| 0:37 | Intro. [Recording date: October 21st, 2025.] Russ Roberts: Today is October 21st, 2025, and my guest is neuroscientist and author Gaurav Suri. He's the co-author with Jay McClelland of the book, The Emergent Mind: How Intelligence Arises in People and Machines, which is our topic for today. Gaurav, welcome to EconTalk. Gaurav Suri: Thank you, Russ. Pleasure to be here. |
| 0:56 | Russ Roberts: So, this is a rather extraordinary book. It's got some very powerful explanations of the nuts and bolts of neural networks and large language models, and the different aspects of those models that have come to be part of our lives now in the form of various ChatGPT [Generative Pre-trained Transformer], and Claude, and others, but it also has some very deep ideas about the human experience and how we should think about it. So, I really encourage people to read this book. We've done a lot of episodes on AI [artificial intelligence] and we've done not a small number on neuroscience, and this book really integrates them in such a fascinating way. So, my hat is off to you and your co-author, Jay McClelland. Gaurav Suri: Wonderful, thank you. Russ Roberts: So, let's start off with the title of the book, The Emergent Mind. Of course, here on the program we talk about emergence now and then, and its various aspects related to economics: how prices emerge from competition, say, between buyers and sellers. They're not caused by any one person; they're not designed or directed. How does that concept apply to the mind? Gaurav Suri: Right. The experience of living is that we do things. We wake up, we go to the mall, we choose a taco instead of a burrito, we perceive things, we understand things. The question is: how does this happen? And a really inviting answer is that it happens because we have intelligence built in to the components of our mind: that there are neurons that know how to understand, that there are neurons that know how to speak a language or perceive. When you actually look at neurons, they don't appear to be doing any of these things. All they do is they activate and maybe they connect with other neurons. So, the idea of the title, and really of the book, is that intelligence, our intelligence, but also intelligence of our machines, emerges from the interaction of simple processing units that themselves are not intelligent in the way we understand intelligence, that they themselves are doing something very simple, and yet the whole has properties that none of these neurons have. So, Russ, this is the definition of emergence: Emergence is a property that's present in the whole system, but not in the component parts in the same way. And, the thesis of the book is that intelligence in our brains and in our machines emerges in that it's not present in the units--the neurons or the neuron-like computing entities--but is present in the whole. And, emergence, as you know, is not a rare thing in the universe, that it applies to brains. Emergence is everywhere, from galaxies to water, to ants. It exists everywhere, and the book makes the case that that is how we should approach our pursuit of understanding our intelligence. |
| 4:31 | Russ Roberts: So, let's go a little deeper into that for a minute, because in nature you use the example of flocks of birds; you can use the example of schools of fish. We know that there's no king or czar, and even when there's a queen bee or a queen ant, they're not giving orders. And somehow--and this is a profound, I think, insight and challenge of economics--the system acts as if the person is giving orders. And that's very deceptive. It causes people to misunderstand what's going on. And, I think one of the challenges of science, and one of the challenges of economics--which it's not done a great job at, but we try--is to understand: Well, if it's not that, how could it possibly work the way it looks? And, you have a beautiful example--and I thought we'd digress on this, but not too long because we could spend a long time on it--on how two ants, when they come to a barrier, a little stick that's in the way, and the stick's not centered on the path. The stick is short and one way around is long. And, eventually when you have a group of ants, they all seem to manage to find the short way. But no one gives them a map. They don't have Waze, we're pretty confident. How do they do that--using pheromones, using chemical secretions? Because you need that micro-mechanism to understand the macro properties of the whole system. Gaurav Suri: I think this is a really interesting example and a seminal example, and I am delighted to lead our conversation with this. Because I, after this example, Russ, began to think of our brains as a colony of ants. That's a metaphor, but I think a good metaphor for what's going on. So, I grew up in India and I grew up in an era where there was no social media and bad TV. And so you were outside. And one of the things that I did when I was outside was look around in the world, and one of the things I saw were ants. This example that you were talking about consists of ants going from their nest to a source of food and coming back. And, the line of ants can become pretty dense. So, they're going from their nest, they're getting the food, and they're coming back maybe with a particle of food. And, as a young person, what I did was I placed an obstacle on that train. So, if you imagine a band of ants, you put an obstacle there and you see what the ants do. Well, you do something interesting, which is you make the obstacle have a short way and a long way. And initially--it's quite dramatic to actually see this happen--initially, the ants go 50-50, right? So, 50 going the long way, 50 going the short way. And in a few minutes something amazing happens. Nearly all the ants, not all, but nearly all the ants are going the short way. And, as you nicely ask, 'Well, what's going on?' And, when my son was young, I would ask him this question and he'd say, 'Well, maybe the queen ant is communicating to the other ants,' and well, first of all, they're not communicating in that way, and second of all, they don't have a conception of distance. Right? Ants by themselves do not distinguish themselves by their intelligence. You put an ant on a flat surface and the ant is pretty much going to walk around until it dies of exhaustion. And yet, together they're capable of extreme intelligence, like building cities and finding the short way around obstacles. So the question is: What are they doing? Well, they're doing something very simple. Just like we talked about neurons doing something very simple, which was activate and connect, ants do something very simple, which is secrete pheromone and follow the pheromone trail. That's what they do. Now, it turns out that if there's a long way and a short way and the ants are 50-50 and one ant goes the long way and one ant goes the short way, the ant that goes the short way is going to reach the destination first. It's going to have a choice, 'Gee, which way should I go back?' Well, it's going to go back the way it came because it laid the pheromone while coming, and so it's got more pheromone on the short path than on the long path, so it follows the short path. A little bit later, the ant from the long path comes and it has to decide, 'Which way am I going to go?' Well, it has pheromone on its path, but now there are two trails of pheromone on the short path, the first ant coming and going. So, it comes back the short way, and boom, that's all it takes. It's like there's a greater pheromone concentration on the short path, and that leads to the phenomenon. Russ Roberts: And, of course, we assume that birds and schools of fish do the same thing. They have a very simple behavioral rule. They don't have a rule that says, 'Let's go 40 degrees over here and then swoop around and we'll spin twice, and--.' They just try to keep not too far and not too close to each other, and that creates a flock. I make the analogy in my book, The Price of Everything, to the Blue Angels. The Blue Angels does an incredible thing. They have a choreographed flight pattern incredibly close to each other. But of course, they can communicate--and they have to, or they're going to have problems, or they at least have a plan in advance. The birds manage to do it without a leader, without a plan, without a flight path, etc. It's an amazing thing. |
| 10:39 | Russ Roberts: So, tell me and tell us: what's that have to do with the brain? By the way, you should just explain--let's step back a minute. Let's talk a little bit about the anatomy of the brain. There are these on-off things--electrical signaling--that's going back and forth between various neurons, jumping distances sometimes, often reaching out to multiple neurons from their source, from the original neuron. Give us a little bit of a picture of what's happening in there, and how could that possibly lead to intelligence. It's just like an electrical storm, that's all it is. Just a lot of electricity jumping around--a lot of it because there's a lot of neurons. Gaurav Suri: Right. There's a hundred billion neurons, give or take. The latest number is 86--I think that's subject to revision--86 billion neurons. But, let's-- Russ Roberts: That's in your head. I only have 84 billion. You clearly have more than I do, but-- Gaurav Suri: I am not at all sure about that, and I'm also not sure that counting neurons is a good way to count intelligence. But, Russ, what you're saying is fascinating because our brain has electricity in it. This is not like a line from a Frankenstein sort of story. Our brain is composed of these neurons. There's other things in the brain, but the principle information processing engine of the brain are these neurons. And, a neuron--think of a neuron like it's sort of like a tree. So, up top there are these dendritic branches. So, imagine the massive branching happening, and then there's a long trunk with some shallow roots. The long trunk is called the axon, and, like, a tree ends in the ground, but a neuron--imagine the axon of a neuron, the tail of a neuron opening into the branching--the dendrites--of another neuron. So, there are these structures that are in close proximity to each other. And neurons, at a high level of abstraction, do two things. They generate bursts of electricity called action potentials. These are actual bursts of electricity. So, you can hook them up to a device that transforms electrical signals to audio, and you can hear bing, bing, bing, pshh. It sounds like that. And, that's the sound of a neuron having electricity. It's quite marvelous. So, that's the one thing they do; and the other thing they do is they connect with each other. They may be connected from birth. So, we are born with pathways of neurons that lead from our retina to our visual cortex in the brain, and in fact, that is how it was first discovered--this is a slight tangent, just 30 seconds. It was first discovered that it is the brain that is the seat of intelligence because a gladiator or someone died and they followed his optic nerve. And, some people thought--I think some Greek philosophers thought--that it is the heart that is the seat of cognition. But, you follow the optic nerve to the train--the axons that are strung together--that make a pathway, the gray matter track, and that goes to the visual cortex. That's in the back of the brain. So, these connections are innate, meaning some of them exist from birth; but they're also modified by experience. So, when we learn something new--like, I learned something when you were talking about the Blue Angels just now, and that was a nice analogy for me because oh yeah, murmuration and communication and emergence, and here the emergence is maybe the birds looking at each other, and there the emergence is maybe the pilots to each other, right? But it is an emerging system. So, I learned that. And, when we learn some things, we represent knowledge in our brain by making new connections. Neurons that weren't connected before are getting connected. And that's the only two things that neurons are doing. And from these two things, the emergence, the intelligence, is emerging. Russ Roberts: Of course, as you're talking, over a hundred thousand people are having an electrical storm in their brain while they're listening because they've never maybe heard this before. Or they've heard it, but they haven't heard it quite this way. And, those who might be watching, your visual use of your hands to show what a neuron might look like was about two feet tall. They're smaller than that, of course. 'Oh, yeah'--and then, another thing fires somewhere else as I embed in my brain this new knowledge. Of course, I don't embed: it gets embedded. I don't have any control over it. I can try. I can say, 'Oh, I have to remember this.' Maybe that helps, maybe it doesn't. Maybe we'll talk about that later. |
