デイビッド・イーグルマン:脳は「神経の議会」、AIの幻覚は仮説、そしてAIで民主主義を考え直す

It's just like all of us are humans with human flaws, right? And if you could have a system, which we already have, that can extract massive data patterns and simulate millions of possible future trajectories and figure out how to optimize things for a giant population, why wouldn't we be asking to be led by that? And one of the things that I think is really special here, the more I thought through this, is this could finally be the truest democracy that we've ever had. Yo, it's the future, don't hit the snooze. AI is here giving humans a boost. Team humans rolling, brains in the game. With Dan Turchin, y'all remember the name. People ain't in the mix, no stopping to blow. Better work, better life, yeah, we're ready to go. Robots and humans are collab, you know. Okay, let's start the show. Good morning, good afternoon, or good evening. Welcome back to a special episode of AI in the Future of Work. This is Dan Turchin, CEO of PeopleRang, the AI platform that automates employee service.

And you know the reason why this is a special episode? Because we're taking this one face-to-face, eyeball-to-eyeball, or based on today's guest, brain-to-brain. These conversations get intense, and so it's kind of a pleasure to be in the same place together. It kind of humanizes the conversation. And today's guest needs no introduction, but I will indulge him and I will introduce him anyway. Dr. David Eagleman is one of the most renowned neuroscientists on the planet. Among other things, he is a professor of neuroscience at Stanford, go Cardinal. He is the host and produced my favorite all-time science documentary called The Brain on PBS, nominated for an Emmy. he's a perennial New York Times bestseller he's a serial neurotech entrepreneur if I were to list his many accolades, awards, honors there would be no conversation because that would take all day

Lex Friedman called him one of the best science communicators of our generation one of my personal role models and heroes Stuart Brand said that Dr. Eagleman is the best scientist and fiction writer of our time. He has accomplished so much, and yet one thing that I can confidently claim Dr. Eagleman has never done is appeared in an episode of AI in the Future of Work that is all about to change. And, David, it's my pleasure to welcome you at what we call Team Human. Let's get started by having you share a bit about your backstory and how you got to where you are today. Great. Well, first of all, thanks, Dan, for having me here. I'm glad to check this off my list of things that I wanted to do. How did I get here? I, you know, I grew up in New Mexico outside the Albuquerque city limits in the middle of nowhere.

And so I'm sort of surprised that I made it out of there. But my father had an extensive library in our house. He built stacks of books like a library. And so I spent my whole childhood reading. And I actually, as an undergraduate, I majored in British and American literature. That was my first love. And then I was always taking a lot of science, though. I was always invested in science. And I finally discovered neuroscience my last year of college. And I was so hooked. And I thought, this is the greatest thing. And now, all these years later, I still think it's the greatest thing. So that's how I got here. We met recently at Dr. Ben Gertzel's, a good previous guest on both of our shows. I should have mentioned, Dr. Eagleman's also the host of the number one science podcast, Inner Cosmos. Amazing podcast, if you don't listen to it. But we met at Dr. Ben Gertzel's Singularity Net event recently. And that conference is intriguing in many ways. it's all about engineering a brain.

Let's just get started by having you share your perspective on should we engineer a brain? Can we engineer a brain? Great. Let me start with the can we. So my whole career in neuroscience, it always felt like AI was a cool idea, but it was really far away. And that's because when you look at the brain, you've got 86 billion neurons. Each neuron in the head is unbelievably complex. It's an entire cell. It's got the whole human genome in it. It's trafficking millions of proteins around in very complicated cascades. It's got all these little compartments where it's doing chemical reactions and little crucibles. I mean, a cell is as complex as a city. Okay. Okay, so what AI did when it started really rolling in the 50s and 60s is it said, look, let's forget all that complexity and just assume it's a unit, and the units are connected because, of course, neurons are connected on average, about 10,000 of its neighbors you're connected to.

