Pod Insights Lex Fridman Podcast Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs About this episode Juergen Schmidhuber is the co-creator of long short-term memory networks (LSTMs) which are used in billions of devices today for speech recognition, translation, and much more. Over 30 years, he has proposed a lot of interesting, out-of-the-box ideas in artificial intelligence including a formal theory of creativity. Video version is available on YouTube https://www.youtube.com/watch?v=3FIo6evmweo. If you would like to get more information about this podcast go to https://lexfridman.com/ai or connect with @lexfridman on Twitter https://twitter.com/lexfridman, LinkedIn https://www.linkedin.com/in/lexfridman/, Facebook https://www.facebook.com/lexfridman, or YouTube https://www.youtube.com/lexfridman where you can watch the video versions of these conversations.
Transcript 10,708 words · 2 speakers
Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Lex Fridman? The following is a conversation with Jürgen Schmidhuber. He's the co-director of a CS Swiss AI lab and a co-creator of long, short-term memory networks. LSTMs are used in billions of devices today for speech recognition, translation, and much more. Over 30 years, he has proposed a lot of interesting, out-of-the-box ideas on meta-learning, adversarial networks, computer vision, and even a formal theory of, quote, "creativity, curiosity, and fun." This conversation is part of the MIT course on artificial general intelligence and the Artificial Intelligence podcast. If you enjoy it, subscribe on YouTube, iTunes, or simply connect with me on Twitter at
0:44 Lex Fridman? Lex Fridman, spelled F-R-I-D. And now, here's my conversation with Jürgen Schmidhuber.
1:09 Lex Fridman? Early on, you dreamed of AI systems that self-improve recursively. When was that dream born?
1:17 Speaker 01 When I was a baby. No, that's not true. When I was a teenager.
1:22 Lex Fridman? And what was the catalyst for that birth? What was the thing that first inspired you?
1:29 Speaker 01 When I was a boy, I'm...
1:33 Speaker 01 I was thinking about what to do in my life, and then I thought the most exciting thing is to solve the riddles of the universe. And that means you have to become a physicist. However, then I realized that there's something even grander. You can try to build a machine that isn't really a machine any longer, that learns to become a much better physicist than I could ever hope to be. And that's how I thought, maybe I can multiply my tiny little bit of creativity into infinity.
2:10 Lex Fridman? But ultimately, that creativity will be multiplied to understand the universe around us? That's the...
2:18 Lex Fridman? The curiosity for that mystery that drove you?
2:21 Speaker 01 Yes, so if you can build a machine that learns to solve more and more complex problems, a more and more general problem solver, then you basically have solved all the problems, at least all the solvable problems.
2:41 Lex Fridman? So how do you think, what does the mechanism for that kind of general solver look like? Obviously, we don't quite yet have one or know how to build one, but we have ideas. And you have had throughout your career several ideas about it. So how do you think about that mechanism?
2:59 Speaker 01 So in the eighties,
3:02 Speaker 01 I thought about how to build this machine that learns to solve all these problems that I cannot solve myself. And I thought it is clear, it has to be a machine that not only learns to solve this problem here and this problem here, but it also has to learn to improve the learning algorithm itself.
3:25 Speaker 01 So it has to have the learning algorithm in a representation that allows it to inspect it and modify it so that it can come up with a better learning algorithm. So I call that meta learning, learning to learn, and recursive self-improvement, that is really the pinnacle of that, where you then not only learn
3:50 Speaker 01 how to improve on that problem and on that, but you also improve the way the machine improves, and you also improve the way it improves, the way it improves itself. And that was my 1987 diploma thesis, which was all about that hierarchy of metal learners that.
4:11 Speaker 01 Have no computational limits except for the well-known limits that Gödel identified in 1931 and for the limits of physics.
4:22 Lex Fridman? In the recent years, meta-learning has gained popularity in a specific kind of form. You've talked about how that's not really meta-learning with neural networks, that's more basic transfer learning. Can you talk about the difference between the big, general meta-learning and a more narrow sense of meta-learning, the way it's used today, the way it's talked about today?
4:46 Speaker 01 Let's take the example of a deep neural network that has learned to classify images. And maybe you have trained that network on 100 different databases of images.
5:01 Speaker 01 And now a new database comes along and you want to quickly learn the new thing as well.
5:09 Speaker 01 So one simple way of doing that is you take the network, which already knows 100 types of databases, and then you just take the top layer of that and you retrain that using the new labeled data that you have in the new image database. And then it turns out that it really, really quickly can learn that too, one-shot basically, because from the first 100 data sets, it already has learned so much about computer vision that it can reuse that and that is then almost good enough to solve the new task, except you need a little bit of adjustment on the top.
5:54 Speaker 01 So that is transfer learning, and it has been done in principle for many decades. People have done similar things for decades.
6:04 Speaker 01 Meta learning, true meta learning is about having the learning algorithm itself open to introspection by the system that is using it. And also open to modification such that the learning system has an opportunity to modify any part of the learning algorithm and then evaluate the consequences of that modification and then learn from that to create a better learning algorithm and so on recursively. So that's a very different animal where you are opening the space of possible learning
6:48 Speaker 01 algorithms to the learning system itself.
6:51 Lex Fridman? Right. So you've, like in the 2004 paper, you described GEDOM machines, programs that rewrite themselves.
6:59 Speaker 01 Yeah.
7:00 Lex Fridman? Philosophically and even in your paper mathematically, these are really compelling ideas. But practically, do you see these self-referential programs being successful in the near term to having an impact where sort of it demonstrates to the world that this direction is a good one to pursue in the near term?
7:24 Speaker 01 Yes. We had these two different types of fundamental research, how to build a universal problem solver. One basically exploiting proof search and things like that that you need to come up with asymptotically optimal, theoretically optimal, self-improvers and problem solvers. However, one has to admit that through this proof search...
7:56 Speaker 01 Comes in an additive constant, an overhead, an additive overhead, that vanishes in comparison to what you have to do to solve large problems. However, for many of the small problems that we want to solve in our everyday life, we cannot ignore this constant overhead. And that's why we also have been doing other things, non-universal things, such as recurrent neural networks, which are trained by gradient descent and local search techniques, which aren't universal at all, which aren't provably optimal at all, like the other stuff that we did, but which are much more practical
8:41 Speaker 01 as long as we only want to solve the small problems that we are typically trying to solve in this environment here. Yeah, so the universal problem solvers, like the Gödel machine, but also Marcus Hutter's fastest way of solving all possible problems, which he developed around 2002 in my lab, they are associated with these constant overheads for proof search, which guarantees that the thing that you're doing is optimal. For example, there is this fastest way of solving all problems with a computable solution, which is due to Marcus, Marcus Hutter.