| 16:01 | Russ Roberts: But, a big theme of this book is--and it's deeply disturbing by the way--I said this is a really extraordinary book. It's a deeply disturbing book, because it's the picture you paint of our inner life, which is our consciousness and our experience of our mind, and how we make decisions and act in the world, you're arguing, is not so much in our control. You pick on economists, correctly, perhaps; we could talk later about that if we have time. But, economists say, 'You look at the world, you have a goal, utility maximization, certain things give you pleasure, certain pain, you have alimited amount of money, and you make these, quote, so-called "rational decisions" trading off pleasure and pain--pleasure from multiple sources--which one is the biggest bang for the buck?' That's the caricature of an economist's way of looking at decision-making. You have a very different view, and in that view, I am somewhat at the mercy of my past experiences that are bouncing around in my brain and causing me to do things that, I might explain them with some words, but they're not really in my control. So, try to give us a flavor of that perspective on how we act. Gaurav Suri: Yeah, that's a lovely, beautifully posed question. Thank you for it. Let's take a simple example. This experiment that I'm going to describe was done in the 1970s by two, I think, brilliant psychologists, Nisbett and Wilson. What they did was they presented four stockings--let's label them A, B, C, D--to people. And, they asked them to choose whichever stocking they like, and people chose the fourth stocking in general, and when asked, 'Why did you pick this stocking?' some people said, 'I like the thread count. The hue of the color is slightly different.' They all had their reasons. Meantime, all the stockings were identical, and it's a well-known fact that when you give people a list of identical things, they'll most likely pick the last thing. That's just a choice pattern that people have. And so, nobody said, 'I picked it because this was the last thing in the order, and these were identical.' They had a reason. And, there are some very beautiful experiments that--so there are two hemispheres in our brain, the left and the right, and they're connected by a bundle of nerves called the corpus callosum, and for some purposes, a while back, patients, they would cut off the corpus callosum to prevent fits and whatnot. So, now you have two hemispheres that are not talking with each other, and it's possible to show the visual field of the right hemisphere, which, for many people, does not dominate in language--show them a message like: 'Start crawling on the floor.' And they can't process that they're crawling on the floor because you showed a message, because the corpus callosum is cut off and that message hasn't traveled to the language center, that hasn't made awareness, and whatnot. And, they come up with a reason. So, they will say things like, 'I think I dropped my keys,' or 'go to the bathroom.' 'Why are you going to the bathroom? You just went.' 'I think I need to wash my hands.' So, the thesis of the book is that this system of associative neurons that are interacting with each other, rather like ants interact with each other, are also producing our justification and stories for why we do the things that we do. It's the same system that's producing the choice is also producing the reason for the choice. And sometimes the reason has a lot to do with the choice. 'Johnny, why don't you go and play in your friend's house?' 'Oh, last time I was there, he beat me up.' That reason probably has a lot to do with the underlying neural net, also known as the brain. But, the astounding thing is that there are different systems, and they may not--and we humans are master storytellers, and we often tell stories that put us in a good light and make us virtuous and heroic. And, these are accounts; these are stories produced by the underlying network. So, let's take a very simple example. Let's say you go to the refrigerator for snack, and somebody says, 'Russ, why'd you go to the refrigerator?' 'Oh,' you might say, 'I'm hungry.' Or, 'I'm thirsty.' And, true, that's true. What's happening underneath? Well, you have thirst-detecting neurons in your brain. The way these neurons work is they detect the salinity level in your blood, and if the salt level is too high, that means you don't have enough water. And they start to activate, right? Bing, bing, bing, bing. Or hunger neurons might start to activate. And they're connected with experience. They're connected to another network that initiates signals to the muscles to move in the direction of the refrigerator. So, these neurons that are interacting are you go to the refrigerator. They don't necessarily have to do anything with the linguistic conscious feeling 'I'm hungry' or 'I'm thirsty and I'm going to the refrigerator.' You might have that feeling, or you may go to the refrigerator out of habit because you usually go at 6:00 when you come back from work. The point is that these systems do their thing. They make us move, they make us do things in the world, they make us choose--and we'll talk about economists and pleasure and pain in a minute. But, they are busy interacting with each other, kind of like ants, and they are producing our thought and our actions. Now, is that disturbing? Well, I'd say no. And, I want to mark a reason for hoping we can come back to that later in the conversation. |
| 22:53 | Russ Roberts: Well, at the end, we'll talk about the end of the book, which I think is deeply uplifting and profound. But here's the depressing part. So, let me take your refrigerator example. So, I'm reading your book this morning at 10:45, and I'm ashamed to say I'm reading it on my phone. 'Ashamed'--why am I ashamed? I'm not. I happen to have a romantic association with physical books. But, it's an interesting question. Why did I have to say that I'm ashamed? Well, that's part of what we're talking about here. There's something--in my own set of neurons, I feel a need, not rational, to say something nice about physical books. Why? I don't know. I could tell you a story; I just tried to. But, at any rate: so I'm reading the book on my phone and it's 10:45, and I'm going to eat lunch around 11:45 because that's about when I usually eat lunch. But, at 10:45, I got up from the couch where I was reading, and I went into the kitchen, and I got something to eat. Now, if my wife had come home at that moment, she say, 'What are you doing?' I'd say, 'Well, I'm a little hungry before lunch.' A reasonable thought: but not true. And your book forces me to confront that. I got up, I got some peanuts, and I know why I did that--really, your book forced me to confront my act, which is that when I read, I like to eat at the same time. It's not because reading makes me hungry. It doesn't even make any sense. I remember my dad, when I was a young boy saying it's fun to eat and read. He would eat, say, a bowl of popcorn while he was reading a book. So, I watched that, and I do the same thing. Fortunately, there wasn't a bowl of popcorn. If there had been, I probably would have consumed the entire bowl, regardless of the size. And, as I'm walking to the kitchen to get the peanuts or whatever I'm going to rummage and find--and this is where I think the ideas of the book are so helpful, but also disturbing--by the way, some things that are true are disturbing: we decide whether we want to consume them or not, or try to think about it. But, anyways, I head there; I'm saying to myself, 'It's almost 11:00, I'm going to be eating lunch in less than an hour, I really shouldn't have a handful of peanuts.' But I did. And I went and got them as I was thinking that, I didn't--so what happened there? Gaurav Suri: Well, that's a great example. I love that example. I think this--it's not quite a habit. So, a habit is something that you almost--it's more automatic. I call it action readiness. So, the way I describe it, Russ, is that you have a greater action readiness to snack when you're eating-- Russ Roberts: When I'm reading. Gaurav Suri: When you're reading, sorry. When do we get action readiness? Well, watching somebody would do it. A particular context would do it. Lots of people eat popcorn while watching a movie. Even if they don't want popcorn, even if the popcorn is stale, they will eat it, right? Now, interestingly, if you offer people stale popcorn in a conference room, they won't eat it, but in a movie house, they'll eat it. Russ Roberts: Oh, yeah. Gaurav Suri: So, here's the amazing thing--this struck me when I was doing my Ph.D. So, a little bit of a biography, because I think this is important context. I worked in industry for many years, I was a management consultant, I was a partner at Deloitte Consulting, and I saw people deciding in boardrooms a lot, right? And I would--and these are big decisions: What is the price of our inkjet printer in South Korea? And I'm thinking, was this really decided based on the data that I'm presenting, or was it because this person showed up to the meeting and that person didn't, and there's light streaming in, and so-and-so wants to go to the bathroom? There are this conspiracy of interacting things: Just like neurons are interacting, people are interacting. And I was really struck by people have patterns, people have greater action readiness to do certain things. So, I come, in midlife--consulting was kind to me--and in midlife, I had the luxury of coming back to academics, and I did my Ph.D. And one of the first experiments I did--this is crazy, I still can't believe this experiment--if you show people two pictures, one of a beautiful scene in nature, and one of, say, a horrific mutilation, and ask them to pick one picture, 90%-plus of them will pick the beautiful scene in nature. Not a surprise. A little bit of a surprise that there's still about 10% that want to see the other picture, but 90% is close enough, as close as you get to 100% in human behavior. But, if you change the experiment slightly, what you do is you show them the horrific picture and teach them that if they press the S key--S for switch--they can switch to the beautiful nature picture. It's the same choice, essentially, right? It's just that now, a proactive action is needed. It isn't framed in the context of choice. Now, in the case of proactive action, only 50% switch from the bad image to the good image. But here's the amazing thing. If you, instead of S, make it a forward slash key--which people don't usually press--now, instead of 50%, only 30% switch. And here's the--this is the experiment I did, which is if you pretend them[?] to copy the HTTP line of code, which has a bunch of forward slashes--so they're copying code. They don't know why they're copying code, but they're copying code, so they're writing a bunch of forward slashes. Now, if you ask them to do it, that 30% goes back to 50%, because they pressed slash a bunch on the previous task. So, things that we do, things that we're used to doing, profoundly influence the things that we do. The things that we do are influenced by preference, yes; but they're also influenced by what we're looking at, what our attention is directed to, what our action readiness is, what we saw somebody else doing, what the incidental associations are. |