So AI said, look, let's just get rid of that, make a neural network where you've got units of connections, you can change the strength of the connections. And so they ignored all of the complexity of the brain. And the truth is that most of us in neuroscience sort of snickered at AI for a long time, thinking this is just never going to work that well. And then suddenly it worked really, really well. I mean, there were, of course, hints of that. I've been following this field for, you know, decades. But, you know, it just started with Perceptron and then different sorts of ways of doing pattern recognition in neural networks and deep learning and so on. And it was just getting better and better. But obviously everything exploded with LLMs. That's when all of us really had our hair blown back by this. And so that's gotten me, I think, I mean, I'm just going to speak for myself, but I suspect this is true for a lot of my colleagues that we felt like, oh, wow, maybe, you know, this is all Mother Nature had to work with, with cells, which are inherently really complicated.

They're like little animals, each of them. but maybe it is some sort of simpler algorithmic thing where you just get intelligence out of that you get really deep pattern recognition out of it so there are many unanswered questions just from the point of view of the brain we don't know if AI is conscious we don't know if the current way the transformer models are built if that could be conscious we don't know if they actually have any sense of an emotion even though they certainly claim to. Anyway, there are a million unanswered questions. Of course, there are differences too, like an LLM is unbelievably energy intensive, whereas the human brain runs on about 20 watts. And most importantly, the human brain can do continuous learning, meaning every second of my life I'm taking in new information and learning stuff, whereas LLMs have a training phase, and then you ask the questions after that. So, there are differences, but I will say I am mightily impressed by what has happened with AI.

And there's something in neuroscience, for decades we've been talking about this, the computational hypothesis of the brain, which is, hey, could you replicate what's going on in the brain with a machine or in silico or whatever? Like, Dan, if I took your brain and I replicated all the important stuff out of beer cans and tennis balls, and it's running the same algorithms as your cells, then I should be able to say, like, hey, Dan, how are you feeling? You say, oh, I'm a little cold in here. Maybe I'm a little hungry or whatever. But it should be you and perhaps even conscious and so on. That's the computational hypothesis of the brain. And we don't know that that's right, but that's certainly the theory for us to chase all the way to ground level before we dismiss it. Anyway, what we see is a real boosting of that possibility by what's happening in AI. We see that, wow, if you really just capture the basic algorithms, something amazing is happening.

happening. Now, the question of should we, I don't know. I think I'm going to default on that question because it is what's happening. And so whether or not we should, that's clearly the way we're going. You are a world expert in neuroplasticity, or you describe it as liveware. Liveware. That's not so much software or hardware, it's liveware. And here's where I feel tension with the brain AI analogy. the LLAM essentially becomes static after it's been trained. All of the inference, all the compute that's happening at inference time is once these quote, very much finger quotes, neuronal pathways are hardened. But that's fundamentally not the way the brain works. To you, does that create tension in terms of that analogy? Exactly right. I mean, there are real differences between the way it's working and we're working. But to my mind, it's, you know, like clearly the proof is on the table that we're close,

that we're figuring stuff out. But this issue about the LLMs not changing their weights, this is the issue about continuous learning. This is why they can't do the continuous learning that we can. Every moment that we are walking through the world from cradle to grave, our weights, our connection weights are changing as we take in new information. So there are these big differences, but boy, I am really optimistic that we now have good reason to think, hey, you know what, this is computationally capturable. And we can surpass the human brain. I mean, it's already the case that, you know, just if you had a friend that read every single thing on the planet, you could ask him about anything, you'd be really impressed with that friend. So even before we hit AGI it already mind good Is recall equivalent to intelligence You know I don think it is but LLMs are doing so much more than recall I think a big part of intelligence, or let me say for sure, human

creativity, is the remixing of ideas. So what we do is we absorb things our whole lives, and we remix them. We bend ideas, we break them, we blend them, and we're coming out with new versions of things. That basic cognitive operation is what has led to our civilization. We build skyscrapers, we compose symphonies, we get off the planet, we do all kinds of stuff because we can take in the data and do something new. And LLMs do that too. They're really good at remixing information. And of course, you can prompt it any way you want. You know, give me a poem about Godzilla standing on Mars and petting a cat and whatever, and it does it effortlessly, it's really good at this, and I think that is one of its greatest strengths. In fact, just as a side note, a lot of people complain about AI hallucinations. I think often hallucination is the same thing as hypothesizing, which is what we do all the time. As scientists, we say, hey, what if this crazy thing were true or something?