9:24 Speaker 01 And to explain what's going on there, let's take traveling salesman problems. With traveling salesman problems, you have a number of cities, N cities, and you try to find the shortest path through all these cities without visiting any city twice. And nobody knows the fastest way of solving traveling salesman problems, TSPs.
9:54 Speaker 01 But let's assume there is a method of solving them within N to the five operations, where N is the number of cities.
10:06 Speaker 01 Then the universal method of Markov's is going to solve the same traveling salesman problem also within N to the five steps. Plus, O of 1, plus a constant number of steps that you need for the proof searcher, which you need to show that this particular class of problems, the traveling salesman problems, can be solved within a certain time bound, within, or N to the five steps, basically. And this additive constant doesn't care for N, which means as N is getting larger and larger, as you have more and more cities,
10:50 Speaker 01 the constant overhead pales in comparison. And that means that almost all large problems are solved in the best possible way. Already today, we already have a universal problem solver like that. However, it's not practical because the overhead, the constant overhead, is so large that for the small kinds of problems that we want to solve in this little biosphere.
11:20 Lex Fridman? By the way, when you say "small", you're talking about things that fall within the constraints of our computational systems. They can seem quite large to us mere humans, right?
11:30 Speaker 01 That's right, yeah. So they seem large and even unsolvable in a practical sense today, but they are still small compared to almost all problems because almost all problems are large problems, which are much larger than any constant.
11:47 Lex Fridman? Do you find it useful, as a person who has dreamed of creating a general learning system, has worked on creating one, has done a lot of interesting ideas there, to think about P versus NP, this formalization of how hard problems are, how they scale, this kind of worst-case analysis type of thinking, do you find that useful? Or is it only just a mathematical, it's a set of mathematical techniques to give you intuition about what's good and bad?
12:21 Speaker 01 So, P versus NP, that's super interesting from a theoretical point of view. And in fact, as you are thinking about that problem, you can also get inspiration for better practical problem solvers. On the other hand, we have to admit that at the moment, the best practical problem solvers for all kinds of problems that we are now solving through what is called AI at the moment, they are not of the kind that is inspired by these questions. No. There we are using general-purpose computers such as recurrent neural networks, but we
13:01 Speaker 01 have a search technique, which is just local search, gradient descent, to try to find a program that is running on these recurrent networks such that it can solve some interesting problems, such as speech recognition, or machine translation, and something like that. And there is very little theory behind the best solutions that we have at the moment that can do that.
13:26 Lex Fridman? Do you think that needs to change? Do you think that will change? Or can we go, can we create a general intelligence systems without ever really proving that that system is intelligent in some kind of mathematical way? Solving machine translation perfectly, or something like that, within some kind of syntactic definition of a language? Or can we just be super impressed by the thing working extremely well, and that's sufficient?
13:50 Speaker 01 There's an old saying, and I don't know who brought it up first, which says, "There is nothing more practical than a good theory." And a good theory of problem solving...
14:08 Speaker 01 Under limited resources, like here on this universe or on this little planet, has to take into account these limited resources. And so probably there is lacking a theory which is related to what we already have, these asymptotically optimal problem solvers, which tells us what we need in addition to that to come up with a practically optimal problem solver. So I believe we will have something like that. And maybe just a few little tiny twists are necessary to change what we already have to come up with that as well. As long as we don't have that,
14:53 Speaker 01 we admit that we are taking suboptimal ways and we can use networks and long-term memory for equipped with local search techniques. And we are happy that it works better than any competing method. But that doesn't mean that we think we are done.
15:16 Lex Fridman? You've said that an AGI system will ultimately be a simple one. A general intelligence system will ultimately be a simple one. Maybe a pseudocode of a few lines will be able to describe it. Can you talk through your intuition behind this idea? Why you feel that at its core, intelligence is a simple algorithm?
15:42 Speaker 01 Experience tells us that the stuff that works best is really simple. So the asymptotically optimal ways of solving problems, if you look at them, are just a few lines of code. It's really true. Although they are these amazing properties, just a few lines of code. Then the most promising and most useful practical things.
16:09 Speaker 01 Maybe don't have this proof of optimality associated with them. However, they are also just a few lines of code. The most successful recurrential networks, you can write them down in five lines of pseudocode.
16:24 Lex Fridman? That's a beautiful, almost poetic idea. But what you're describing there is the lines of pseudocode are sitting on top of layers and layers of abstractions, in a sense. So you're saying at the very top, it'll be a beautifully written sort of algorithm. But do you think that there's many layers of abstractions we'll have to first learn to construct?
16:52 Speaker 01 Yeah, of course. We are building on all these great abstractions that people have invented over the millennia, such as matrix multiplications and real numbers and basic arithmetics and calculus and derivations of error functions and derivatives of error functions and stuff like that. So without that language that greatly simplifies our way of thinking about these problems, we couldn't do anything. So in that sense, as always, we are standing on the shoulders of the giants
17:35 Speaker 01 who, in the past, simplified the problem of problem solving so much that now we have a chance to do the final step.
17:45 Lex Fridman? So the final step will be a simple one. If we take a step back through all of human civilization and just the universe in general, how do you think about evolution? And what if creating a universe is required to achieve this final step? What if going through the very painful and inefficient process of evolution is needed to come up with this set of abstractions that ultimately lead to intelligence? Do you think there's a shortcut? Or do you think we have to create something like our universe in order to create something like human-level intelligence?
18:25 Speaker 01 So far, the only example we have is this one, this universe.
18:30 Lex Fridman? Do you think you can do better?
18:36 Speaker 01 Maybe not, but we are part of this whole process. So apparently, so it might be the case that the code that runs the universe is really, really simple. Everything points to that possibility because gravity and other basic forces are really simple laws that can be easily described also in just a few lines of code, basically. And then there are these other events, the apparently random events in the history of the universe, which as far as we know at the moment, don't have a compact code.
19:16 Speaker 01 But who knows, maybe somebody in the near future is going to figure out the pseudo-random generator, which is computing whether the measurement of that spin up or down thing here is going to be positive or negative.
19:34 Lex Fridman? Underlying quantum mechanics.
19:35 Speaker 01 Yes.
