot solve those problems. And those are physics like actual physical problems that no fit forward system can do.
So and that one, when I release it, we'll talk again later.
But there is that fundamental limit and there's also another aspect is that can we force ourselves to think about again, like, you know, let's look at biology.
Why does biology uses this sort of like recurrence all the time? You know, whether it's a simple system, as you talked about it, like a bacteria around me bow whatever, or a very complex system, like a complicated system, because those are also complex systems. A very complicated system like our brain.
It's full and abundant with recurrence, right? It's everywhere. And again, this goes back to that idea we were talking about, about the idea of closure, that you actually build a world model. You build a world model and you have now an experience. And I think this is something that is fundamentally missing from feedforward systems. They don't, you know, you can have a large language model now tell you everything about rain, right?
It knows the temperature much better than you do by precisions that, you know, you may not know because you couldn't even connect it to some sensors. And it knows something about it, about it. It can tell you about models of the rain. It can tell you about that. It can write poems in, I don't know, it can write haiku in Japanese, I don't know Japanese.
It can do that, right? It can paint something, but rain in the style of, I don't know, Monet. It can write a symphony about it. But it has no experience of what rain is like, you know. You know what? It just absolutely doesn't know what rain is. My dog knows what rain this. He knows what it is. He has experience. Very like. So there is that sort of like being an environment is a.
Is the place where basically you need now that sort of like machinery that we were talking about. Now you need to have this sort of machine, the recurrence machinery has to come in to be able to have this sort of different assignment of environment.
But yet again, I was saying is that like, you know, there are fundamentals, mathematical limits of what these faithforward systems do. And you know, that's. I'll be releasing this. I'm hoping that will release by the end of the year, but, you know, maybe get to next year. I don't know.
So that's one thing and then the other one is actually, it's a paper that I'm working with a friend of mine. I meant to send to You. But like, you know, for some things happened in his personal life, then he couldn't finish it.
So with Baktosh Babadi, who is a computational neuroscientist, but also is a physician too, and he worked with Larry Abbott and Heimson Polinski, and now we're collaborating with each other on this topic. Basically, we created this sort of like, machine that we operationalize all things that the back propagation does, but we do it explicitly with recurrent system, no backprop at all.
Right. And the idea is that we want to show, hey, you can actually do this. It's, you know, more energy and it's a very sparse system.
It's a highly sparse matrix that can do all what back fabrication does.
But we don't need to change the weights. The whole thing actually happens. The weights are the recurrent state of the system.
Right, okay. So and the modifications of it.
And that kind of also shows you that. I mean, we are not claiming, okay, look, this is what biology does, but it's a clue that maybe that's what we should look at.
[01:23:03] Speaker C: So the learning happens in the dynamics, then the weights get updated, like after the learning. Is that the way.
[01:23:10] Speaker A: So basically think of them, we have some modules, again, the idea of composition, we have some modules that these modules are doing the basically operational. It's all the mathematics of back propagation, but only with few neurons, like three neurons doing this sort of like multiplication and so on and so forth. And the way the whole system is integrated and put together and what it shows is that it basically can mimic what you can do with single layer perceptron or multi layer perceptron, and so on and so forth.
And anything with that sort of like a learning rule is adaptable by the system.
And, you know, the paper, I think we should be able to like finish it hopefully in just a few weeks.
We're just like doing the language edits now and just like, you know, beautifying it a little bit.
[01:24:05] Speaker C: Well, you said, I mean, so. But you said that you're not claiming that this is the way that brains do it. But the back propagation has been a sort of a thorny issue in neuroscience because.
[01:24:16] Speaker A: Exactly,
[01:24:17] Speaker C: it is what it has. What is what enabled neural networks to function so well, even though famously, it takes tons and tons of examples because the weights have to change so slowly. And so people started looking for back propagation in brains. So surely brains must do it because these artificial networks do it. Right.
[01:24:35] Speaker A: So this has been kind of a difficult challenge. Exactly. People always jump into the transplant ideas from different fields without really trying to think about the limitations that exist in that sense system. And I think that's a big issue whether, you know, you come from physics and you don't look at the real biology, you come from computer science, you don't look at real biology, you could end up with that sort of like picture. And of course there's that idea of some are saying, well you know what, I can see that the connection of some neurons and you can run it does something similar to what we do, you know, in terms of the alpha function. Right, so that level of
[01:25:10] Speaker C: abstraction.