And then the only difference is that as trained scientists, what we do is we make a hypothesis and then we follow it out. We say, okay, what if that were true? Okay, this is what would happen. And then you see if it explains the data better. You know, this is when Einstein says, hey, what if I were riding on top of a photon? What would the world look like? I mean, that's great. So anyway, when I hear about AI hallucinations, I sort of feel like, okay, well, that's a good example of remixing things. And maybe once in a while I get something good out of that. Most of the community feels like we're rapidly approaching AGI, Artificial General Intelligence, also feels that LLMs won't get us there and that there's maybe a neurosymbolic architecture or something post-LLM. You're pretty complimentary of LLMs. Do you think that this is the, I shouldn't say end state, that's not the right terminology, but how far can we go with LLMs? Yeah, great question. And I think LLMs won't be the final thing.

And already, exactly as you said, a lot of people are working on ways of mixing things. And I think the models currently by Anthropik and OpenAI and others are actually mixtures of things. A lot of this is proprietary and locked away exactly what they're doing. But it's not your grandmother's LLM. And I do suspect that the transformer model, this particular architecture, has its limits. But, boy, everybody is working on this now. I mean, every kid I teach at Stanford, that's all they're doing is, hey, what's the next thing that I can do here? How can I contribute? And so this is going to move very rapidly. But, yes, I do think we'll need other pieces of architecture. I think when we look at the human brain, what we see is it ain't one thing. It's not like, oh, here's the structure, here's the architecture. Instead, it's this weird melange of all these different architectures smushed together. This is how evolution always works. Evolution never says, okay, I want to solve this problem and then stops.

Evolution just keeps going and often comes up with many, many solutions to the same problem. And this is what we see in the brain is that what I described in my book Incognito is that the brain is a team of rivals. That's how I put it. I was basing this off Doris Kearns Goodwin. What's that? The reference to Abraham Lincoln. Exactly right. His cabinet was it. Exactly right. I love that. Yeah, when he got elected president, he said, you know, look, I want to get all my enemies into my cabinet so I can have this diversity of viewpoints. So anyway, I've suggested that that's really the way to think about the brain. if I put some chocolate bonbons in front of you, part of your brain wants to eat that, part of your brain wants to not eat it because it's all sugary, part of your brain says I'll eat it but I'll go to the gym tonight. You can argue with yourself and cuss at yourself and contract with yourself. And I think this is often overlooked that we are made of lots of different drives.

And by the way, this is something I wrote about some years ago, But I just recently found a quotation from Friedrich Nietzsche, who cast this in his own language in the 1880s. He said, essentially, that the brain is made up of lots of drives, and each drive philosophizes in its own spirit, by which he meant, you know, each one tries to stand on a soapbox and say, look, this is why you should eat the bonbons. Another one says, hey, look, this is a really good reason why you shouldn't eat the bonbons. They're all philosophizing trying to get what they want. And so the way I think this comes together is that the right way to think about the brain is like a neural parliament where you're made up of all these different political parties, all of which love their country, but have different ways of steering the ship. And so they all vote. And at different times, by the way, the vote will tip in different directions. and that's the direction that the ship of state moves in, you being the ship of state.