| 29:31 | Russ Roberts: But, if you're listening to this and you haven't read the book, and if I was listening to this conversation and had not read the book, I would have said, 'Yeah, yeah, yeah, people are irrational, people can be manipulated.' But, you're really saying something more profound. I want you to try to give an example and go back to the neural network in the brain for trying to understand what's going on. The models that you provide in the book emphasize--without going into the nuts and bolts of the models themselves--they emphasize frequency, and almost the fact that certain behaviors get grooved through these neural connections that get made, that multiple factors will often interact in interesting ways. I want to give you a chance to defend what you just said from a different perspective, because if I didn't know you're a neuroscientist and you told me about that experiment, I'd say, 'Okay, yeah, people don't always follow what they say they're going to do, or they don't always act in their own self-interest.' But, you're really saying something, I think, much deeper than that from the neuroscience. So, give us the flavor of that. Gaurav Suri: Yeah, that's great. So, one metaphor that my co-author, Jay McClelland, recounts, is to start thinking of neural networks as reservoirs of interconnected pools of water. Like, you know how in a stream, how water falls: there's little pools, and there's a channel from one pool to the other pool. So, think of these little micro pools as units, which are populations of neurons, and think of the channels between them as the connections between these neurons. And, the more water that goes from one pool to the other pool, the deeper the channel. Right? And, the thought here is that the more a neuron is co-occurring with another neuron--so if I say green and you say grass, what happened? Well, what happened is that neurons corresponding to the word green have a deep channel or a connection with neurons connected with saying the word grass or thinking of grass. And, what made this channel? It's the frequency of exposure. So, this was Canadian neuroscientist Donald Hebb, came up with this idea that if two things are co-active, they become connected. So, neurons that fire together wire together. It's a simplification of Hebb's rule. And it's a beautiful rule. I'll repeat it: Neurons that fire together--meaning they are co-activated together--wire together. So, imagine your brain: We've been imagining your brain as an ant colony, and now imagine it as the system of pools of water in a stream, maybe in a waterfall, and the pools fill up. Some pools fill up, depending on how the water is falling. Water comes and goes, but the traces of the traffic between the pools, that stays; and that's the knowledge of the system. Right? So, this is a startling idea, but it's useful to think of your thoughts as patterns of electricity in your brain, patterns of activation. Thoughts come and go, just like water comes and goes in a stream, but your knowledge are the channels, are the connections, between neurons. And, this is why people respond differently to the same input, because they have different connections. They have different channels that takes the input that they're seeing or smelling or tasting and sends it to different parts of the brain based on, in part, their experience. And, if you've had different experiences than mine, then your channels are going to be different, your pathways in your brain are going to be different, and you'll respond differently to the same input than I will. Now, what makes these channels? Well, the point here, the central idea, is that channels can get made just by repetition. Eating while you're reading, do it a few times, and you've got that action readiness, you've got that channel. |
| 34:36 | Russ Roberts: But I don't want to overstate the implications of what you're saying for habit, which is important. So, part of what you're saying is that we have good habits as well, and we have bad habits, that we have things we've become accustomed to that might make it more likely for me to do something than for you. So, the choice I might make in that reading setting is different. You might not eat while you're reading, and I will, because I've done it many times. My dad ate popcorn while he read, and yours didn't. Or whatever it is. But, it really goes, for better or for worse, it goes way beyond that, to suggesting that-- Let me give a rapid movie version of the human experience; and you do this in the book. You're an infant, you're out of the womb, things are blurry, some colors, stimulus all over the place, and you begin to accumulate these channels by the stimulus that you're getting and the firing that goes on your brain. I don't know if we'll get to it, but of course, some of it's already set up. We don't know how much, but some of the channels are already set, because we have some innate behavior that we call instinctual or hard-wiring--and I'm fascinated by that; maybe we'll talk about it. But, most of what we are at our current age as adults, it's just the accumulation of all the firing that happened in the past, which produces a likelihood of certain behaviors that are not modeled effectively, you're arguing, by, say, an economist's worldview or a psychologist's worldview. It's simply the: I'm at the mercy of the chemistry and electricity of my brain. There's a certain anti-free-will aspect of this that you don't talk about in the book. You don't write--I don't think, the phrase 'free will'--I didn't notice it. But, when you're reading it, you're realizing, 'Oh my gosh, so many of the things that I think I'm controlling, really, I'm not.' I'm just moving through the world like an animal moves, because I am an animal. Gaurav Suri: Right. So, let's talk about free will, because it's come up, and it's centrally related to what we're talking about here. The picture that we're painting about these channels shaped by experience is a picture of a deterministic system-- Russ Roberts: It feels that way-- Gaurav Suri: And, 'deterministic,' what does that word mean? The deterministic system means that a system is--operations of a system are predictable based on its current state and its input. Right? Now, well, we learn from experience, so we're changing our brain, and so we're not going to respond the same way to the same input. Literally, if I learn to juggle, I've got so many new connections in the brain that you can see them. You can actually see white matter tracts that didn't exist before. So, the brain is plastic. It's changing. But, the thing is, it's a deterministic system, and yes, absolutely, the case we're making in The Emergent Mind is that it's a deterministic system. But there's a big 'but,' and this is why I'm really hesitant to say free will is not real. There's three 'buts,' so bear with me. The first 'but' is that free will is a very useful construct. It's a construct, because if we don't have free will and we don't hold people accountable for what they do, there will be chaos. So, even if the system itself is deterministic, the construct is very useful, and we need that construct, we need to live by that construct. That's Point Number One. Point Number Two is that for all practical purposes, even though it's a deterministic system, it is such a complex system that it's computationally pretty close to not being easily tractable. It's an extremely complex system, right? You can't tell whether something's going to land heads or tails, even though that's a deterministic system, depending on the force of your-- Russ Roberts: Laws of physics, laws of physics, there's nothing magical going on there. Gaurav Suri: There's nothing magical. The third thing is that we have goals; and goals, too, are deterministic. It's predictable where these goals come from. Goals are basically, I think it's very useful to think of them as memory structures that influence action. So they're a sub-network. And, how do goals come? We will talk about that later. But, goals are essentially sub-networks of activation that are influencing what you do at a current state. Now, goals--when people talk about free will, they really are talking about the pursuit of goals. And goals live in a deterministic system. And, I think it's really a disservice for a neuroscientist to say that, 'Yeah, we live in a deterministic system, and free will is not real, and you guys are full of it,' because free will is a very nuanced thing. It's a useful thing. It reflects on goals. It's talking about things that are complex. So, I don't think it's useful to go around saying that, 'Look, free will isn't real.' I think it's a useful construct. But, the point is that our goal pursuit is coming in a deterministic context. Now, we are a process, Russ, absolutely--you said like an animal running around or like water flowing downstream--we're a deterministic process. And, I think that we are a part of the universe, no more than the trees or the stars, but also no less. We're a process. And I find great comfort in that. My understanding of myself--I've stopped thinking of so-and-so as just fundamentally evil or so-and-so as--like, we are processes. We're processes shaped by our context, and we get to have goals based on our experience and our bodies and the demands, and we follow those goals, and we watch ourselves unfold in this life, we see ourselves unfolding, and that's pretty magical. |
| 41:35 | Russ Roberts: Let me try to restate what you said about determinism, because I think I understand, but you can correct me if I have it wrong. It's deterministic in the sense that it's subject to the laws of nature--chemistry, biology, physics. But it's so complicated that those laws cannot be understood by an outsider: you need to be God, you need to have some infinite/unlimited ability to connect all the multiple causes that lead people to do things. So, in some sense, it's irrelevant whether there's free will or not, because it's not like it's so simple and deterministic that you can figure it out the way you can figure out whether the ball is going to roll downhill or not. That's deterministic also, but it's not interesting: it's not complex. At the same time, when I got up to get the snack, because I'm reading your book, I'm thinking about it a different way than on a normal Tuesday, and I'm wondering: Oh my goodness, I am the pawn of my past. Of course, I have some freedom to say no, and I don't get up and have the snack every time--there are times I don't do that. So that's interesting. And, you write about that in the book as well, that there's motivation and discipline that's imperfect: sometimes it works, sometimes it doesn't. But, in theory, the processes that allow me to sometimes say no are no different from the other ones. It's just that that day, that context, that weather, that sunlight you're talking about--and to see yourself as that moving through the universe in that way, there is something really quite beautiful about it. It's also troubling, to some extent. Gaurav Suri: Well, yeah, I could see how one would think it's troubling if one thinks of this internal spirit that is the source of one's actions. Russ Roberts: Me. Me. Me. Gaurav Suri: Yeah. You, Russ, you are an emergent consequence of the little parts of biology that are made; and I think there is a lot of beauty in watching yourself unfold. Now, by the way, there are things that we do that seem effortful to us--so just paying