19:36 Lex Fridman? Do you ultimately think quantum mechanics is a pseudo-random number generator? So it's all deterministic. There's no randomness in our universe. Does God play dice?
19:47 Speaker 01 So a couple of years ago, a famous physicist, quantum physicist, Anton Zeilinger, he wrote an essay in Nature. And it started more or less like that. One of the fundamental insights of the 20th century was that the universe is fundamentally random on the quantum level. And that whenever you measure spin up or down or something like that, a new bit of information enters the history of the universe.
20:30 Speaker 01 And while I was reading that, I was already typing the response and they had to publish it because I was right. That there is no evidence, no physical evidence for that. So there's an alternative explanation where everything that we consider random is actually pseudo-random, such as the decimal expansion of pi, 3.141 and so on, which looks random, but isn't. So pi is interesting because every three-digit sequence, every sequence of three digits
21:09 Speaker 01 appears roughly 1 in 1,000 times and every five-digit sequence appears roughly 1 in 10,000 times, what you would expect if it was random. But there's a very short algorithm, a short program that computes all of that. So it's extremely compressible. And who knows, maybe tomorrow, somebody, some grad student at CERN goes back over all these data points, better decay and whatever, and figures out, oh, it's the second billion digits of pi, or something like that. We don't have any fundamental reason at the moment to believe that this is truly random and not just a deterministic video game.
21:54 Speaker 01 If it was a deterministic video game, it would be much more beautiful because beauty is simplicity. And many of the basic laws of the universe, like gravity and the other basic forces, are very simple. So very short programs can explain what these are doing. And it would be awful and ugly. The universe would be ugly, the history of the universe would be ugly if, for the extra things, the random, the seemingly random data points that we get all the time, that we really need a huge number of extra bits to describe all these, these extra bits of information.
22:40 Speaker 01 So as long as we don't have evidence that there is no short program that computes the entire history of the entire universe,
22:51 Speaker 01 we are, as scientists, compelled to look further for that shortest program.
22:59 Lex Fridman? Your intuition says there exists a shortest, a program that can backtrack to the creation of the universe. So compute the shortest path to the creation.
23:10 Speaker 01 Yes, including all the entanglement things and all the spin up and down measurements that have been taken place since 13.8 billion years ago. And so, yeah. So we don't have a proof that it is random. We don't have a proof that it is compressible to a short program. But as long as we don't have that proof, we are obliged, as scientists, to keep looking for that simple explanation.
23:43 Lex Fridman? Absolutely. So you said simplicity is beautiful or beauty is simple. Either one works. But you also work on curiosity, discovery. You know, the romantic notion of randomness, of serendipity, of being surprised by things that are about you. Kind of in our poetic notion of reality, we think as humans require randomness. So you don't find randomness beautiful. You find simple determinism beautiful.
24:20 Speaker 01 Yeah.
24:22 Lex Fridman? Okay.
24:23 Speaker 01 So why?
24:24 Lex Fridman? Why?
24:24 Speaker 01 Because the explanation becomes shorter. A universe that is compressible to a short program is much more elegant and much more beautiful than another one, which needs an almost infinite number of bits to be described. As far as we know.
24:47 Speaker 01 Many things that are happening in this universe are really simple in terms of short programs that compute gravity and the interaction between elementary particles and so on. So all of that seems to be very, very simple. Every electron seems to reuse the same sub-program all the time as it is interacting with other elementary particles.
25:14 Speaker 01 If we now required an extra oracle injecting new bits of information all the time for these extra things which are currently not understood, such as better decay, then the whole description length of the data that we can observe of the history of the universe would become much longer and therefore uglier.
25:49 Lex Fridman? And uglier. Again, simplicity is elegant and beautiful.
25:54 Speaker 01 All the history of science is a history of compression progress.
25:58 Lex Fridman? Yeah, so you've described sort of as we build up abstractions and you've talked about the idea of compression. How do you see this, the history of science, the history of humanity, our civilization, and life on Earth as some kind of path towards greater and greater compression? What do you mean by that? How do you think about that?
26:22 Speaker 01 Indeed, the history of science is a history of compression progress. What does that mean? Hundreds of years ago, there was an astronomer whose name was Kepler, and he looked at the data points that he got by watching planets move. And then he had all these data points, and suddenly it turned out that he can greatly compress the data by predicting it through an ellipse law. So it turns out that all these data points are more or less on ellipses around the sun.
27:01 Speaker 01 And another guy came along, whose name was Newton, and before him Hooke. And they said the same thing that is making these planets move like that is what makes the apples fall down. And it also holds for stones and for all kinds of other objects. And suddenly, many, many of these compressions, of these observations, became much more compressible. Because as long as you can predict the next thing, given what you have seen so far, you can compress it, but you don't have to store that data extra. This is called predictive coding.
27:43 Speaker 01 And then there was still something wrong with that theory of the universe, and you had deviations from these predictions of the theory. And 300 years later, another guy came along, whose name was Einstein. And he, he was able to explain away all these deviations from the predictions of the old theory through a new theory, which was called the General Theory of Relativity, which at first glance looks a little bit more complicated, and you have to warp space and time, but you can phrase it within one single sentence, which is, no matter how fast you accelerate
28:24 Speaker 01 and how fast or hard you decelerate, and no matter what is the gravity in your local framework, light speed always looks the same. And from, from that, you can calculate all the consequences. So it's a very simple thing, and it allows you to further compress all the observations because suddenly there are hardly any deviations any longer that you can measure from the predictions of this new theory. So art of science is a history of compression progress. You never arrive immediately at the shortest explanation of the data, but you're making progress.
29:08 Speaker 01 Whenever you are making progress, you have an insight. You see, oh, first I needed so many bits of information to describe the data, to describe my falling apples, my video of falling apples. I need so many data, so many pixels have to be stored. But then suddenly I realize, no, there is a very simple way of predicting the third frame in the video from the first two. And maybe not every little detail can be predicted, but more or less, most of these orange blobs that are coming down, they accelerate in the same way, which means that I can greatly compress the video. And the amount of compression progress, that is the depth of the insight that you have at that moment.
29:53 Speaker 01 That's the fun that you have, the scientific fun, the fun in that discovery. And we can build artificial systems that do the same thing. They measure the depth of their insights as they are looking at the data, which is coming in through their own experiments. And we give them a reward, an intrinsic reward in proportion to this depth of insight. And since they are trying to maximize the rewards they get, they are suddenly motivated to come up with new action sequences, with new experiments that have the property that the data that is coming in as a consequence of these experiments
30:36 Speaker 01 has the property that they can learn something about, see a pattern in there, which they hadn't seen yet before.