So coming back to your question, what we're going to get with AI is if it comes out like the human brain, we're probably going to have lots of different kinds of architectures bolted together and some sort of voting, and that's how it will operate. And by the way, that's, as I said, currently how a lot of systems already are. For example, if you ask ChatGPT a math question, it doesn't go through the LLM. It detects its math question and sends it all through the calculator. It saves a lot of resources that way. I'm proud to share details from today's sponsor. It is the AI Playbook, Mastering the Rare Art of Machine Learning Deployment, a wonderful book published by repeat guest on this podcast, Dr. Eric Siegel, A prolific speaker, writer, professor, rapper, I dare say. It is a great read, very entertaining, very relatable.

It presents the gold standard six-step playbook for ushering predictive AI initiatives from concept to deployment. It is just coming out in paperback form. It's been a crowd pleaser since it launched a few years back. and you will so enjoy learning from the great Dr. Eric Siegel. Go ahead and head out to bizml.com to get your copy of the AI playbook, Mastering the Rare Art of Machine Learning Deployment from Dr. Eric Siegel. Now back to today's amazing episode. We are going to spend some, not Dave and Dan, but other people will spend somewhere around a trillion dollars on compute infrastructure, building up data centers and things in the next 12 months. And this conversation, and particularly I'm so glad you referenced the brain running at 20 watts. It's just a reality check about, you know, you call it the three-pound machine, right? The inner cosmos, right?

It's such a stark reminder about what's unique about the brain and why kind of synthetically generating them in data centers It feels a little bit like a brute force approach, a little bit naive. What's your sense about the approach we're taking and kind of what I'll call brute force? You can choose to use a different term, but is that the best path to synthetically generating intelligence? I think I'd probably share the intuition that a lot of us have, which is that we're going to build these data centers, and there will be scientific advances in the coming years that make them really obsolete. But that said, I'm sure there will be plenty of compute that we need for plenty of other things, so it will get used somehow. But, yes, maybe it will turn out that we've got these giant data centers. I suspect that with many scientific trajectories you see this sort of thing. Just as an example, you know, I was in labs when the Human Genome Project was really getting rolling,

and people were spending a ton of money on these machines that pretty quickly became outdated, and now you can do sequencing with new tricks and techniques and just do it faster and better and cheaper. So this probably happens with everything, no exception here. By the way, we should mention that you studied with Dr. Crick. That's right, that's right. A little before your time, but you actually studied his talk. Yeah, exactly. I did my postdoctoral fellowship at the Salk Institute. I studied with Francis Crick there, among other mentors. And that was really a once-in-a-lifetime experience for a young scientist. because, I mean, for anyone who doesn't know, Francis Crick was the guy who co-discovered the structure of DNA but then switched over by the 1990s to try to figure out the brain. And it's a real opportunity for somebody like him. If you win the Nobel Prize when you're still young, you don't have any of the pressures of, oh, I've got to get this grant and run this lab and graduate these students and whatever.

You just get to think big thoughts the rest of your life. And that's what he did. He was the most big thought thinking guy that I've ever had the pleasure to hang out with. And your obituary about him was beautiful. Thank you. Beautiful writer. I'll share a link to that. The audience can appreciate it as well. Thanks. Right before you came off stage when we met, I think I interrupted you mid-sandwich. Perfect timing. You gave this brilliant keynote at singularity.net. very thought provoking kind of at the intersection of geopolitics and democracy and AI and you pose this question I might paraphrase but you can of course correct me should we ask AI to help us lead or should we ask AI to help lead us yeah I been thinking about it a lot maybe if you could share a little bit about the thesis Right I been thinking about this a lot too Really, what I'm wondering is, when are we going to elect an AI as president?