attention--that feel an awful lot like an inner us is exerting control. So, there's this famous task in psychology; it's called the Stroop task, and the Stroop task is in one condition, you're supposed to read the word, say green, and you say, 'That's green, that's black, that's white,' whatever. But, in the other case, you're supposed to ignore the word, which is a color word, and instead identify the ink that the word is written in. So, if the word green is written in red, you're supposed to say red, not green. Now, it turns out--this is reliable; it's been done thousands and thousands of times--and reliably it's true that in that color-naming task, people are slower than in the word-reading task, because for all of us, word reading is much more automatic than color naming. But, what's happening when we're naming the color? This is the question. It's fundamentally an attention task. Now, what does it mean to pay attention? It means to pay attention is to increase the activation in a particular stream. And, again, this is a deterministic thing. It's guided by your goals; your goals of doing this task is guiding it. But, you have the capacity to pay attention and you have the feeling of effort. So, this is where this 'me,' this thing you're talking about, this self-is coming, because you, the self, are paying attention and pursuing your goals, and that's what's giving you this notion of achievement. And, the point is that that's valuable and beautiful and no less beautiful than it's deterministic. Does the number three exist in nature? What is the concept of three? So, when we are children, your mom says, 'That's three ducks,' or, 'That's three trees,' or, 'That's three dogs.' And, we form a concept that is the intersection of all these experiences we've had with three things, and this concept emerges of three. Three is a concept. So, this effortful goal pursuit is a concept that emerges, again deterministically, but it's a concept. And, just because it's a concept, doesn't mean it's less real. Is three less real to us? Where is three? Show me the platonic idea of three in nature. No, it's a concept; and yet, it feels as real and we honor it and we use it. Free will is a concept that emerges in this deterministic system, so let's honor it and use it. |
| 47:01 | Russ Roberts: So, let me give you an example from this morning. A different time, a little earlier. I'm laying in bed and I'm thinking what is usually called stream of consciousness, a bunch of erratic, unpredictable, seemingly random things are flashing through my mind. I'm trying to decide whether to get out of bed or not, it's before my alarm. As I lie in bed, there's all these different thoughts. There's a worry about a thing I didn't get done the day before. I'm thinking about my favorite sports team. I might be looking forward to lunch with my wife. I don't seem to control them: they're just the stream at this time of thought. And at some point, I realize, 'Oh, I'm awake, and I really should be thinking about,' let's say, 'my interview with you today.' And I start thinking about what I should ask. And, it appears that in the first set of thoughts, they're not under my control, I'm just being bombarded by them. But, in the second part, I've taken charge. I have stopped that flow. I've said, 'Enough of that: let's get purposeful, let's go to a goal'--in this case, 'of having a good interview.' What's happening in my brain in those two different types of consciousness? I know you think it's the same thing, because you're a very focused guy on the neural network. Which is plausible. But, what's different? Why'd that happen? Any ideas? Gaurav Suri: Yeah, yeah, lots of ideas. This is what I think about all the time, Russ. So, yes, I'm a neural network guy, and I'm going to think about this in neural network terms. Now, first, you're laying in bed and it's a stream of consciousness, and one thought is connecting and cascading to the other, and a few things are happening. Maybe light is coming in through the blinds and that's triggering the day is started. Maybe you're hearing the birds outside, the day has started. And, with the day started, a common action-ready thought is, what are my goals for today? Now, somebody else might not get that thought. You, Russ, from your life, are in a place where maybe you get up and think that thought, which is, what are my goals for today? And, that thought, like any other thought, is a pattern of electricity, it's a pattern of activation in your brain. Okay. Now, when that pattern happens, it's activating potential goals. Right? And, you remember that you're going to talk to me, and you finished reading the book, and you're framing your questions. That, what are my goals now stream of consciousness, like, still is coming to: what are my goals for today? And, now all of a sudden it's, like, 'Oh yeah, yeah, I need to ask that guy about Adam Smith.' I'm making this up. Russ Roberts: I was just going to say that--that terrible caricature of Adam Smith in the book. It's the only thing I didn't like, but we'll-- Gaurav Suri: No, no, the quote is not a caricature. I'll talk about that in a second. I mean, the quote is Adam Smith--I think the quote is him reflecting on how powerful body the [?]-- Russ Roberts: Oh, that quote's fine. It's a different part. But, sorry, keep going. Gaurav Suri: Okay. So, now you have this goal, and this goal is: you want to ask me a specific question. You want to be precise about that question. So, what's happening is, this goal is directing your attention to that particular question, and you've focused on that question. Still a very deterministic system. By the way, the act of giving attention has different experiential qualities, right? Russ Roberts: For sure. Gaurav Suri: And, Daniel Kahneman talked a lot about that. And, Daniel Kahneman's answer was that, 'Well, it's two different systems.' I said, 'No, no, it's the same system.' It's the same system, but different kinds of things are happening in the same system. In this system, you are influenced by the goal, which was activated not through magic, not through the inner self, some magical inner self, but through a deterministic flow of activation. And, now that goal is parsing your activation to really zone in on that particular question where there's ambiguity, and you are examining those chains of thought, and that feels different. And, all of a sudden you are in this different mode experientially without the underlying pattern of connectivity, the pattern of neural network, changing. It's the same neural network. Because, what this does for us that the inner spirit doesn't do, is it explains how we might arise from the things that we see in our brain. The option is to be a dualist. So, there's a story in the book where Descartes--the French philosopher, mathematician--he's walking around in a garden, and he steps on a stone, and all of a sudden the arm of a nearby statue moves. And Descartes is, like, 'Wait, what happened here? How did the arm move?' So he talks to the gardener, and the gardener explains--in French, I'm sure--that there are hydraulic pipes that go from the stone to the statue, and that when he steps on it, the water pressure through the pipe makes the arm move. And, many of us would have said, 'Oh, okay.' But Descartes, he's struck by this. He's, like, 'Ah, maybe this is the mechanism by which we move our hand away from a fire: that maybe there's some signal that the fire, like a water, transmits to our brain, and that the brain then sends a signal back into the pipe, and it's all done with water-like pressure.' And, he imagines the nervous system to have microtubules filled with fluid. And Descartes is wrong in every way. But he's brilliantly wrong, because now he's after a mechanistic way of explaining the things we do. And he makes a lot of progress. So he says, 'I can imagine this machine'; but he stops short because he says, 'When we do mathematics or fall in love or write poetry, it can't be these tubes.' He can't imagine these tubes. Well, with a neural network, you can. And you not only can imagine them, you can give useful accounts of how these things emerge from the interactions of simple processing units. And, Descartes's answer was, 'Yeah, moving away from a fire mechanistic--animals have this mechanistic stuff, but we have this inner spirit that lives in our pineal gland.' And so, that's how dualism started. And, in some ways, that tradition, Descartes was making progress. But he didn't go far enough. He didn't make the possibility that perhaps it's useful to think of our entire brain or all our actions coming from mechanistic processes, like water in the pipes that he was thinking about. And so, he is left with 'the spirit.' And people have tried to weigh dead bodies after--the spirit is not that useful because if the spirit knows that you want a taco instead of a burrito, well, how does the spirit know that? How does the spirit know that you like this and not that? It's not explaining anything, right? I mean it's basically a restatement of the problem. Russ Roberts: Yeah--well, there's something unsatisfying about it. Let's call it God for the moment. Actually let's call it the soul, which is the Western way of thinking about this mystery. Because, for Descartes and many others, it's just inconceivable that it could all be the same: it could all be mechanics and deterministic. So, it's true that the soul is a tautological, unhelpful answer. But of course, in some sense, the neural network is also, in that it's--well, it's all that, right? It's always that. That's your explanation every time. There's no room-- Gaurav Suri: Well, let me stop you there. Russ Roberts: Sure. Gaurav Suri: The neural network is not the same as the soul because what is possible-- Russ Roberts: No, I didn't mean to say it say it was. Sorry. Go ahead. Gaurav Suri: But, it's more helpful because what it shows is that it can--with the same inputs, with architectures of neural networks, you can see how the output emerges. You can see, like you can see a ball rolling down in your example earlier. You can trace the activations and you can say, 'Oh, this is why false memories happen, or context happen. You can trace these activations; and it's magical. Like, when you see this emergence happen, and you trace it, and you understand operations of the mind like you understand a garage door opener, for me, this is transformative. And so, while the soul might be a black box, the neural network isn't. You can trace it. Russ Roberts: Well, parts of it are a black box, which in the scientific worldview we assume it'll all be revealed eventually. It's just a matter of time. But, I still have some room for mysticism, and I think at the end we'll talk about that, I hope. |