Yeah,
[01:25:11] Speaker A: exactly. And then they take that as granted that therefore the biology is doing the same thing. This is this sort of like mistaking the, the sort of equivalent system for what that system is. Yeah. So, and that equivalence, of course, as I said, is, is very simplified and you know, it's a wrong interpretation but you know, yes, truly the LLMs are powerful and amazing.
They're very cool. Like they can do a lot of things with them, but it doesn't mean that that's what the brain is.
[01:25:45] Speaker C: So earlier you said, when I was asking like, why shouldn't we just do it in a bigger feed forward network and you said there were two reasons. One is the efficiency.
And then I think I lost a thread on the second reason, but you were talking about the experience. The
[01:25:59] Speaker A: second is the fundamental limitation.
[01:26:01] Speaker C: Fundamental limitation.
[01:26:03] Speaker A: Fundamentally there's some physical systems that you cannot understand them with feedforward and you know, I'll be using that mathematical paper on this.
But also what we just talked about, the other aspect of it is also that sort of biological systems building a world model that they can act on it and the true agents. Right. Like, you know, as we talked about it, like, you know, right. You don't need to have, have a mathematical description of how a rock falls or whatever to be able to understand that it's falling. Like, you know, my dog doesn't do that. Like he never went to, I don't know, school of Galileo or Newton or whatever. Right. But he can understand those things. When an LLM can give you lessons with all sorts of things about those aspects, but it has zero understanding of what actually real fault of an object is or rain is or so on and so forth, that for that to happen you have systems that need to build in recurrence in them to be able to do all the sort of things that we talked about
[01:27:09] Speaker C: and have the computational closure. My worry is that. So you and I were laughing earlier about the idea of calling these things Agents. Right. And that has like, repercussions societally. Like, well, it must be conscious because it's an agent, you know, But I don't know that I want, want an AI system with what, what you and I agree on is real, quote unquote, agency. Right. Do you think it's a good idea to build that into the system?
[01:27:35] Speaker A: Okay, so that's, that's, you know, that's like a Blade Runner, right?
[01:27:38] Speaker C: Yeah, at
[01:27:39] Speaker A: some point. And I think it will happen. Like it, I assume at some point it will happen. It may not happen in our lifetime, but it may happen at some point that, you know, you just need to build a system that can change itself, you know, and it's connected to the outside world. Right now you can, you can think of, okay, we can enable these elements by attaching some sort of sensors to it. Then it will have some information about the environment. But again, that's the, that's again the raw thinking, not the deep thinking and rigorous thinking. Like what is it happening inside the system? Is it changing anything inside the system or not? Right.
So there's two different things. But could it happen at some point? Point.
I mean, it could happen.
I leave that to science fiction and I'm just not sure I want it. Future stories.
[01:28:25] Speaker B: Yeah.
[01:28:26] Speaker A: Do I have a fundamental opposition to it? I don't actually.
I like Blade Runner. You know, I could be on. Yeah,
[01:28:33] Speaker C: sure, yeah, yeah, it'll make. Humanity's ending, will make for a good sci fi story.
[01:28:38] Speaker A: No, I don't like humanity to end, but it could actually happen.
[01:28:42] Speaker C: Yeah, that's, that's true.
[01:28:44] Speaker A: But much simpler things like, you know, elements like disrupting the whole Internet and banking system and you know. Right, right. Like idiotic things. Not, not because they're very intelligent, but you know, some bad actor using them to do that.
[01:28:57] Speaker C: Right, yeah.
So sorry, I'm kind of jumping around a little bit now, but is one of the things that you're interested in, so you're talking about like when you compose a system, multiple different companies, computational modules like divisive normalization and winner all and the interesting things happen when you compose them and you're talking about this sort of explosion potentially of the capacity of the system, is that one of the fundamental things that you're interested in is figuring out the capacity of what's possible.
[01:29:36] Speaker A: That, so that when. So there are two aspects of it. One is on one side I want to understand what the brain does.
On the other side I want to be able to translate that to build something novel and in Doing so, both of them converge on these aspects. For example, understanding the mathematical aspects of recurrent systems, understanding what composition is in these physical systems, and how can we then reinterpret it into subphysical devices?
Again, because if you understand it, then you can also transplant it on the other side, right?
And composition is a huge aspect of it, because composition is what enables us to go from like lower level abstraction to bigger and much richer abstractions. But it also can give us the aspect of being able to build systems that they are very much tuned to what they need to solve. That's also another aspect, right? Like, you know, we talk about this sort of LLMs and the scaling and so on and so forth, but also think about it that you can build machines that are super good at, I don't know, predicting something very specific, right?