30:44 Lex Fridman? So there's an idea of power play that you've described, a training a general problem solver in this kind of way of looking for the unsolved problems.
30:53 Speaker 01 Yeah.
30:53 Lex Fridman? Can you describe that idea a little further?
30:56 Speaker 01 It's another very simple idea. So normally what you do in computer science, you have, you have...
31:04 Speaker 01 Some guy who gives you a problem. And then there is a huge search space of potential solution candidates. And you somehow try them out. And you have more or less sophisticated ways of moving around in that search space until you finally found a solution which you consider satisfactory. That's what most of computer science is about. PowerPlay just goes one little step further and says, let's not only search for solutions to a given problem, but let's search to pairs of problems and their solutions,
31:46 Speaker 01 where the system itself has the opportunity to phrase its own problem. So we are looking suddenly at pairs of problems and their solutions, or modifications of the problem solver that is supposed to generate a solution to that new problem. And this additional degree of freedom allows us to build curious systems that are like scientists in the sense that they not only try to solve, try to find answers to existing questions, no, they are also free to pose their own questions. So if you want to build an artificial scientist,
32:31 Speaker 01 we have to give it that freedom, and PowerPlay is exactly doing that.
32:35 Lex Fridman? So that's a dimension of freedom that's important to have. But how hard do you think that? How multi-dimensional and difficult the space of then coming up with your own questions is?
32:49 Speaker 01 Yeah.
32:50 Lex Fridman? So it's one of the things that as human beings, we consider to be the thing that makes us special, the intelligence that makes us special, is that brilliant insight.
33:00 Speaker 01 Yeah.
33:00 Lex Fridman? That can create something totally new.
33:03 Speaker 01 Yes. So now let's look at the extreme case. Let's look at the set of all possible problems that you can formally describe, which is infinite. Which should be the next problem that a scientist or a PowerPlay is going to solve? Well, it should be.
33:28 Speaker 01 The easiest problem that goes beyond what you already know.
33:33 Speaker 01 So it should be the simplest problem that the current problem solver that you have, which can already solve 100 problems, that he cannot solve yet by just generalizing.
33:47 Speaker 01 So it has to be new, so it has to require a modification of the problem solver, such that the new problem solver can solve this new thing, but the old problem solver cannot do it. And in addition to that, we have to make sure that the problem solver doesn't forget any of the previous solutions.
34:06 Lex Fridman? Right.
34:07 Speaker 01 And so by definition, PowerPlay is now trying always to search in this pair of, in the set of pairs of problems and problem solver modifications for a combination that minimizes the time to achieve these criteria. So it's always trying to find the problem which is easiest to add to the repertoire.
34:30 Lex Fridman? So just like grad students and academics and researchers can spend their whole career in a local minima,
34:38 Lex Fridman? stuck trying to come up with interesting questions, but ultimately doing very little. Do you think it's easy, in this approach of looking for the simplest unsolvable problem, to get stuck in a local minima? Is not never really discovering new, you know, really jumping outside of the 100 problems that you've already solved in a genuine, creative way?
35:03 Speaker 01 No, because that's the nature of PowerPlay, that it's always trying to break its current generalization abilities by coming up with a new problem which is beyond the current horizon. So just shifting the horizon of knowledge a little bit out there, breaking the existing rules, such that the new thing becomes solvable, but wasn't solvable by the old thing. So like adding a new axiom, like what Gödel did when he came up with these new sentences, new theorems that didn't have a proof in the formal system, which means you can add them to the repertoire, hoping that they are not going to damage the consistency of the whole thing.
35:51 Lex Fridman? So in the paper with the amazing title Formal Theory of Creativity, Fun and Intrinsic Motivation, you talk about discovery as intrinsic reward. So if you view humans as intelligent agents, what do you think is the purpose and meaning of life for us humans? You've talked about this discovery. Do you see humans as an instance of power play agents?
36:20 Speaker 01 Yeah, so humans are curious and that means they behave like scientists, not only the official scientists, but even the babies behave like scientists and they play around with their toys to figure out how the world works and how it is responding to their actions. And that's how they learn about gravity and everything. And yeah, in 1990, we had the first systems like that would just try to to play around with the environment and come up with situations that go beyond what they knew at that time and then get a reward for creating these situations and then becoming more general problem solvers and being able
37:02 Speaker 01 to understand more of the world. So yeah, I think in principle, that curiosity strategy or more sophisticated versions of what I just described, they are what we have built in as well because evolution discovered that's a good way of exploring the unknown world. And a guy who explores the unknown world has a higher chance of solving problems that he needs to survive in this world. On the other hand, those guys who were too curious, they were weeded out as well. So you have to find this trade off. Evolution found a
37:43 Speaker 01 certain trade off. Apparently in our society, there is a certain percentage of extremely explorative guys and it doesn't matter if they die because many of the others are more conservative. And so, yeah, it would be surprising to me if that principle of artificial curiosity.
38:10 Speaker 01 Wouldn't be present in almost exactly the same form here.
38:14 Lex Fridman? In our brains. So you're a bit of a musician and an artist. So continuing on this topic of creativity. What do you think is the role of creativity in intelligence? So you've kind of implied that it's essential for intelligence, if you think of intelligence as a problem solving system. As ability to solve problems, but do you think it's essential? This idea of creativity?
38:44 Speaker 01 We never have a program, a sub-program that is called creativity or something. It's just a side effect of what our problem solvers do. They are searching a space of problems, a space of candidates, of solution candidates, until they hopefully find a solution to a given problem. But then there are these two types of creativity and both of them are now present in our machines. The first one has been around for a long time, which is human gives problem to machine, machine tries to find a solution to that. And this has been happening for many decades and for many decades, machines have found
39:24 Speaker 01 creative solutions to interesting problems where humans were not aware of these particularly creative solutions, but then appreciated that the machine found that. The second is the pure creativity. That I would call, what I just mentioned, I would call the applied creativity. Like applied art, where somebody tells you, now make a nice picture of of this pope and you will get money for that. Okay, so here is the artist and he makes a convincing picture of the pope and the pope likes it and gives him the money.
40:01 Speaker 01 And then there is the pure creativity, which is more like the power play and the artificial curiosity thing, where you have the freedom to select your own problem. Like a scientist who defines his own question to study. And so that is the pure creativity, if you will. As opposed to the applied creativity, which serves another.
