Daniel Kahneman: Thinking Fast and Slow, Deep Learning, and AI | Lex Fridman Podcast #65

Daniel Kahneman 0:00 The following is a conversation with Daniel Kahneman, winner of the Nobel Prize in Economics for his integration of economic science with the psychology of human behavior, judgment, and decision-making. He's the author of the popular book, Thinking, Fast and Slow, that summarizes in an accessible way his research of several decades, often in collaboration with Amos Tversky......

Daniel Kahneman 6:03 and what, what people can do.

Lex Fridman 6:06 So, the effect of the in group and the out group?

Daniel Kahneman 6:09 You know, the- it's clear that those were people, you know, you could- you could shoot them. You could- you know, they were not human. They were not- there was no empathy or very, very little empathy left. So occasionally, you know, they might have been- and, and very quickly, by the way, uh, the empathy disappeared if there was initially. And the fact that everybody around you was doing it, that, that completely- the group doing it and everybody shooting Jews, I think that, that, uh, makes it permissible. Now, how much, you know, whether it would- it could happen, uh, in every culture or whether the Germans were just particularly efficient and, and disciplined so they could get away with it.

Lex Fridman 7:06 Mm-hmm.

Daniel Kahneman 7:08 That's-

Lex Fridman 7:08 It's a question.

Daniel Kahneman 7:09 It's an interesting question.

Lex Fridman 7:10 Are these artifacts of history or is it human nature?

Daniel Kahneman 7:14 I think that's really human nature. You know, you put some people in a position of power relative to other people and, and then they become less human. They- they become different.

Lex Fridman 7:28 But in general, in war, outside of concentration camps in World War II, it seems that war brings out darker sides of human nature, but also the beautiful things about human nature.

Daniel Kahneman 7:41 Well, you know, I mean, what it- what it brings out is the, the loyalty among soldiers. I mean, it brings out the bonding. Male bonding, I think, is a very real thing that- and that happens. And so- and, and there is a certain thrill to friendship and there is certainly a certain thrill to friendship under risk-

Lex Fridman 8:04 Yeah.

Daniel Kahneman 8:04 ... and to shared risk. And so people have very profound emotions up to the point where it gets so traumatic that, uh, that little is left, but-

Lex Fridman 8:18 So,

Lex Fridman 8:20 let's talk about psychology a little bit. Uh, in your book, Thinking, Fast and Slow, you describe two modes of thought system. One, the fast, instinctive and emotional one, and system two, the slower, deliberate, logical one. At the risk of asking Darwin to discuss (laughs) , uh, theory of evolution, uh, can you describe distinguishing characteristics for people who have not read your book of the two systems?

Daniel Kahneman 8:49 Well, I mean, the word system is a bit misleading, but it's- at the same time it's misleading, it's also very useful.

Lex Fridman 8:58 Yes.

Daniel Kahneman 8:58 But what I call system one, it's easier to think of it as, as a family of activities. And primarily the way I describe it is there are different ways for ideas to come to mind, and some ideas come to mind automatically. And the example- a standard example is two plus two, and then something happens to you. And, and in other cases you've got to do something, you've got to work in order to produce the idea. And my example, I always give the same pair of numbers as 27 times 14, I think.

Lex Fridman 9:36 You have to d- perform some algorithm in your head, some steps.

Daniel Kahneman 9:39 Yes, and, and it takes time.

Lex Fridman 9:41 Yeah.

Daniel Kahneman 9:42 It's a very different. Nothing comes to mind except something comes to mind which is the algorithm, I mean, that you've got to perform, and then it's work-

Lex Fridman 9:53 Right.

Daniel Kahneman 9:53 ... and it engages short-term memory and it engages executive function and it makes you incapable of doing other things at the same time. So, uh, the- the main characteristic of system two that there is mental effort involved and there's a limited capacity for mental effort, whereas system one is effortless essentially. That's the major distinction.

Lex Fridman 10:16 So, you talk about their- you know, it's really convenient to talk about two systems, but you also mentioned just now and in general that there's no distinct two systems in the brain from a neurobiological, even from psychology perspective. But why does it seem to, uh ... from the experiments you've conducted, there does seem to be kind of emergent two modes of thinking. So, at some point these kinds of systems came into a brain architecture, maybe mammals share it, but- or, or do you not think of it at all in those terms that it's all a mush and these two things just emerge?

Daniel Kahneman 11:02 I mean, you know, evolutionary theorizing about this is cheap and-

Lex Fridman 11:07 Yeah. Fair enough.