And when I first wrote this title down to make this talk for the Singularity Conference, I thought, you know, this is sort of, how do I put this, it's sort of a splashy title and whatever, but I actually, you know, I went through, hopefully I went through this talk very thoughtfully and constructed my argument carefully. I actually think this is inevitable and it's going to happen much faster than any of us think. I think, I mean, I'll just make up a number, but I think like within 20 years, my argument is why would we be forced every time to vote for, let's say, Kamala or Donald when both of them are humans, full of flaws, full of tribalism and, you know, 100 reasons to do things that don't have to do with the nation but have to do with who their donors are and what they're trying to accomplish and they want to get reelected, whatever. It's just like all of us are humans with human flaws, right? And if you could have a system, which we already have, that can extract massive data patterns and simulate millions of possible future trajectories

and figure out how to optimize things for a giant population, why wouldn't we be asking to be led by that? And one of the things that I think is really special here, the more I thought through this, is this could finally be the truest democracy that we've ever had because the Anthropic published a paper in March where they demonstrated that you could do qualitative research on people where you interview people in real depth, And what they did is 45-minute interviews with people, 81,000 people. And, of course, the AI is just asking questions and digging in deep and so on. They really get to know these people. Imagine that, let's say, once a year, once every six months, each of us sits down with our American agent, the individual agent to you, who quizzes you about everything. hey, Dan, how do you feel about immigration, about economy, about this, and manufacturing, and whatever, and taxes, and jobs?

Asks you everything, and it gets to know you. It gets to know you very well, what you care about. And by the way, there may be issues that you don't care about at all. You say, I don't care about that. This is cool. That's fine. Dan doesn't have really an opinion on that one. And then your agent and my agent and 350 million other agents converge, and you get this thing that actually represents all of us and what we actually want. I mean, just as a side note, one of the things that I looked up when I was preparing that talk was how many write-in votes happen. And there are millions of write-in votes every year. Mickey Mouse still is the winner, but Bart Simpson, Darth Vader, all this stuff. People are expressing their frustration when they do these write-in votes. But I think we would get a lot less frustration because when people talk with AI, they do so with much more genuineness and honesty than they do with their fellow human. There's been lots of studies on this so far. For example, if I'm debating something with a fellow human, I don't want to look stupid.

If I realize I'm getting cornered or whatever, maybe I make an excuse or whatever. I mean, everybody does this. But when you're talking with an AI and the AI brings something up, you say, oh, wow, I never thought about it that way. So we're much more likely to listen to AI and speak honestly with AI. We don't have to express frustration by doing silly things with AI. We can really have our agent get to know what our political opinions are. The way I think about this is I think we should have like a big meeting with our agent once every six months or a year. But you can always check in any time. Like if you have a conversation with someone, change your mind. You just come home and you tell your agent, hey, by the way, I changed my mind on this point. I feel differently about abortion than I did or gun control than I did. and great, and then it's representing that. And so now imagine you have this system that's not trying to get reelected and please donors and not, you know, do public mistakes or whatever, and it just actually implements some sort of utilitarianism, which is trying to seek the maximum good for the most people,

and you have a certain pedestal that you've put in place to make sure that it can't do any harm to people and there's certain thresholds at the bottom that it has to maintain. And after that, it just tries to maximize things for everyone. Over the last year, we've seen bots granted citizenship and various kinds of PR stunts, essentially. All relating to tricking humans into thinking that bots are... And those are kind of creepy and dystopian and just kind of silly. What I can tell of all the listeners who weren't in attendance is this was a thoughtful talk that wasn't about advocating rule of law to bots. It was really about thinking about what society expects from elected officials and just a well-researched, as far as I could tell also, fairly unrehearsed. You got up there with like a cell phone, and for 40 minutes, this was a very interesting talk.