40:30 Lex Fridman? And in that distinction, there's almost echoes of narrow AI versus general AI. So this kind of constrained painting of a pope seems like
40:41 Lex Fridman? the approaches of what people are calling narrow AI. And pure creativity seems to be,
40:49 Lex Fridman? maybe I'm just biased as a human, but it seems to be an essential element of human level intelligence. Is that what you're implying?
41:00 Lex Fridman? To a degree.
41:01 Speaker 01 If you zoom back a little bit and you just look at a general problem solving machine, which is trying to solve arbitrary problems, then this machine will figure out in the course of solving problems that it's good to be curious. So all of what I said just now about this pre-wire curiosity and this will to invent new problems that the system doesn't know how to solve yet should be just a byproduct of the general search. However,
41:33 Speaker 01 apparently, evolution has built it into us. Because it turned out to be so successful, a pre-wiring, a bias, a very successful exploratory bias that that we are born with.
41:49 Lex Fridman? And you've also said that consciousness, in the same kind of way, may be a byproduct of problem solving.
41:56 Speaker 01 Yeah.
41:57 Lex Fridman? Do you think, do you find this an interesting byproduct? Do you think it's a useful byproduct? What are your thoughts on consciousness in general? Or is it simply a byproduct of greater and greater capabilities of problem solving? That's that's similar to creativity in that sense.
42:16 Speaker 01 Yeah, we never have a procedure called consciousness in our machines. However, we get as side effects of what these machines are doing, things that seem to be closely related to what people call consciousness. So for example, already in nineteen ninety, we had simple systems, which were basically recurrent networks and therefore universal computers, trying to map incoming data into actions that lead to success. Maximizing reward in a given environment, always finding the charging station in time. Whenever the battery is low and negative signals are coming from the battery,
43:00 Speaker 01 always find the charging station in time, without bumping against painful obstacles on the way. So complicated things, but very easily motivated. And then we give these little guys
43:15 Speaker 01 a separate recurrent neural network, which is just predicting what's happening if I do that and that. What will happen as a consequence of these actions that I'm executing? And it's just trained on the long and long history of interactions with the world. So it becomes a predictive model of the world, basically. And therefore also a compressor of the observations of the world, because whatever you can predict, you don't have to store extra. So compression is a side effect of prediction. And how does this recurrent network compress? Well, it's inventing little subprograms, little subnetworks, that stand for everything that frequently appears in the environment,
43:57 Speaker 01 like bottles and microphones and faces, maybe lots of faces in my environment. So I'm learning to create something like a prototype face, and a new face comes along and all I have to encode are the deviations from the prototype. So it's compressing all the time the stuff that frequently appears. There's one thing that appears all the time, that is present all the time when the agent is interacting with its environment, which is the agent itself. So just for data compression reasons, it is extremely natural for this Recur network to come up with little sub-networks that stand for the properties of the agents,
44:39 Speaker 01 the hand, the other actuators, and all the stuff that you need to better encode the data which is influenced by the actions of the agent. So there, just as a side effect of
44:54 Speaker 01 data compression during problem solving, you have internal self-models. Now you can use this model of the world to plan your future. And that's what yours have done since 1990. So the Recur network, which is the controller, which is trying to maximize reward, can use this model of the network, of the world, this model network of the world, this predictive model of the world, to plan ahead and say, "Let's not do this action sequence, let's do this action sequence instead, because it leads to more predicted reward." And whenever it's waking up, these little sub-networks that stand for itself, then it's thinking about itself.
45:38 Speaker 01 Then it's thinking about itself and it's exploring mentally the consequences of its own actions. And now you tell me why it's still missing.
45:53 Lex Fridman? Missing the next, the gap to consciousness.
45:55 Speaker 01 Yeah.
45:56 Lex Fridman? There isn't. That's a really beautiful idea that, you know, if life is a collection of data and life is a process of compressing that data to act efficiently,
46:10 Lex Fridman? in that data, you yourself appear very often. So it's useful to form compressions of yourself. And it's a really beautiful formulation of what consciousness is, as a necessary side effect. It's actually quite...
46:25 Lex Fridman? Compelling to me. You've described RNNs, developed LSTMs, long short term memory networks. They're a type of recurrent neural networks. They have gotten a lot of success recently. So these are networks that model the temporal aspects in the data, temporal patterns in the data. And you've called them the deepest of the neural networks, right? So what do you think is the value of depth in the models that we use to learn?
46:59 Speaker 01 Since you mentioned the long short term memory and the LSTM, I have to mention the names of the brilliant students who made that possible. First of all, my first student ever, Sepp Hochreiter, who had fundamental insights already in his diploma thesis. Then Felix Gers, who had additional important contributions. Alex Gray, the guy from Scotland, who is mostly responsible for this CTC algorithm, which is now often used to train the LSTM to do the speech recognition on all the Google Android phones and whatever, and CAV, and so on. So these guys, without these guys, I would be nothing.
47:42 Lex Fridman? It's a lot of incredible work.
47:45 Speaker 01 What is now the depth? What is the importance of depth? Well, most problems in the real world are deep in the sense that the current input doesn't tell you all you need to know about the environment. So instead, you have to have a memory of what happened in the past. And often, important parts of that memory are dated. They are pretty old. So when you're doing speech recognition, for example, and somebody says 11, then that's about half a second or something like that, which means it's already 50 time steps.
48:27 Speaker 01 And another guy, or the same guy, says 7. So the ending is the same, 11. But now the system has to see the distinction between 7 and 11. And the only way it can see the difference is it has to store that 50 steps ago, there was an s or an l, 11 or 7. So there you have already a problem of depth fifty. Because for each time step, you have something like a virtual layer in the expanded, unrolled version of this recurrent network, which is doing the speech recognition. So these long time lags, they translate into problem depth.
49:09 Speaker 01 And most problems in this world are such that you really have to look far back in time to understand what is the problem and to solve it.
49:22 Lex Fridman? But just like with LSTMs, you don't necessarily need to, when you look back in time, remember every aspect. You just need to remember the important aspects.
49:30 Speaker 01 That's right. The network has to learn to put the important stuff into memory and to ignore the unimportant noise.