| 57:10 | Russ Roberts: But, let me shift gears for a minute. Let's talk about flattery. So, I've never read Dale Carnegie, but my understanding is, is that Dale Carnegie says when you're talking to someone, use their name a lot. 'Gaurav, I really like your book, Gaurav.' So, I keep invoking your name. And I think he didn't know this, but I think he was suggesting that that's starting to groove your channel there to see me in an attractive way. And, again, this is the kind of thing your book's done to me. Before we started recording, I told you how much I liked your book. I was flattering you. Now, I didn't think I was flattering. I thought I was actually connecting with a thinking human being whose work I had enjoyed and respected. And, I'm in a similar--I like to think, at least--I'm in a similar profession of intellectual understanding and science. So, I wasn't flattering you, Gaurav, and trying to make this a better interview. I was connecting with you on a human level. And then I start, in the middle of this, I'm thinking, 'Now wait a minute, is that what--' and, all these nice things I've said along the way about how good the book is. Now, you haven't listened to thousand-plus episodes of EconTalk, so you don't know that actually I don't do that very often, so it's actually kind of special. But, on the other hand, maybe I'm just trying to suck up to you and have a better interview. Gaurav Suri: So, there's pause here for dismay. I think there is a pause for a lot of hope as well. Just because it's that you are connecting at me, or at least a small part is you were connecting at me at a human level, does not negate the fact that you like the book. Now, right? So, one of the things that a neural network view allows is it allows for multiple causes of the same thing. It's not saying that either this is true or that's true. Either this person is a good person or they're a bad person. Either this is justified or it's not justified, Russ Roberts: Or I'm an insincere flatterer or-- Gaurav Suri: You like the book-- Russ Roberts: a deep thinker who appreciates your book like few interviewers do. Gaurav Suri: Right. So, the thing is that, because we've got this analogy of many pools--like, a lot of the pool in your early comment about loving this book--which I'm now honored by because you don't often say that--came from the pool of the fact that you were reading this book and maybe engaged in it. And, maybe a little trickle also came from, 'Let's connect with this guy. This is a nice thing to say to break the ice.' But, that doesn't negate the water in the other pool. Russ Roberts: True. I love that. |
| 1:00:12 | Russ Roberts: I want to ask you about a topic that I've noticed is on the rise in people's thinking, which is intuition. So, if you'd asked me five years ago about intuition, I would have said, 'Oh, that's just your gut, that's just guessing, that's--' etc. And, I wrote a book called Wild Problems, which basically said we overestimate our ability to apply analytical thinking and data to the decisions we make, and we're fooling ourselves that we think we can do that. And then I read Patrick House's neuroscience book, 19 Ways to Look at a Consciousness, where he says intuition is when your brain is processing data in ways that you don't fully understand. And then, I interviewed David Bessis--that hasn't come out yet; it's coming out next week, by the time this interview comes out listeners will have had a chance to hear it--who says, 'Actually math isn't equations. It's vision. It's seeing things that are almost--imagined, but you see in your mind's eye--and the equations are just to make sure they're right or not. And sometimes they're not, and you need to check, and that's how you hone your intuition.' I've got an interview coming up soon with Gerd Gigerenzer who has a book called The Intelligence of Intuition. So, intuition, which I see as sort of a pushback against the algorithms of computer science, and the extremes of social science run amok with data analysis and overemphasis on analytics when so many things can't be measured--how do you as a neuroscientist think about this way of thinking about what we're capable of as human beings in cognition? Gaurav Suri: Yeah. Do you watch Jeopardy? Russ Roberts: I've seen it. Don't watch it often, but I've seen it. Sure. Gaurav Suri: So, I watch Jeopardy, and I have this experience of knowing that I know the answer, but not being able to verbalize the answer. Knowing that this is knowledge I have. Or, I might be sitting in a roomful of professors, and I might get the feeling that I have something important to say here, and yet not know what I want to say. Well, what is that? Well, in the neural network view, that's the awareness of activation. There's a lot of activation flowing, and you have awareness of it, and that might even guide your actions. You might take this exit instead of staying on the freeway. And you might call that intuition. The thing is, that we haven't talked about consciousness. What's non-intuitive? Non-intuitive is when we have a conscious experience and when we can trace logic from A to B. Right? All men are mortal. Socrates is man. Therefore Socrates is mortal. We can trace that logic in a conscious and explicit way. Well, neural networks don't demand that at all. In fact, neural networks think that this formal way of thinking arises with learning from our basic associative machinery, which at its core is intuition. Like, nothing in this set requires us to be consciously thinking thoughts. There is a passage in the book where we show a symbol that is halfway between an H and an A. In one case--the phrase is 'THE CAT'--and in one case, the H/A is surrounded by the T and the E of 'the,' and in the other case, the same symbol is surrounded by the C and the T of 'cat.' And, we read that symbol as H in 'the,' and A in 'cat' because context is traveling to that ambiguous symbol and informing us what it is. Now, none of this reached--like, we process this not thinking, 'Oh, is this a the or is this an H or an A?' Right? Many of us don't look at that phrase and don't process it. And a lot of our processing happens like that. Automatically, context, multiple causes, and answer starts to form, actions start to happen. Sometimes they rise to consciousness. Sometimes we can do formal logic with it. But, the point is that this system allows you to know answers even if they haven't yet risen to the conscious level. We don't mention consciousness in this book till much later. It's in Chapter 10. It's because the point is that activation is the currency of thought. Activation, which is the electricity in neurons influencing other neurons, which make you eventually get off the couch and go to the refrigerator or take this exit and not that or choose taco instead of burrito. These are activations. Sometimes these activations make it to consciousness. Some thoughts even happen within the formal system. But we only have a lot of focus on those thoughts because we only have access to our consciousness--by definition. And so, we think that this strange thing happening in the bedrock is some mysterious intuitive knowledge. Meantime, it's very possible that it's most of our knowledge, right? It's--most of the things that we do, like wearing your pants in the morning, you are not consciously thinking, 'Okay, I did one leg, now I'm going to do the other leg.' Right? And so, a lot of these things are happening automatically. They're chains out there. And sometimes we pause and we reflect, 'Gee, how did I know that?' And, we can't put a word on it. We call that intuition. Meantime, that's the engine. |
| 1:06:36 | Russ Roberts: Yeah, and it's a fantastic, I think, important realization. We've talked many times in the program about how if you can't remember something--which being 71 happens more frequently than it did 10 years ago--you just put it down. And, what we often do is we think, 'Oh.' I often go through the alphabet to help me try to remember a name of an actor in some movie or somebody I met. And, the better system, of course, is forget about it for a minute and then try to remember a minute or so--to think about it again, and it snaps into your mind immediately. It's the same thing as you're working on a problem in the shower, and you stop thinking about it, and sure enough it comes, you realize what the answer is. Or, I love this one: I got-- Alan Watts in The Wisdom of Insecurity, which is a book I kind of love. It's a fascinating book, I'll mention it for a second. It was written in 1951 or 1952, I think. It could have been written yesterday. It's all about how the pace of change, technology--we're not happy. We're struggling to deal with reality. And it's a very thought-provoking book. But, at one point he gives you an 11-letter jumble of letters, and he says, 'If you want to know what word this forms, you could sit and ponder it and try starting it with different syllables and different letters.' He says, 'It's much better just to look at it and not think about it and you'll see the answer.' And he says that, and then he doesn't tell you the answer, first of all. And then he says, 'If you try and it doesn't work, please turn the page because you're going to get really mad at me.' So, I'm reading the book the other day for the second time, and I come to that page, and I forgot what the word is. So, I think, 'Oh, that's okay. I've got the trick. I'll just look at it and not think; and it'll just come from the intuition to the back of my brain.' And I can't get it. Eventually I got it. It came back to me eventually. Gaurav Suri: What was the word? Russ Roberts: I'm not going to tell you because I want listeners to open the book. I don't have it here, but I think it's a-- Gaurav Suri: Okay. So, I have one other memory trick for you. So, let's imagine you're trying to think of an actor--let's say Jennifer Aniston--and you can't recall what her name is. But you start thinking about her movies, and you say, 'Oh, she was in that. She was in that. She was in that.' And, you start thinking about her looks; and you start thinking about maybe where she's from. You recall that she's from New Jersey. When you're trying to recall something, if you recall many variables about that thing that you know, those things have channels into the name Jennifer Aniston. So, if you get activation into those channels, invariably the name will pop into mind. |
| 1:09:18 | Russ Roberts: Love it. So, I'm just going to pick on you for a sec about the Adam Smith thing. The Adam Smith quote you give is correct; it's from the Theory of Moral Sentiments. But then you have a mock dialogue of Smith's with, I think it's Freud? Is it Freud? I think. And, you have Smith talking like a modern economist. He wouldn't talk that way. And not just he wouldn't use the jargon: he actually didn't think that way. But, it's just a cheap shot at Smith that you didn't intend. I'm giving you the benefit of the doubt, Gaurav. Gaurav Suri: Yeah. So, let me tell you about those dialogues. What I've realized is that when you're trying to do science, you've got to have a human element in there. And so, the book is filled with these dialogues between Bertrand Russell and Wittgenstein, between one neuroscientist saying one thing and the other. They're all fictitious-- Russ Roberts: They're beautiful. They make the book. It's a sweet part of the book-- Gaurav Suri: We wanted someone to represent the 'humans decide things rationally.' So, when you do this, there is an element of caricature involved. Because, people are nuanced. No one person takes this extreme view. But, we picked Smith and Freud, Adam Smith and Freud. And I think Freud is an interesting case because, I think, Freud wanted to be a neural network guy. He was dissecting frogs and whatnot, and there was no--that system of thought hadn't been developed. He didn't have the tools. And so, he comes to this conclusion--and Freud got a lot wrong--but what the conclusion he comes to by talking to people, his patients, that what they're saying is not really responsible for what they're doing. These appear to be stories. And he has this revolutionary view that is the unconscious. He doesn't know how. He's not got this model of units activating other units. But he's got this Copernicus-like idea that it's happening under the conscious radar. And, you know, he was right about that. Maybe he got a lot wrong. I mean, sometimes a pipe is just a pipe, and sometimes all of this emphasis on childhood trauma is sometimes overstated--sometimes not, but sometimes overstated. But, what he got right is really revolutionary because: because we are conscious creatures it's so easy for us to think that, 'Oh, this is why I am doing this, because that's my conscious thought.' And, Freud is saying, 'No, no, no. Wait a minute.' That, for me, it has a magnitude equal to saying that the earth is not the center of the universe. Because it's something that appears to be true: The earth is flat, everything moves around us. That appears to be true, but it's not. It appears to be true that all our thoughts and actions are guided by this conscious logic. It appears to be true, and it's not. And, Freud was the first to point that way. And, what we argue is not that everything is unconscious. We think that the conscious thoughts themselves are activations that are interacting with the rest of the system. But there is a rest of the system. |