It made me think a lot, and I wanted to personally say thank you, and I hope more people have the opportunity to hear you talk about this topic. Thank you. I'm actually going to do an episode of my Intercosmos podcast on this topic because I'm now obsessed with it. And so much would have to change. For example, constitutionally, it has to be a person born in the United States. I mean, what happens when it's not even a person anymore? So there's going to have to be all kinds of constitutional changes. Obviously, one of the big things is who controls this AI. My answer to that, I think, is it's got to be open source and transparent, meaning anybody can look at the weights, anybody can ask it any question they want, and we really have to have an emphasis on explainability. So it says, look, here's this piece of legislation that I'm putting into place, and here's why. Here's a bunch of graphs and charts. I've done 2 million simulations on this. This one ends up being optimal, keeping the economy healthy, while preserving security, while blah, blah, blah, blah, 50 other things,

and here's all the data. And that's so much more transparent and consistent with our values than what happens when we elect a human with an ego and an agenda and deep-pocketed lobbyist, right? I mean, there's so many reasons why it is not a sensationalist idea. Exactly. I mean, imagine that any citizen could ask Donald Trump, like, hey, why did you make that decision? and he gave a well-reasoned explanation for everything. I mean, first of all, he probably couldn't give a well-reasoned explanation for everything. I'm not saying anything about Trump except to say that he's a human. He's not going to have a well-reasoned explanation. There's probably lots of levels of reasons why he does stuff that maybe he doesn't even have access to. But, you know, also, what president has time to talk to anyone who wants to ask a question? It just seems so obvious to me that this is where we have to go. Now, a gentleman at the conference came up to me and said something so interesting that I hadn't thought of,

which is I suggested in my talk that this would be, you know, whatever, 20 years off. And he said, I think this could happen much sooner if you get someone like Taylor Swift to run for president and say, look, I don't know anything about politics, but I've got this great AI here, and I'm the face. Just elect me because I'm beautiful and popular, and I'll use this AI. And I thought that was a really great idea because that's how we can sneak this in much faster. The next version of this podcast, maybe it will be David and Dan in Taylor Swift headquarters, right? Facing signs or knocking on doors or something because that's very astute if you ever said that. When we were just getting to know each other before starting the recording, We were kind of going through kind of run a show stuff. I sent you a bunch of topics. And you said, I had this epiphany. I think I know why we dream. And you recently recorded an amazing conversation with Sam Harris.

So you pretended to ask my permission, but I was like, oh, my gosh. Like, yes. So great. Well, we were talking about talk topics. Yes. Explain. Okay, great. So about six years ago, I published a book called Live Wired, which is about brain plasticity. And one of the things that strikes me so amazing about the brain is that, for example, if you go blind, the other senses take over the visual part of your brain, the visual cortex in the back of your head. And so it's not really the visual cortex. It's just the part of the cortex that happens to be used for vision. As long as your eyes are working and if they're not working, the brain doesn't let any territory lie fallow. It takes it over. Okay. Now, the surprise was, to me, in 2013, some of my colleagues at Harvard did a study where they blindfolded people and put them in the brain scanner. And then they would make noises or touch the person on the hand, and they were looking at the signals. But to their surprise, what they found out is that after about 60 minutes,

the visual cortex would start to respond to sound and touch. Now, that was mind-blowing because how could this takeover start happening after just 60 minutes? And what I realized, my student and I were talking about this, and we realized that because our planet rotates into the darkness for half of the time, your visual system is at a unique disadvantage because in the dark, you can still hear and smell and taste and touch, but you can't see. And obviously, I'm talking about evolutionary time before we made electric lights. Okay, so, I mean, it was really dark, and you can't see, and that means that that vision is disadvantaged compared to the other senses, meaning all night long your visual system is at risk of getting taken over. You're essentially blind, I mean, especially when you sleep. There are many hypotheses for why we sleep, but one of them is that it's just to stay out of trouble. You curl up in the corner of a cave and you shut down and restore energy and stuff like that Anyway you know even when you asleep you can feel if something crawls on your skin you can hear a noise things like that but you can see

So what we realized was that dreams are the brain's way of defending the visual territory at nighttime. Every 90 minutes, if you look up the circuitry of dreaming, it's this very ancient circuit that just blasts random activity into the visual cortex and only the visual cortex every 90 minutes. So every night in a minute you just slam some random activity in there. And so it's essentially a screensaver, dreaming. And so we did quantitative studies on this. We compared 25 different species of primates, looked at how plastic their brains are, in other words, how flexible their brains are, and how much dream sleep they get, and it turns out it correlates perfectly, which is to say, like humans, we have very flexible brains and we do lots of dreaming to defend against this takeover. But other monkeys like lemurs, they have very not plastic brains. Their visual cortex is no danger of getting taken over. They have very little dream sleep. So anyway, that seems to be why we dream at night.