49:39 Lex Fridman? So, but in that sense, deeper and deeper is better? Or is there a limitation? Is there... I mean, LSTM is one of the great examples of architectures that do something beyond just deeper and deeper networks. There's clever mechanisms for filtering data, for remembering and forgetting. So do you think that kind of thinking is necessary? If you think about LSTMs as a leap, a big leap forward over traditional vanilla RNNs, what do you think is the next leap? Within this context? So LSTM is a very clever improvement, but LSTM still don't
50:24 Lex Fridman? have the same kind of ability to see far back in the future, in the past, as us humans do. The credit assignment problem across way back, not just fifty time steps or a hundred or a thousand, but millions and billions.
50:40 Speaker 01 It's not clear what are the practical limits of the LSTM when it comes to looking back. Already in 2006, I think, we had examples where it not only looked back tens of thousands of steps, but really millions of steps. And Juan Perez Ortiz, in my lab, I think was the first author of a paper where we really, was it 2006 or something? Had examples where it learned to look back for more than 10 million steps. So... For most problems of speech recognition, it's not necessary to look that far back. But there are examples where it does. Now, the looking back thing,
51:25 Speaker 01 that's rather easy, because there is only one past. But there are many possible futures. And so a reinforcement learning system, which is trying to maximize its future expected reward and doesn't know yet which of these many possible futures should I select, given this one single past, is facing problems that the LSCM by itself cannot solve. So the LSCM is good for coming up with a compact representation of the history so far, of the history and of observations and actions so far. But now, how do you plan in an efficient and good way among all these? How do you select
52:10 Speaker 01 one of these many possible action sequences that a reinforcement learning system has to consider to maximize reward in this unknown future? So again, we have this basic setup where you have OneViconetwork, which gets in the video and the speech and whatever, and is executing actions and is trying to maximize reward. So there is no teacher who tells it what to do at which point in time. And then there's the other network, which is just predicting what's going to happen if I do that and then. And that could be an LSCM network. And it allows to look back all the way
52:54 Speaker 01 to make better predictions of the next time step. So essentially, although it's predicting only the next time step, it is motivated to learn to put into memory something that happened maybe a million steps ago, because it's important to memorize that if you want to predict that at the next time step, the next event, you know? Now, how can a model of the world like that, a predictive model of the world, be used by the first guy? Let's call it the controller and the model. The controller and the model. How can the model be used by the controller to efficiently select among these many possible futures? The naive way we had about 30 years ago
53:38 Speaker 01 was let's just use the model of the world as a stand-in, as a simulation of the world, and millisecond by millisecond we plan the future. And that means we have to roll it out really in detail, and it will work only if the model is really good, and it will still be inefficient because we have to look at all these possible futures, and there are so many of them. So instead, what we do now, since 2015, in our CN systems, controller model systems, we give the controller the opportunity to learn by itself how to use the potentially relevant parts of the M, of the model network, to solve new problems more quickly. And if it wants to,
54:23 Speaker 01 it can learn to ignore the M. And sometimes it's a good idea to ignore the M because it's really bad. It's a bad predictor in this particular situation of life where the controller is currently trying to maximize reward. However, it can also learn to address and exploit some of the subprograms that came about in the model network through compressing the data by predicting it. So it now has an opportunity to reuse that code, the algorithmic information in the model network, to reduce its own search space such that it can solve a new problem more quickly than without the model.
55:09 Lex Fridman? Compression. So you're ultimately optimistic and excited about the power of R, of reinforcement learning in the context of real systems?
55:21 Speaker 01 Absolutely, yeah.
55:22 Lex Fridman? So you see RL as a potential, having a huge impact beyond just sort of the M part is often developed on supervised learning methods. You see RL as a, for problems of self-driving cars or any kind of applied side robotics. Is that's the correct, interesting direction for research in your view?
55:50 Speaker 01 I do think so. We have a company called NaSense.
55:53 Lex Fridman? NaSense.
55:53 Speaker 01 Which has applied reinforcement learning to little Audis.
55:59 Lex Fridman? Little Audis.
56:00 Speaker 01 Which learn to park without a teacher. The same principles were used, of course. So these little Audis, they are small, maybe like that, so much smaller than the real Audis, but they have all the sensors that you find in the real Audis. You find the cameras, the LiDAR sensors. They go up to 120, 20 kilometers an hour if you, if they want to. And, and they have pain sensors, basically, and they don't want to bump against obstacles and other Audis. And so they must learn like little babies to park. Take the raw vision input and translate that into actions that lead to successful parking
56:44 Speaker 01 behavior, which is a rewarding thing. And yes, they learn that.
56:48 Lex Fridman? They learn.
56:48 Speaker 01 So we have examples like that, and it's only in the beginning. This is just the tip of the iceberg, and I believe the next wave of AI is going to be all about that. So at the moment, the current wave of AI is about passive pattern observation and prediction. And, and that's what you have on your smartphone and what the major companies on the Pacific Rim are using to sell you ads, to do marketing. That's the current source of profit in AI. And that's only 1% or 2% of the world economy, which is big enough to make these companies pretty much the most valuable companies in the world, but there's a much, much bigger.
57:34 Speaker 01 A fraction of the economy going to be affected by the next wave, which is really about machines that shape the data through their own actions.
57:44 Lex Fridman? Do you think simulation is ultimately the biggest way that those methods will be successful in the next 10, 20 years? We're not talking about 100 years from now. We're talking about sort of the near term impact of RL. Do you think really good simulation is required? Or is there other techniques like imitation learning, observing other humans operating in the real world? Where do you think this success will come from?
58:13 Speaker 01 So at the moment, we have a tendency of using
58:17 Speaker 01 physics simulations to learn behavior for machines that learn to solve problems that humans also do not know how to solve. However, this is not the future because the future is in what little babies do. They don't use a physics engine to simulate the world. No, they learn a predictive model of the world, which maybe sometimes is wrong in many ways, but captures all kinds of important abstract, high level predictions, which are really important to be successful. And that's what was the future 30 years ago when you started that type of research. But it's still the future.
59:01 Speaker 01 And now we know much better how to go there, to move there, to move forward and to really make working systems based on that, where you have a learning model of the world, a model of the world that learns to predict what's going to happen if I do that and that. And then the controller uses that model to more quickly learn successful action sequences.
59:27 Speaker 01 And then, of course, always this curiosity thing. At the beginning, the model is stupid. So the controller should be motivated to come up with experiments with action sequences that lead to data that improve the model.
59:40 Lex Fridman? Do you think improving the model, constructing an understanding of the world in this connection is the, now, the popular approaches have been successful or, you know, grounded in ideas of neural networks. But in the eighties with expert systems, there's symbolic AI approaches, which, to us humans, are more intuitive, in the sense that it makes sense that you build up knowledge and this knowledge representation. What kind of lessons can we draw into our current approaches in, for, from expert systems, from symbolic AI?