| 1:12:42 | Russ Roberts: So, I want to read a quote, and a couple more things I want to touch on if we can. This quote is a pretty good summary of the perspective of the book. And, you don't know me: I lead a religious life. I would say I believe in God, but that's an awkward--those are words. I'm not quite sure what those words mean; but I do act as if I believe in God. So, let's just leave it at that. I do certain things that other people don't do. I don't use my phone from sunset on Friday night to dark on Saturday night--because I'm a religious Jew. So, I don't eat certain things, right? Now, in a certain sense, your book is an attempt--not an attempt--your book strips away any possibility of a religious--a role for God in human action. For a whole bunch of reasons, which I'm very sympathetic to despite what I've just said. I think when we do science, invoking God is not helpful. But, you have a quote here I want to read and I want to react to it. The quote starts off talking about LLMs--Large Language Models--again, the things that we have become a little bit dependent on in our daily life, using ChatGPT and Claude. Here's the quote: So LLMs are capturing some aspects of human thought today and will likely capture even more aspects of human thought in the future. Importantly, we remain committed to the fundamental principle that thinking processes in humans and LLMs can be understood as physical processes. In our view, neither natural nor artificial intelligence relies on a touch of ineffable magic or spirit stuff. Both processes unfold within a neural network, and both processes are deeply emergent in many different senses of that word. They emerge from the interactions of neurons, from exposure to experience that guides their learning, and from the scale of the neural networks involved. And I think about that, and we started off with a conversation about electricity jumping through dendrites and other neurons, sometimes multiple neurons, and then certain things reach a certain level of activation, and my arm stretches out or I'd have a certain thought. And, I think about what we've achieved as human beings with LLMs--which is insane, absolutely insane. And we don't fully understand it, which is even more insane. We've created something we don't fully understand. And, I'm thinking, 'Isn't that magical enough for you?' Because you use the word 'marvel,' and I think about it all the time: the human brain and an LLM are both two extraordinary things. Where did they come from? Okay, we can tell a story. But there's something magical there. It's not natural to us that those trillion, or not trillion, that the 86 billion mechanical things in my mind can create Beethoven's Ninth. And-- Gaurav Suri: I completely agree. I completely agree-- Russ Roberts: "Two roads diverged in a wood/ I took the one less traveled"-- Russ Roberts: That's--and maybe an LLM will someday reach something transcendent like that, but I'm pretty impressed when it helps me brainstorm for a meeting about what the agenda ought to be. So, that's an amazing, miraculous thing to me. So, the wonder and awe that I think we both feel for this, I think they're related. I'm not going to push it, but I think they're related. Gaurav Suri: So, I share the wonder and awe, right? So, I share humble bewilderment and humble gratitude to see the majesty of the things around me. I am blown away by this. And you could call that an experience akin to a religious experience. I have a lot of awe for the universe, and I have a nearly holy desire to understand it, right? Russ Roberts: Yeah. Gaurav Suri: And one of the things that I've realized is that complexity comes from the application, often comes from the application of simple rules to simple infrastructure. But, again and again and again, apply--so you apply a simple rule, the thing changes a little. You apply it again, it changes a little. Then you apply it again, it folds on itself. It's like the murmuration of swallows: and knowing that the murmuration of swallows is occurring because each bird is looking at its nearest neighbor, for me, does not decrease the majesty. Russ Roberts: I agree. Gaurav Suri: It actually increases the majesty, right? Russ Roberts: Yeah. Gaurav Suri: I mean, this is my point: I am a part of this universe, no less, but also no more than the trees and the stars; and that there's a deeply awe, rich, human gratitude that I have for this experience, akin to a religious experience. Now, I want to say that the book is actually silent about God, just as it is silent about free will. Because, you know, these are concepts, right? Is the number three a real thing or not a real thing? Well, where is it? Where is it in nature? It guides us. Free will--we're deterministic systems--what is it? Well, it's a useful construct. God can be a useful construct. I mean, through history, looking at the positive side, it's caused wars and whatnot, but it's given people community. It's given people a reason to come together. It's given people courage and fortitude when faced with despairing circumstances. It's given people the capacity to somehow go on when suffering is unbearable. What? We should call this construct useful or useless? It's a useful construct in many contexts. All I'm saying is, for me, it's not a useful construct to understand what's happening around me. Russ Roberts: Yeah, I'm just [inaudible 01:19:35]-- Gaurav Suri: It can be a useful, even a beautiful construct for many other purposes. Before this--before The Emergent Mind--I wrote a mathematical novel called A Certain Ambiguity. And A Certain Ambiguity is really a conversation between a secular, non-religious mathematician and a judge who is religious, who is seeing whether this guy committed blasphemy or not. And, the whole book is this guy saying, 'Look, you can understand the universe.' And, the other guy saying, 'Well, there is majesty.' Well, not to give you the book--this was 2007--but the point is, it's both. One doesn't impinge on the other. It's crazy to think that a concept that's a useful concept--a construct--is true or not. Is it useful? Does it serve a purpose? Do we want it in society? Does it touch me? I mean, these are very important things. So, far from--we're not antagonistic to God or [inaudible 01:20:36]-- Russ Roberts: No, you're not, I don't mean to accuse you, although the line about magic stuff is, I think, a, is a-- Gaurav Suri: Yeah. Magic is a problem when you're using magic to understand the universe, right? We think the neural networks is an attempt to understand ourselves--who we are, what we do--with a mechanistic lens. And, for me, what's transformative is that, 'Gee, I can make progress.' You know, for the first many decades of my life, I saw this strange creature. I'd look in the mirror, and this somehow knows he wants this and not that, and somehow--and I'd think what's going on up there? And, we have the analogy of a computer, but it's not quite a computer, and then, oh, yeah, there's emotions and they're interacting. But it's all vague. It never came together. And this neural network provides a framework for you to trace the activations. And one other thing, Russ: You said that LLMs are, they go outside our understanding. That is true, but it's a very specific way in which that's true. So, the point there is that an LLM, just like our brain, is a neural network. It's a very different kind of neural network. Its learning is different. How it makes the channels between units is different. We do it somehow more efficiently. And we don't know how the brain does it. We don't yet know. We know how the neural network, how the LLM does it. Russ Roberts: The point being, we don't need to train on a hundred billion-- Russ Roberts: documents to figure out the cat is black means something that to us then-- Gaurav Suri: Yes. That's right. And we also don't need to use 5% of the energy of the whole country to run our data centers. Our brain is 20 watts of--a very small amount of energy. But the similarity is that they're both neural networks. Now here's the thing: These are complex neural networks, and they have emergent properties. And, for some properties, you can trace the input, and you can see how it got to the output. But, when you make your network really complex, and you hook it up with simple rules, you get emergent patterns: you can understand the big-picture causes, but you can't understand the specific causes. That's what's going on with LLMs. So, when people say that you can't really know what the LLM is doing, they're right in the sense that you can't really know why the flock of birds turned this way and not that. In principle, you can; but that'll be very complex, and it'll involve tons of effort, right? |
| 1:23:17 | Russ Roberts: Yeah. That may be way beyond us for a very, very long time. All I was trying to say--I was teasing you--but all I was trying to say is that I don't have any problem thinking about what historically in the West has been called the soul as this miraculous consequence of electricity going on in this little tiny skull mine. It's extraordinary-- Russ Roberts: And, I don't have--it blows me away. I find it deeply moving. And to think of it as having a divine element, even though I can't weigh the soul--and I can't, by the way, we can't find the emergence, either. Emergence is a concept we use to try to put a word to this inexplicable richness that comes from 86 billion things doing their thing. It's crazy. Gaurav Suri: Yeah. I have no problem with the soul. And for me, the soul is an emergent consequence of activity in our neural network. My problem, which is a personal problem, is I don't find the soul a useful device to understand who we are and what we think. I understand it as an emergent consequence, a word that reflects the depth and the intricacy of our neural network. And, the misunderstanding that a lot of people, I think, have is that understanding something reduces its majesty. Russ Roberts: Oh, I agree with you-- Gaurav Suri: I think quite the opposite. Russ Roberts: I agree. |