It's to defend the territory from takeover from the other senses. I didn't realize that the parts of the brain that are being unused, like by a sense that's disabled, it can be used by another sense. Yeah. So the whole, the full capacity of brain power, even if it's not consumed by all the senses, say equally. Exactly right. Yeah, the cortex, which is the wrinkly outer bit, that's essentially a one-trick pony. It's the same circuit repeated over and over and over and over the whole way across. And so even though we look at it and we say, oh, that's visual cortex, that's auditory cortex, So that's sensory cortex for your body. It just happens to be that based on the wiring that comes in. But if something breaks about that wiring, then it says, oh, that's cool. I'll use this territory for something else. I've read about studies that you've conducted where people who are deaf, maybe they have a heightened sense of touch, and so they can learn to quote here through haptic feedback.

That's exactly right. So what we did is we made a vest with vibratory motors on it that captures sound. And so for a deaf person, we're capturing sound, breaking it up into frequencies from low to high, and deaf people can learn how to understand the language of that and come to hear the world. We shrunk that down to a wristband. I actually ran a company on this for the last 10 years, a wristband with vibratory motors on it. It does the same thing. It's capturing sound, turning into patterns of vibration on the wrist, and deaf people can understand the world that way. because it doesn't matter how the information gets into the cortex as long as it gets there, the brain will figure out what to do with it. One of the things that's fascinating to me about brain research, this applies to a lot of areas of technology, but specifically to the brain, is that no matter how much we think we know, as a community it seems like we're well aware of how much we don't know. So maybe just as a thought experiment, I know I've got to get you off the hot seat and I'm going to respect your time, but you've got to answer this one last important question for me. when David and Dan are back here

and it's 2036 and not 2026 what is something that will be just foundational about our understanding of the brain that today just would seem like science fiction yeah I mean you want the whole list of the top ten here's what I think we're at the foot of the mountain in terms of understanding the brain what's funny is we have this book in the field called Principles of Neuroscience which is this big thick textbook about 880 pages and it's not I look at that and I think that's not principles if it were principles of neuroscience it would be like a pamphlet right but what we clearly are doing is just core dumping all this data that we are getting but we're not seeing the big picture yet and what I love so much about scientific revolutions is when suddenly there's a collapse of a bunch of data and we say oh, that's a thing. For example, imagine that you had a magical microscope

that would allow you to peer into a cell, all the complexity in the cell. And then, you know, you'd say, there's no way I'm ever going to figure this out. But then Crick and Watson come along and say, look, the important part is to keep the sequence of the base pairs, and all the rest is housekeeping. And you'd say, oh, got it. That's what we're waiting for in the brain. I don't mean we're waiting for it passively. We're actively trying to figure out such principles. And we've discovered a number of things, but boy, there is so much more to be discovered. So what kind of foundational things? I mean, the big central question is consciousness. Why does it feel like something to be alive? What's the difference when you are asleep and your brain is running along and doing all kinds of stuff, but then you're awake and it's running a slightly different algorithm and you're consciously aware? You have the impression that you are here and you see colors and you feel pain and all this stuff of being awake. Can science answer that? No, we don't know the answer to consciousness. We're totally, not only do we not have an answer to consciousness, we don't even know really what such an answer would look like in scientific language