1:00:16 Speaker 01 So I became aware of all of that in the eighties. And back then, logic program, logic programming was a huge thing.
1:00:24 Lex Fridman? Was it inspiring to you yourself? Did you find it compelling? Because most, a lot of your work was not so much in that realm, right? It was more in the learning systems.
1:00:34 Speaker 01 Yes and no, but we did all of that. So we, my first publication ever, actually, was nineteen eighty-seven, was the implementation of genetic algorithm of a genetic programming system in Prolog. So Prolog, that's what you learned back then, which is a logic programming language. And the Japanese, they had this huge fifth generation AI project, which was mostly about logic programming back then, although neural networks existed and were well-known back then. And deep learning has existed since nineteen sixty-five, since this guy in the Ukraine, Ivanenko, started it.
1:01:18 Speaker 01 But the Japanese and many other people, they focused really on this logic programming. And I was influenced to the extent that I said, okay, let's take these biologically inspired algorithms like evolution, our programs, and, and...
1:01:35 Speaker 01 And implement that in the language which I know, which was Prolog, for example, back then. And then, in many ways, this came back later because the Goethe machine, for example, has a proof searcher on board, and without that, it would not be optimal. Or Marcus Hutter's universal algorithm for solving all well-defined problems has a proof searcher on board. So that's very much logic programming. Without that, it would not be asymptically optimal. But then, on the other hand, because we are very pragmatic guys also, we focused on recurrent networks and suboptimal stuff,
1:02:18 Speaker 01 such as gradient-based search and program space, rather than provably optimal things.
1:02:25 Lex Fridman? The logic programming certainly has a usefulness in when you're trying to construct something provably optimal or provably good or something like that. But is it useful for practical problems?
1:02:37 Speaker 01 It's really useful for our theorem proving. The best theorem improvers today are not neural networks.
1:02:43 Lex Fridman? Right.
1:02:43 Speaker 01 No, they are logic programming systems, and they are much better theorem improvers than most math students in the first or second semester.
1:02:54 Lex Fridman? But for reasoning, for playing games of Go or chess, or for robots, autonomous vehicles that operate in the real world, or object manipulation?
1:03:05 Speaker 01 Yeah.
1:03:05 Lex Fridman? You think learning...
1:03:07 Speaker 01 Yeah, as long as the problems have little to do with theorem improving themselves, then as long as that's not the case, you would just want to have better pattern recognition. So to build a self-driving car, you want to have better pattern recognition and pedestrian recognition and all these things. And you want to minimize the number of false positives, which is currently slowing down self-driving cars in many ways. And all of that has very little to do with logic programming, yeah.
1:03:43 Lex Fridman? What are you most excited about in terms of directions of artificial intelligence at this moment in the next few years, in your own research and in the broader community?
1:03:57 Speaker 01 So I think in the not-so-distant future, we will have, for the first time...
1:04:05 Speaker 01 Little robots that learn like kids. And I will be able to say to the robot, "Look here, robot, we are going to assemble a smartphone. Let's take this slab of plastic and the screwdriver and let's screw in the screw like that." No, not like that, like that. Not like that, like that. And I don't have a data glove or something. He will see me and he will hear me and he will try to do something with his own actuators, which will be really different from mine, but he will understand the difference and will learn to imitate me, but not in the supervised way,
1:04:50 Speaker 01 where a teacher is giving target signals for all his muscles all the time. No, by doing this high-level imitation, where he first has to learn to imitate me and then to interpret these additional noises coming from my mouth as helpful signals to do that pattern. And then it will, by itself, come up with faster ways and more efficient ways of doing the same thing. And finally, I stop his learning algorithm and make a million copies and sell it. And so, at the moment, this is not possible, but we already see how we are going to get there. And you can imagine, to the extent
1:05:35 Speaker 01 that this works economically and cheaply, it's going to change everything. Almost all our production is going to be affected by that. And a much bigger wave, a much bigger AI wave is coming than the one that we are currently witnessing, which is mostly about passive pattern recognition on your smartphone. This is about active machines that shapes data through the actions they are executing, and they learn to do that in a good way.
1:06:08 Speaker 01 So many of the traditional industries are going to be affected by that. All the companies that are building machines will equip these machines with cameras and other sensors. And they are going to learn to solve all kinds of problems through interaction with humans, but also a lot on their own to improve what they already can do.
1:06:36 Speaker 01 And lots of old economy is going to be affected by that. And in recent years, I have seen that old economy is actually waking up and realizing that this is the case.
1:06:47 Lex Fridman? Are you optimistic about that future? Are you concerned? There's a lot of people concerned in the near term about the transformation of the nature of work. The kind of ideas that you just suggested would have a significant impact of what kind of things could be automated. Are you optimistic about that future? Are you nervous about that future? And looking a little bit farther into the future, there's people like Elon Musk, Stuart Russell, concerned about the existential threats of that future. So in the near term, job loss, in the long term, existential threat. Are these concerns to you or are you ultimately optimistic?
1:07:31 Speaker 01 So let's first address the near future.
1:07:38 Speaker 01 We have had predictions of job losses for many decades. For example, when industrial robots came along, many people predicted that lots of jobs are going to get lost. And in a sense, they were right. Because back then, there were car factories and hundreds of people. And these factories assembled cars. And today, the same car factories have hundreds of robots and maybe three guys watching the robots.
1:08:14 Speaker 01 On the other hand, those countries that have lots of robots per capita, Japan, Korea, Germany, Switzerland, a couple of other countries, they have really low unemployment rates. Somehow, all kinds of new jobs were created. Back then, nobody anticipated those jobs. And decades ago, I always said, it's really easy to say which jobs are going to get last, but it's really hard to predict the new ones. Thirty years ago, who would have predicted all these people making money as YouTube bloggers,
1:08:58 Speaker 01 for example? Two hundred years ago, sixty percent of all people used to work in agriculture. Today, maybe one percent. But still, only, I don't know, five percent unemployment. Lots of new jobs were created, and Homo ludens, the playing man, is inventing new jobs all the time. Most of these jobs are not existentially necessary for the survival of our species. There are only very few existentially necessary jobs, such as farming, and building houses, and warming up the houses,
1:09:43 Speaker 01 but less than ten percent of the population is doing that. And most of these newly invented jobs are about interacting with other people in new ways, through new media and so on, getting new types of kudos, in forms of likes and whatever, and even making money through that. So Homo ludens, the playing man, doesn't want to be unemployed, and that's why he's inventing new jobs all the time. And... He keeps considering these jobs as really important and is investing a lot of energy and hours of work into into those new jobs.