| 1:25:04 | Russ Roberts: But, I want to make a small defense of the soul for--if you still have time. Gaurav Suri: Yeah. Of course, yeah. Russ Roberts: You made a list of what was good about religion. You were kind, and you made a short list of what was bad. It's a long list of what is bad. Gaurav Suri: It's a long list. Russ Roberts: Just, like, you mentioned it creates community; and of course community can be glorious and exhilarating, and it can be violent and dangerous. Tribalism is a big part of who we are, and there are wonderful things about being part of a tribe, and there are cruel things. But in your list of good things, you left one out that I would have listed, and I want to use it to think about the implications of neural networks for our lives, not just for our understanding. And that will transition us to the end of the book, which is really glorious. One of the things religion is good for is it demands that we be better. It demands that we seek to surpass our selves, both in the sense of recognizing our smallness--it feels like we're the center of the universe--but religion forces you to say otherwise. Annie Lamott has a great line. She says, 'God's name is Not-me.' That's God's name: Not-me. So, if I can get out of myself and my own head--and God asks me to do many things. Not using my cell phone on the week, on Friday night, is one of them, but much more than that: Visiting the sick, and comforting the lonely, and a thousand things that I think are important about a religious life lived well. Is that in my neural network? Without this concept of God and religion, at least as I understand it--there are a lot of religions, so I just can only speak about the one I know well, Judaism. God expects me to be a good person, to improve the world. And I know you, Gaurav, because I've read your book, you feel the same way. Is that entered into the culture of our lives? Would that have entered without this idea of God? Even if it's not real? I think it's real. I like to think it's real, at least. I can't prove it. I have some evidence, but it's not very good. Self-improvement, making the world better: talk about it. Gaurav Suri: So, let me start with where I completely agree. Having guideposts to do good will make us do good. If we hold each other accountable to behaving certain ways, those are inputs into our neural network, right? It's an input into the neural network that--you know, you pick up the phone on Friday and your family, your wife, maybe comes in and says, 'What the heck are you doing?' And, there's an unpleasantness associated with that. Or, the unpleasantness is from your conception of what it means to be religious, and that will make you feel bad. So, there are multiple constraints that we put on each other. The law is one constraint; and we need the law. And, religion is a vehicle for putting constraints on each other. And, I'm dead with you that we need constraints, because in a neural network, our actions are shaped by many things, including those innate connections that can be prosocial, but can also cause us to do great harm to each other and ourselves. Right? So, how do you control that? Well, what you do is you build a society in which there are mutual constraints that nudge us towards building a better world and building a better ourselves, and emphasizing, as Lincoln called, the better angels of our nature. And, that doesn't happen automatically. It happens with thought. That's what culture is. And religion has been the glue of putting this--thou shall not kill or do this and that--for many, many societies. And, it has been, that part of it has been incredibly useful. And it has a great synergy with the idea, the neural network view, that our behavior is an outcome of multiple constraints. And if you can get one of the constraints to be the constraints that we put on each other, you're doing good. Now, there's a deeper point here. Religion serves another purpose: because of its appeal to the majesty of the universe and a higher power, it sort of gets these neuromodulators, these emotional systems, going that make for greater activation, right? So, we respond more in the presence of neuromodulators than we do without. It makes for better learning. Neuromodulators make deeper channels. So, if you've got emotion--like, music does the same thing--the reason Jaak Panksepp, a neuroscientist, used to say that the reason we so are beguiled by opera is because it sounds like the wailing, the suffering, the crying, right? It's got that neuromodulatory pull. And so, religious experience, I think, is like that. It puts constraints, which is great. It uses the power of emotion, which is great. It puts these checks and balances on each other. It gives us quests. These are all great things. These are fabulous things. If we can avoid the bad thing, that's even better, right? None of that is my--as a scientist, my question is in understanding myself, does religion play a useful role? My answer to that particular question is: I haven't found it to. But, that's a limited thing, right? If you are moved to answer the eternal question of who you are, I had, say, that thinking with neural networks can be a transformative experience. If your goal is to put constraints on each other through the church, or the synagogue, or the temple, then that's a function that religion is doing; and neural networks are silent on that. In fact, we're celebrating the fact that mutual constraints on each other are going to be important inputs into our neural network, and we're explaining that people are not fundamentally good or bad, right? This is the power. It is not like, 'Oh, so-and-so just is evil,' or 'So-and-so is just good.' No. They are a product of their biological and lifespan, developmental destinies. And so, the neural network is saying, 'Well, yeah, instead of being so quick to say this person is good and this person is bad, let's think about these constraints that we can put on each other that make us do better things.' And, that realization doesn't happen until you get inside and realize there is no--they're neurons that are not good or bad. They're interacting neurons. There isn't this well of good or bad in us. And that view only comes by thinking about the mind in this way. Russ Roberts: I love that. I want to see if there's a practical implication of it for my behavior, which there is a little bit at the end. We're going to talk to--well, we'll get to it, I promise. |
| 1:33:16 | Russ Roberts: But I want to think about the following. Suppose I look into my soul, my behavior, my thoughts, my actions, and I'm disappointed: I can be better. I want to be a better friend. I want to be a better interviewer. I want to be a better husband, better father, better son. And I'm motivated to do that. And, you have some nice thoughts on motivation in the book; we're not going to get into them. But, here's my question: If I see myself and my actions as the product of a neural network, does it have any practical lessons for how I can be a better person, if that's my goal? Because it's not unrelated. It's not a small thing. In fact, you could say it's half--or close to more than half--of the whole thing as a human being, as I go through the world. So, I've read this amazing book, and it says, 'Actually, because all these neurons are firing, and sometimes they get to a high enough level and enough of them are firing, they do certain things,' and I don't like the outcome. I want to do better. Russ Roberts: What do I do? Gaurav Suri: Yeah. So, what we're talking about is: What does the neural network perspective offer for making a better life? Russ Roberts: Self-transformation. Gaurav Suri: Self-transformation. But, I want to start, though, not with self-transformation, but with kindness. And, let me say why. There's a lot of-- Russ Roberts: This is the end of the book, by the way, that I love that, that I said we'd get to. So, you've jumped in. Go ahead. It's beautiful. Gaurav Suri: So, when I realize that--we grew up in different parts of the world with different experiences, our ethnicities are different--but at a deep level, our neurons are doing some very similar things. Your neurons are connected differently than mine. And, what that's done for me, Russ, is when I see people do things that I don't agree with, I don't go to that they're evil or they're stupid. I go to thinking about what are the causes in their environment that's making them do that. And so, I'm able to lead with kindness, because, you and me, and every person, are deeply, in a deep sense, we're the same thing, right? Our neural networks are different. Our responses to the same input is different. Some of us do many prosocial things, and others do very anti-social things. But, this idea that there are channels--there's input, there's water flowing through those channels that moves to the muscle--the fundamental mechanisms are the same. And, but for accidents of context, we are the same. And, that makes me lead with kindness. And for me, that's transformative, right? That changes how I think about the world around me. The second thing is: What if I want to improve myself? Right? What if you want to be a better interviewer? Well, yeah, it's a deterministic system, but goals are celebrated in the neural network view. We, in fact, want to understand what the heck it means to get a goal, and set a goal, and follow a goal. And, you can explain this in a deterministic fashion, but you realize that the goal is really about attention, right? That, the more you activate the goal, the more influence the goal is going to have on your action. And, that attention means connecting different things to the goal, and making time and space to pay attention to that goal, and looking at other people who follow goals and think about goals. And, these are all--we're all influencing each other, and the environment is influencing us, and within this deterministic system, we can make goals and we can attend to our goals. And, those goals have the power to shape our actions, which are ultimately the root of transformation. The fact that it's a deterministic system doesn't take away from the fact that we can make goals and follow them. In fact, understanding what a goal is and what a goal isn't makes us more likely to follow that goal. So, whether it's transformation, whether it's kindness, whether it's leading with forgiveness, first--I don't mean forgiveness of behavior. No, shoot, we can't forgive behavior, right? Because we need a functioning society. What I mean by forgiveness is making room to understand somebody's context can lead to greater understanding, can lead to greater good will, can lead to greater cooperation. Understanding how our goals come about can lead to self-transformation. Understanding that we are creatures that have this magic of thought floating on this planet, hurtling through the galaxy, which itself is hurtling through deep space. We've got this island of thought, and it's precious and likely very rare, and that we can be kind to each other and build something glorious that makes us thrive as a species, is majesty in its rawest form. And, for me, understanding who we are enables this journey. Russ Roberts: My guest today has been Gaurav Suri. The book is The Emergent Mind. Gaurav, thanks for being part of EconTalk. Gaurav Suri: A great pleasure. |
How is your brain like an ant colony? They both use simple parts following simple rules which allows the whole to be so much more than the sum of the parts. Listen as neuroscientist and author Gaurav Suri explains how the mind emerges from the neural network of the brain, why habits form, why intuition often knows before language does, and why our post-hoc explanations can mislead us. The conversation then grapples with free will and responsibility without mysticism. Ultimately, Suri remains in awe of the emergent mind and at the end of the conversation makes the case for the essential importance of kindness and forgiveness.
READER COMMENTS
Eric
Nov 17 2025 at 12:26pm
If we supposed that neural nets in the brain provided a complete mechanistic picture for understanding our perception, memory, and decision making and that this is the whole story of our nature, then what possible neural net story could be told when…
the heart has stopped,
blood is no longer being pumped,
the brain is not receiving oxygen,
neural activity has ceased,
the body’s eyes are closed,
… and yet the person is able to make observations about events that are happening (sometimes in other locations not with their body) that are later verified after the person is revived and can report their observations?
The advance of medical technology has made near-death experiences (NDEs) more common and they have become a subject of research over more than 35 years. These events include experiences in which people have made observations of details that are later verified, even though they could not have been making those observations with their physical eyes or processing them through their brain’s neural networks that could not have perceived the events.
For an introduction to the topic of NDEs and the research that has been done on them, here is one possible place to start. It includes a summary of results from “35 years of research into NDEs by doctors and other professionals, fully documented for those who want to study further.”