because how do you say, look, do a double integral on this and carry the two and do this, and that equals the beauty of a sunset or the redness of red or the feeling of a pain in your knee. Like, how do you even translate those languages into one another? So it's very difficult, this problem. You know, myself and lots of colleagues are working on this, and I've been working on this for decades, but nobody's got an answer yet. Might we someday decode consciousness like we could decode DNA? Oh, I certainly hope so, yeah. Oh, I totally hope so. Otherwise, what are we doing? Yeah, so, and by the way, I think we've got to do that in the brain before we can answer this question that everyone's asking, which is, are LLMs conscious? Because who the hell knows? I mean, everyone's coming out with their impression. Oh, I think LMs are conscious. No way. They're not conscious. But who knows, right? What we need to do is understand what's happening biologically as in, ah, okay,

if you put together these recursive loops and you have this and that, that equals consciousness. Once we figure that out, then we can look at the AI and say, okay, God, that's got it. That doesn't have it. So that should be conscious. Yeah, exactly. We should be able to actually know instead of just give our impressions of whether we feel like it's conscious. Is there a two through ten? How in the world are there two through ten? Actually, now that's funny you would ask that because I wrote an article in Discover Magazine called Ten Unsolved Questions of Neuroscience. Now, here's the thing. I wrote that article in 2006, and nine and a half of them are still unanswered. One of the questions was why do brains sleep and dream? And I think I've answered the dreaming question now. But otherwise, they're still totally unanswered. One of them is cautious. One of them is, yeah. What is intelligence? Like, we don't even know. Everyone's got their way of describing what intelligence is. You know, maybe it's the capacity to simulate possible futures.

Maybe it's the capacity to squelch distractors. You know, there's all kinds of ways of thinking about intelligence, but it's not like as a field we have a definition that we agree on. So there's so many things. There's questions about emotions and, you know, what emotions are. We often tend to think about emotions as being one end of the spectrum and reason as being the other end of the spectrum or maybe two different categories. But it might be, yeah, it might be that they're the same thing, reason and emotion. It's just with kind of a wider camera lens, meaning if I'm, you know, trying to think of a chess move that I'm doing something very specific, But if I'm just experiencing like, oh, is this a good or a bad situation? Should I run or should I go towards that? Should I, you know, am I in danger or not? That's sort of a wide-angle lens where I'm taking lots of data to put that together. Anyway, all these things, there's a lot of unanswered questions in neuroscience. I'm so enthusiastic about AI for many reasons,

but perhaps the main reason is that it's going to help us speed up science. all the things that I and my students have been doing all the things that I and my students have been doing for decades and before me, you know, for centuries just takes a long time to get progress, to figure things out and the idea that AI is going to speed that up in a way such that when we're sitting here in 2036 we say, oh yeah, you know that list of 10 things I was talking about in 2026, they're all solved And thanks to AI, like that would be my dream. On behalf of humanity, it's so encouraging to see someone who's at the forefront of where these breakthroughs might come from be so enthusiastic. Not about the dehumanization that's going to take place because of AI, but because problems that have held us back as a species were right on the cusp of addressing. and I can't wait to have more versions of this conversation

and it's not going to take, I don't think, a decade to be talking about knocking off some of the nine and a half remaining things from your 2006 list. That's right. Let's do a mid-decade check-in on this. It's been such a pleasure. Likewise, Dan. Thanks. Just genuinely, really enjoyed it. Where can listeners learn more about you and some of this work that you do? Eagleman.com is my website. Brilliant. Well, gosh, that's an intense conversation. I'm glad we did have a face-to-face. It's all the time we have for this week on AI and the future of work. As always, I'm your host, Dan Turchin from PeopleRain. And, of course, we're back next week with another fascinating guest. Yo, it's the end of this show. Time to power it down. AI's the sense for humans with a crown. Team human in the house, got the vibe so clever. Dan Turchin from PeopleRain, making us better. We learned about AI for human We're ready to conquer the future That's all for now See you next time