And you can utilize this sort of ideas and then build a system not just like, you know, a general intelligence system, system needs that, but also even like an expert system, you can build an expert efficient system when you know how to program it. And for that you need to know what the environment of interest is. Because, you know, the other paper that we talked about through emails and so on, so forth was like this idea like, you know, that the environment itself shapes that machinery when you interact with it. And you want to predict about predictions about that environment, right? It's not just like, okay, let's have a classification system now you're in a changing environment and that environment imposes constraints of your internal machinery when you're building this sort of like internal world model of it.
That is actually the point of this having a definition of what the composition is and what it enables us to be.
And especially in terms of neuroscience, it becomes hugely important because, because every day we know more and more about cell types, right? About wiring diagrams and so too
[01:32:05] Speaker C: much
[01:32:05] Speaker A: this is too many
[01:32:06] Speaker C: cell types.
[01:32:07] Speaker A: What do you say?
[01:32:09] Speaker C: We have, we know too many cell types now there's too many cell types.
[01:32:12] Speaker A: That's right. And you know, that's actually one of my problems with this is that I, I love all the experiments and everything that is coming and they're amazing. But you look at these papers, what you learn at the end, you just read, it's just like a catalog, right?
[01:32:25] Speaker C: Well, stamp, that's, that's what biology was criticized for, is stamp collecting. And in that sense we're still doing that. But you do need to collect some stamps.
[01:32:32] Speaker A: That's right. And that's the key point. Like, and I want to emphasize this, and it is super important that we do measurements because without the measurements, we, you know, we can't say much. Right. So we do, we do need more measurements and more measurements and so on and so forth. That's important.
But to have an understanding of these measures, we need to go beyond just like, hey, this is a catalog of things, right? Because from the catalog you can't really go further. Now. You need to know, I have to have the language of, okay, now what happens when I combine these pieces together in this sort of order?
And that's where basically the sort of, like, work I'm doing is in this direction. Right?
[01:33:16] Speaker C: Yeah. So here's a little quote from your, I think from a blog post. So you put out companion blog posts to the paper, to many of the papers that you write, and then you're still creating these companion videos as well that summarize it. So the closure criterion will not tell you what the brain computes. It will tell you what you have to demonstrate before you're entitled to say that it computes anything at all. So I like that this is going to, you know, related to what you were just talking about, about how to go from the stamps or from the, the catalog to what the catalog could potentially do.
[01:33:56] Speaker A: Yeah, yeah, exactly. And, you know, one of the reasons I was making these videos and I wrote these blogs was I want to have different levels of description for people because not everybody's going to sit down and, and read necessarily like, you know, the multiple equations in a paper or whatever, or just whatever. Right. So people have different interests.
But I want to carry the message forward and then make it more digestible in some ways.
[01:34:23] Speaker C: So in a second, I'm going to ask you, how much of brain processing is computation before I ask you that, because that's an explicit question that you ask on your website. That's what you want to figure out. Part of what you want to figure out. Do you think of all life processes as mostly computation or is it, is it specific? You know, it's not specific to the brain. We were talking about the computational closure criterion in single cells earlier, so. But is. Is all life computation?
[01:34:58] Speaker A: So I, I think so. I think all life has that aspect of competition, closure.
[01:35:07] Speaker C: Okay.
[01:35:08] Speaker A: I really do. And I think that's where basically that's the seed of agency, intelligence, consciousness, you know, these sort of other loaded labels that we use. That's the true seed of it, because that's where basically you go from matter to dynamics of different scales. You have effective states that have abstraction of it that enable that system to do prediction and Control of action, controls on the environment and on themselves. And that requires this sort of like recurrence, inclusion, computation and composition that we talked about. Right. So that's the sort of like view of life that I have.
All fits into this sort of like theory.
[01:35:54] Speaker C: So how much of the brain processing is computational?
[01:35:59] Speaker A: So I don't know.
I don't know. That's the reality.
[01:36:08] Speaker C: Yeah, yeah. But are you comfortable with like non computational, like things that wouldn't fit your definition of computation? Are you comfortable with those kinds of dynamics and processes contributing to the mind or cognition or behavior?
[01:36:24] Speaker A: To me, good explanations are, explanations are hard to vary. So if you have a David Deutsch
[01:36:29] Speaker C: kind of idea as well.