1:10:24 Lex Fridman? That's quite beautifully put. We're really nervous about the future because we can't predict what kind of new jobs will be created. But you're ultimately optimistic that we humans are so restless that we create and give meaning to newer and newer jobs, totally new likes on Facebook, things that get likes on Facebook, or whatever the social platform is. So what about long term existential threat of AI where our whole civilization may be swallowed up by this ultra super intelligent systems?
1:11:01 Speaker 01 Maybe it's not going to be swallowed up, but I'd be surprised if we were, we humans were the last step in the evolution of the universe.
1:11:15 Lex Fridman? And you've actually had this beautiful comment somewhere that I've seen saying that artificial, quite insightful, artificial general intelligence systems, just like us humans, will likely not want to interact with humans. They'll just interact amongst themselves, just like ants interact amongst themselves and only tangentially interact with humans. And it's quite an interesting idea that once we create AGI, they will lose interest in humans and have, compete for their own Facebook likes and their own social platforms. So within that quite elegant idea, how do we know in a hypothetical sense
1:12:00 Lex Fridman? that there's not already intelligent systems out there? How do you think broadly of general intelligence greater than us? How would we know it's out there? How would we know it's around us? And could it already be?
1:12:17 Speaker 01 I'd be surprised if within the next few decades or something like that, we won't have AIs that are truly smart in every single way and better problem solvers in almost every single important way.
1:12:34 Speaker 01 And I'd be surprised if they wouldn't realize what we have realized a long time ago, which is that almost all physical resources are not here in this biosphere. But further out, the rest of the solar system gets two billion times more solar energy than our little planet. There's lots of material out there that you can use to build robots and self-replicating robot factories and all this stuff. And they are going to do that. And they will be scientists and curious, and they will explore what they can do.
1:13:15 Speaker 01 And in the beginning, they will be fascinated by life and by their own origins in our civilization. They will want to understand that completely, just like people today would like to understand how life works and also the history of our own existence and civilization, but then also the physical laws that created all of that. So they, in the beginning, they will be fascinated by life. Once they understand it, they lose interest, like anybody who loses interest in things he understands.
1:13:55 Speaker 01 And then, as you said, the most interesting sources of information for them will be others of their own kind. Mm-hmm.
1:14:14 Speaker 01 So, at least in the long run, there seems to be some sort of protection through lack of interest on the other side.
1:14:27 Speaker 01 And now, it seems also clear, as far as we understand physics, you need matter and energy to compute and to build more robots and infrastructure and more AI civilization and EII ecologies consisting of trillions of different types of AIs. And so, it seems inconceivable to me that this thing is not going to expand. Some AI ecology not controlled by one AI, but by trillions of different types of AIs competing in all kinds of quickly evolving and disappearing ecological niches in ways that we cannot fathom at the moment.
1:15:08 Speaker 01 But it's going to expand, limited by light speed and physics, but it's going to expand. And now, we realize that the universe is still young. It's only 13.8 billion years old, and it's going to be 1,000 times older than that.
1:15:26 Speaker 01 So, there's plenty of time to conquer the entire universe and to fill it with intelligence and senders and receivers such that AIs can travel the way they are traveling in our labs today, which is by radio from sender to receiver.
1:15:47 Speaker 01 And let's call the current age of the universe one eon, one eon.
1:15:55 Speaker 01 Now, it will take just a few eons from now, and the entire visible universe is going to be full of that stuff.
1:16:03 Speaker 01 And let's look ahead to a time when the universe is going to be 1,000 times older than it is now. They will look back and they will say, "Look," almost immediately after the Big Bang, only a few eons later, the entire universe started to become intelligent. Now, to your question. How do we see whether anything like that has already happened or is already in a more advanced stage in some other part of the universe, of the visible universe? We are trying to look out there and nothing like that has happened so far. Or is that true?
1:16:38 Lex Fridman? Do you think we would recognize it? Well, how do we know it's not among us? How do we know planets aren't in themselves intelligent beings? How do we know ants seen as a collective are not much greater intelligence than our own? These kinds of ideas.
1:16:57 Speaker 01 Yeah. When I was a boy, I was thinking about these things. And I thought, hmm, maybe it has already happened. Because back then, I know, I knew, I learned from popular physics books that the structure, the large scale structure of the universe is not homogeneous. And you have these clusters of galaxies. And then in between, there are these huge empty spaces. And I thought, hmm, maybe they aren't really empty. It's just that in the middle of that, some AI civilization already has expanded and then has covered a bubble of a billion light years diameter and is using all the energy of all the stars within that bubble
1:17:42 Speaker 01 for its own unfathomable purposes. And so it always happened and we just failed to interpret the signs. But then I learned that gravity by itself explains the large scale of structure of the universe and that this is not a convincing explanation. And then I thought, maybe, maybe it's the dark matter. Because as far as we know today, eighty percent of the measurable matter is invisible. And we know that because otherwise our galaxy or other galaxies would fall apart. They would, they are rotating too quickly.
1:18:25 Speaker 01 And then the idea was maybe all of these AI civilizations that are already out there, they, they, they are just invisible because they're really efficient in using the energies of their own local systems, and that's why they appear dark to us. But this is also not a convincing explanation, because then the question becomes, why is there, are there still any visible stars left in our own galaxy, which also must have a lot of dark matter? So that is also not a convincing thing.
1:19:02 Speaker 01 And today, I like to think it's quite plausible that maybe we are the first, at least in our local light cone, within the few hundreds of millions of light years that we can reliably observe.
1:19:24 Lex Fridman? Is that exciting to you, that we might be the first?
1:19:27 Speaker 01 And it would make us much more important, because if we mess it up through a nuclear war, then, then maybe this will have an effect on the, on the, on the development of the entire universe.
1:19:46 Lex Fridman? So let's not mess it up.
1:19:48 Speaker 01 Let's not mess it up.
1:19:49 Lex Fridman? Jürgen, thank you so much for talking today. I really appreciate it.
1:19:52 Speaker 01 It's my pleasure.
Previous Pieter Abbeel: Deep Reinforcement Learning Next Tuomas Sandholm: Poker and Game Theory Page updated October 26, 2026 · transcript and summary are machine-generated