Near-Death Experiences as Evidence for the Existence of God and Heaven: A Brief Introduction in Plain Language
By Dr. J. Steve Miller
neil21
Nov 17 2025 at 4:14pm
Great episode, I’ll be straight out to get the book. Extremely clear reconciliation of free will with materialist neuroscience.
And of course, when you closed with our designed environments shaping our pool-pathways (our souls?) I of course think immediately of Kunstler’s now 20-year old rant on urbanism in America. https://m.youtube.com/watch?v=Q1ZeXnmDZMQ
JanP
Nov 17 2025 at 7:08pm
An interesting set of ideas. For an early model of human mind as an emergent phenomenon, see “The society of mind” by Marvin Minsky (1985).
Nick Ronalds
Nov 17 2025 at 9:03pm
A stimulating episode. I also appreciated the inference that, given our limited understanding and baked-in blind spots, kindness is the best default move forward.
One omission surprised me. Professor Suri often spoke as if the brain’s workings, while complex, are essentially understood—an intricate network from which consciousness “emerges.” But philosophers of mind still emphasize what David Chalmers called the “hard problem of consciousness,” and it remains unresolved. (Admission: everything I know about the philosophy of mind is from podcasts.)
The issue, they say, isn’t just mapping neural patterns to particular experiences—coffee versus garlic. Maybe neuroscience will eventually crack that. Isn’t the deeper question why any neural activity gives rise to subjective experience at all? Why should a specific physical pattern produce the smell of coffee rather than nothing “felt” from the inside? Emergence explains correlation (or will eventually, perhaps), not why there is something it is “like” to experience those patterns.
Scott Gibb
Nov 18 2025 at 8:11am
You emphasized the brain’s deterministic nature, and you mentioned goals, attention, and improvement, but you didn’t mention stochastic aspects of the brain.
Electrons are quantum particles, meaning that certain aspects of their behavior are stochastic. How can the brain be as deterministic as you emphasize, if it’s thought is governed by quantum particles that have a stochastic nature?
Alex Goodwin
Nov 18 2025 at 10:56am
Could it have to do with the wiring? I think back to my electrical engineering course in college. They broke down hardware into circuits with logic gates. Based on the voltage through a logic gate, it would act like a 1 due to their being enough voltage or a 0 due to their being a negligible amount of voltage. If an electron runs through a circuit, the process should occur, so maybe the stochastic element of the electrons determine which circuit/ neural network it goes through.
I’d be interested in someone more knowledgeable than me checking that though. It’s been over ten years since I took the course and it might not be the relevant framework to understand it.
Greg McIsaac
Nov 20 2025 at 2:11pm
I think the stochastic behavior occurs at the level of individual electrons, but the behavior of populations of electrons is more predictable, given the various electro-chemical forces acting in a brain cell or across a power grid or the internet. Consider a photovoltaic solar panel, in which photons of light are converted to electrical signals in a highly predictable way. Brain cells are small, but electrons are much, much smaller so that cells still contain a lot of electrons that behave as populations.
Alex Goodwin
Nov 18 2025 at 8:32am
This conversation reminds me of Albert Camus and his philosophy of absurdism. In The Outsider, his character Mearsault is being tried for murder and the interrogator is asking him why he shot the victim. There are several reasons you could have expected him to give, but he remembers back to the moment on the beach and says that the sun was in his eyes.
This obviously doesn’t go well for him, but Camus made it clear during the murder scene that that was the reason why he shot him. It wasn’t anything rational. It was a feeling that Mearsault didn’t understand, but he did correctly identify the cause in this case. He didn’t come up with a rational reason like we typically do when asked why we do things. Camus was able to identify this facet of humanity without even being aware of what a neural network is.
Gaurav Suri
Nov 18 2025 at 12:51pm
Thank you, all, for listening to my conversation with Russ. I am so happy to join this supercool discussion!
@Eric: I find it useful to think of neural networks as providing a helpful framework for making progress towards understanding the mind. I don’t think that they have all the answers. Not by a long shot.
@Neil: Thank you! I will check out Kunstler’s work.
@JanP: This shares some aspects with Minsky, but Minsky thought about the mind as being composed of a collection of specialists. We have tried to emphasize its emergent nature
@Nick: I’m far from making progress on the hard problem. What I am surprised by is how far we can get in understanding the mind without relying on consciousness.
@Scott: Among others, Penrose has spoken about links between quantum phenomena and consciousness. It’s sounds exciting, but I don’t quite see how it gets around the hard problem
@Alex: Perhaps the greatest first lines of all: “Maman died today. Or maybe yesterday; I don’t know.”
Luke J
Nov 19 2025 at 6:05pm
I was thinking that a creator being – “god” – might not appreciate the “construct” conversation. On the other hand, what human mind could really comprehend the fullness of such a being; construct is the best humans can do.
Thank you for this conversation. I truly learned a lot.
John
Nov 19 2025 at 8:24pm
This EconTalk episode with Gaurav Suri is deeply thought provoking. I love how he explains the mind in terms of emergence: simple neural units interacting in ways that give rise to really complex behaviors and thoughts. His ant colony metaphor is super effective how individual neurons follow simple rules, but together they create something way bigger than themselves. The conversation also tackles habits, intuition, and how our post-hoc explanations (the stories we tell ourselves) don’t always line up with what our brain was actually doing. They even wrestle with free will in a deterministic system that nevertheless feels full of purpose and it’s surprisingly hopeful, not nihilistic. In the end, Suri circles back to kindness and forgiveness, suggesting that understanding our emergent minds can help us treat ourselves and others with more compassion.
Gaurav Suri
Nov 24 2025 at 7:26pm
John, Thanks so much for this thoughtful note — you’ve captured the spirit of the conversation beautifully. For me, the hopeful part really does come from that tension you describe: a mechanistic, emergent system that still gives rise to purpose, responsibility, kindness, and forgiveness. I’m glad the ant colony metaphor and the focus on habits and post-hoc stories resonated with you; that’s exactly the bridge I’m trying to build between neural machinery and how it actually feels to be a person.
Blackthorne
Nov 21 2025 at 3:21pm
It’s interesting that in many of these conversations discussing consciousness and the “mind” in terms of emergence there’s always a focus on the tension between our everyday “naive” model of decision making and what decision making might look like in the emergent view. Truthfully though, I think this tension is a bit of an illusion.
If consciousness is an emergent phenomena, our notion of self is also emergent. So the background/subconscious factors that play a role in our decision making are also a part of our self. It’s possible that these factors are at odds with what our more foreground/conscious thoughts are, but being at odds with yourself is nothing new, it dates back to at least Plato’s tripartite soul. So we can say that sometimes we make decisions that our more conscious self would object to, but those subconscious components are also us.
This seems a bit pedantic but I think its important to note when we consider models of decision making. Individuals may make choices that seem irrational when viewed through the preferences of the more conscious component of their self, but the true preferences to be considered are that of the whole. This is all a simplification but I think you get what I mean.
David Bergeron
Nov 22 2025 at 7:23pm
In 1979, I was 19 at Rice University sitting in a physics class. I watched ants trying to move a dead roach on the floor. They were all pushing and pulling with no success. Then one ant started pulling on the roach’s antenna, and the body moved. That ant seemed to touch heads with other ants, and they also started pulling on the antenna.
I assumed the first ant recognized that pulling the antenna worked and communicated this to the others through a chemical signal (pheromone?). But this explanation requires the ant to recognize success, which seems more complex than the simple double pheromone trail you described in your example. What do you think?
I’ve always been fascinated by how intelligence emerges from simple neurons and what rule or rules guide them. Thank you for the podcast.
Gaurav Suri
Nov 24 2025 at 7:23pm
David, Love this observation – it’s exactly the kind of thing that got me hooked on emergence too. Also, where was this physics class being held 🙂
I agree that what you saw probably involves something richer than the ultra-simple “double pheromone trail” toy model, but it doesn’t yet require a reflective ant that conceptually recognizes “success.” You can imagine local rules like: when I pull in some direction and feel the load give a bit, my internal circuits get a little blast of activation; that makes me keep pulling that way and emit a short-range recruitment signal or change how I touch nest-mates. The others don’t need to understand why it works; they just bias toward copying the ant whose action pattern is currently associated with motion.
So the spirit of the example still holds: coordination and “strategy” can emerge from very simple agents whose rules are slightly more nuanced than in the textbook models. Your antenna story is a lovely real-world illustration of how a tiny bit of feedback-sensitive circuitry, multiplied across many ants, can look a lot like insight.
Randy Buchner
Dec 2 2025 at 5:57pm
Hello Russ –
Here is a favorite song which comes to mind from listening to the episode:
John Cale & Brian Eno / Spinning Away
Spinning Away
Brian Eno
Hmm-mm
Up on a hill as the day dissolves
With my pencil turning moments into line
High above in the violet sky
A silent silver plane, it draws a golden chain
One by one all the stars appear
As the great winds of the planet spiral in
Spinning away like the night sky at Arles
In the million insect storm, the constellations form
…
The song, in general, invokes some of the themes you discussed regarding having a sense of awe
The image “a million insect swarm” brings to mind the billions of neurons in our heads
The linkage between drawing a line and the constellations suggests imposing meaning upon the universe
[Hi, Randy. The lyrics to this Brian Eno/John Cale song are under copyright, so we have elided them so as to accord with fair-use laws. Also, I think you have mis-heard the word “storm” as “swarm.” A charming misunderstanding if ever.–Econlib Ed.]
Comments are closed.