[01:36:30] Speaker A: Absolutely, absolutely. I think he's saying something fantastic there. Right. Like, you know, if you have a theory that, you know, you see a new phenomenon, then it's like, oh, well, you know, I need to like tweak this heuristics partly just patch something to it that explains. That's a bad theory.
Right. So if you follow that theory theory and you tease apart the system and look at different aspects of it, like, you know, for example, that's what I'm doing with like one of the things I'm doing is looking at the bacteria. For example, if you see how the chem taxes work and if it tells me, hey, these parts of the earth, just simple information relay, that's the way it is. That part of it is just information relay. So I'm ready to put my neck out and say like, you know what this part of bacteria is? Nothing. Nothing like computation. But as a whole, the system is doing this.
Right. So that's, that's totally okay. And I think the same, same is going to like scale up to brain. Of course he is much harder because now you have a system that is scaly. It's just like, you know, astronomical. Right? Yeah. So yeah, there's plenty of work with people to do.
I see people like disappointed and where it's like, oh, what do I do as a theoretical scientists, computational neuroscientists. Don't worry about it. Because
[01:37:49] Speaker C: it's so.
[01:37:49] Speaker A: Because it's so
[01:37:50] Speaker C: enormous. Because it's so astronomical. No,
[01:37:52] Speaker A: no, because I feel like, oh, there's AI now solving mathematics. Oh, come on. Just like, yeah, it's great. Like use it.
[01:38:01] Speaker C: Yeah. Okay, Nima, this has been fun. I'm going to ask you one more question here. Yes. And that is if I, if there's something that I didn't ask you that you're particularly excited about that we didn't discuss here. I mean, I know you have your fingers in a lot of things, but do you want to close on something that we haven't discussed?
[01:38:20] Speaker A: Well, actually we didn't discuss like anything about like the ushering optimizations or.
[01:38:27] Speaker C: Oh, the recurrent neural network.
[01:38:28] Speaker A: Exactly. Or harnessing like you know, kind of functional connectomics and so on and so forth to build, you know, machines that are very cool and they can happen. Yeah, yeah.
[01:38:39] Speaker C: Do you
[01:38:39] Speaker A: want to talk about those things?
But yeah, I think that just like is a whole different conversation. But so that aside, I would like to say one thing that actually is more about some of the things I'm doing now, which is basically building this sort of systems. Again this idea of bringing reservoir computing and recurrence and so on and so forth, but to do fault detections. And that's the idea is that you have observations of the system and in observation of the system could be limited to some aspects of it or to some, you know, you have partial observation of a system from which you can have a recording. And that, that's like driven by that idea. Like you have measurements. Right. And can you build physics informed neural network that can give you a sense of where something. Let's say that that system has changed. So if you observe it long enough, can you build some sort of of like description of that system that you recognize when it's changing? Right. And that has fewer applications. Application. You know, this was actually driven initially by some application oriented question and that's actually what the funding came from.
So we're working on that one. So that's a really like cool thing that like adds to this basically like using that sort of like recurrence machinery to look at the system as it changes. And you want to know where a shift is happening, even if when you have partial observation. And I think similar set of tools will be very useful for neuroscience because that's our fundamental limitation. Like yes, we record from now from 100 neurons instead of one neuron, but still that's like massive subsampling. Right.
Let's say we go orders of magnitude high, higher, it's still the same thing. And then again we need to be able to figure out like when you manipulate the system, what sort of change do you see? Because at the end it's a system that we need to be able to, you know, poke it right like that. That's the real way of like recognizing the causal. We can have a causal description of the dynamics of the system because that's when you have, we can bring in causality we can bring in and counterfactuals. Right. And much of this is missing in this sort of like a very complicated sort of like information processing systems that we want to study. Because, yes, you know, there's optogenetics and you can poke something, but you know, if you don't embed it in this sort of like, description and this sort of like a bigger system description, then it's very hard even to come up with some true description of what, what that poking did
[01:41:30] Speaker C: to interpret.
[01:41:32] Speaker A: And of course, one thing I want to say also a shout to open neuroscience, because that's what's going to enable us to go forward.
And we really need to move in a way that people share their methods, their tools, their computational methods, experiments, of course, after they publish their paper, to get the credit. So
[01:41:53] Speaker C: that's the whole thing is like, I think that's less of a concern now because there seems to be a lot of fruit for everyone. But I mean, I remember giving a talk as a graduate student. I'm sorry, I'm repeating this for my listeners because I've told this before. And I went to Xiaojing Wang Wang's lab and his students were asked, because I was an experimental neurophysiologist and we were talking about open data and stuff, and his students were saying, well, why don't you guys just share your data? And I'm like, you have no idea how hard that
[01:42:21] Speaker A: data is to get. And I might want
[01:42:23] Speaker C: to model it.
[01:42:26] Speaker A: Whatever you do with it, you need to get your credit for collecting that data. You're doing this for a specific reason and you're doing an experiment. And of course, you know, it goes both ways. Right. Like you can also, like, build models and if you don't release your models, no one knows if it was true or not. Yeah. So, but I think that's something that, as you said, the barriers are getting lower and lower. But I think that this is quite useful because some of the work that I did, for example, for, you know, we didn't get to talk about it, for example, harnessing connectomics data. That, that really comes from the fact that it was collected and released. And, you know, I don't know how to go and read electron microscopy. No, I don't want to spend my time like building an algorithm that goes like, figures out, okay, this is the connection from this slice to data, slice to data slice. Like, you know, there's someone else who's doing that. Well, I don't want to do that. Someone else is doing that. That's fantastic. Right?
And we have, you know, this came up once in a meeting that I was invited to by Jerome de Kirk, who is running the Brain Observatory platform in Allen Institute. So there was a session with him and Carlos Fobude and Krzyszokov and a few others. And then they invited a few of us to go and talk about it. And it was partly because I was also associated with Dandy and doing some stuff around that.
And I think that's a very interesting platform. So maybe.
And you know, one of the things that we were talking about was like, maybe we need a few of those instead of like one of those, like back to physics, the tons of observatories. Right,
[01:44:09] Speaker B: right.
[01:44:10] Speaker A: And there you have.
[01:44:11] Speaker B: That'll be healthy,
[01:44:11] Speaker A: dedicated, like rigorous experimentalists. They know what they're doing, right? And, you know, it's not that they're like some technician. No, they're actual like, you know, proper scientists doing these experiments because that's what's needed, because it needs that expertise. But then the data opens up and many people can pour in and test things and so on and so forth. So maybe in some future, hopefully not far from now, we can have multiple of those centers, right, that people can collaborate, they can propose experiments, perhaps even a group of people together, experimentals and theories that come up with something and propose it there. And then the sort of like dedicated experts in that system produce that and, you know, it goes back to the public. Everybody can pour in and explore it as much as they want to. So, so. And then
[01:45:05] Speaker C: we'll all be holding hands and singing that we solve the brain and no
[01:45:09] Speaker A: competition will always exist. I think that's the key about. Humans will always compete with each other. So, no, I'm not worried about that. We always compete. So. But collaboration also helps us to go in places where we don't go.
[01:45:22] Speaker C: And collaboration can be competitive and that can be a source of productivity. Actually.
[01:45:26] Speaker A: Absolutely. In fact, actually, it was like lately I was saying that someone was saying about, like the OpenAI solving math. I said, like, you know what? I'm not worried about humans. They will create like, maybe we should create something like M1, like instead of Formula One, M1 teams, you know, like crazy mathematicians that, you know, they're now outbid by AI system. They can work with AI and as a group and team. And then they compete and then they try to.
We beat you in solving other than Riemann's hypothesis.
It will happen.
[01:45:58] Speaker C: Okay, Nima, this is. We, we've come, we've covered a lot of ground. We've come a long way. We have, as you alluded to, lots more that we didn't talk about. So I look forward to doing it again down the road. But thanks for spending time with me for this.
[01:46:09] Speaker A: Thank you very much for having me. And thanks to you. By the way, I really love your show and I have been following you for a long time. Sometimes I listen to some episodes again and not I haven't listened to all of them, but there's some episodes that I listen to. I think you've done a service to the company. Of course it comes from, you know, you being obviously a very curious person and thoughtful person, you know, thinking in directions that most people don't necessarily do. You know, it has a cost and you know, it's very demanding. So thank you for doing this. And you know, I'm also very glad that you have a partnership with Transmitter. Transmitter is fantastic. So it's really good. Yep.
[01:46:58] Speaker D: Brain Inspired is powered by the Transmitter, an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives written by journalists and scientists. If you've ever you Brain Inspired, support it through Patreon. To access full length episodes, join our Discord community and even influence who I invite to the podcast. Go to Brain Inspired co to learn more. The music you hear is a little slow jazzy blues performed by my friend Kyle Donovan. Thank you for your support. See you next time.
Sam.