Daniel Kahneman 11:07 ... and easy. So it's- the way I think about it is that it's very clear that animals, uh, have, have a perceptual system and that includes an ability to understand the world-

Lex Fridman 11:22 Mm-hmm.

Daniel Kahneman 11:23 ... at least to the extent that they can predict. They can't explain anything, but they can anticipate what's going to happen and that's a key form of understanding the world. And my crude idea is that we- what I call system two-

Lex Fridman 11:40 Mm-hmm.

Daniel Kahneman 11:40 ... uh, well, system two grew out of this and, you know, there is language and there is the capacity of manipulating ideas and the capacity of imagining futures and of imagining counterfactuals thing that haven't happened and, and to do conditional thinking. And there are really a lot of abilities that without language-... and without the, the very large brain that we have compared to others, it would be impossible. Uh, now, system one is more like what the animals have, but system one, uh, also can talk. I mean-

Lex Fridman 12:21 Right.

Daniel Kahneman 12:21 ... it has language, it understands language. Indeed, it speaks for us. I mean-

Lex Fridman 12:25 Yeah.

Daniel Kahneman 12:26 ... you know, I'm not choosing every word, uh, as a deliberate process. The words... I have some idea and then the words come out, and that's automatic and effortless.

Lex Fridman 12:37 And, uh, many of the experiments you've done is to show that, listen, system one exists and it does speak for us, and we should be careful about its... the voice it provides, because it, uh-

Daniel Kahneman 12:48 Well, I mean, you know, we have to trust it, um, because it's... the speed at which it acts. Uh, system two-

Lex Fridman 12:58 It's useful.

Daniel Kahneman 12:58 ... if we, if we're dependent on system two for survival, we wouldn't survive very long, because it's very slow.

Lex Fridman 13:05 Yeah, crossing the street, in New York.

Daniel Kahneman 13:06 Crossing the street. I mean, many things depend on their being automatic.

Lex Fridman 13:10 Yeah.

Daniel Kahneman 13:11 One very important aspect of system one, uh, that it's not instinctive. You used the word instinctive. It contains skills that clearly have been learned. So that skilled behavior like driving a car or, or speaking in fact, uh, skilled behavior has to be learned. And so it doesn't... you know, you don't come equipped with, with driving. You have to learn how to drive, and, and you have to go through a period where driving is not automatic before it becomes automatic. So...

Lex Fridman 13:47 Yeah, you construct... I mean, this is where you talk about heuristic and biases, is you, uh, to make it automatic, you create a pattern, and then, uh, system one essentially matches a new experience against a previously seen pattern. And when that match is not a good one, that's when the cogni- all the, all the mess-

Daniel Kahneman 14:06 Well-

Lex Fridman 14:06 ... happens, but it's... y- most of the time, it works, and so it's pretty -

Daniel Kahneman 14:09 Most of the time, the anticipation of what's going to happen next is correct.

Lex Fridman 14:14 Yeah.

Daniel Kahneman 14:15 And, and most of the time, uh, the plan about what you have to do is correct. And so most of the time, everything works just fine. What's interesting actually, is that in some sense, system one is much better as, at what it does-

Lex Fridman 14:32 Mm-hmm.

Daniel Kahneman 14:32 ... than system two is at what it does. That is, there is this quality of effortlessly solving enormously complicated problems-

Lex Fridman 14:41 Yes.

Daniel Kahneman 14:41 ... which clearly, uh, exists, so that the chess player, a, a very good chess player, uh, all the moves that come to their mind are strong moves. So, all the selection of strong moves happens unconsciously and automatically and very, very fast. And, and all that is in system one, so you... uh, system two verifies.

Lex Fridman 15:07 So, along this line of thinking, really, what we are, are machines that construct pretty effective system one. You could think of it that

Lex Fridman 15:17 way. So, so we're now talking about humans, but if you think about building artificial intelligence systems, robots, do you think all the features and bugs that you have highlighted in human beings are useful for constructing AI systems? So both systems are useful for perhaps-

Daniel Kahneman 15:37 Well-

Lex Fridman 15:37 ... instilling in robots?

Daniel Kahneman 15:39 What is happening these days is that actually what is happening in deep learning is, is more like a system one product than like a system two product.

Lex Fridman 15:53 Mm-hmm.

Daniel Kahneman 15:53 I mean, deep learning matches patterns and anticipate what's going to happen, so it's highly predictive. Uh, what-

Lex Fridman 16:02 That's right.

Daniel Kahneman 16:03 ... what d- deep learning doesn't have, and, you know, many people think that this is a critical... it, it doesn't have the ability to reason, so it, it does... uh, there is no system two there. But I think very importantly, it doesn't have any causality or any way to represent meaning and to represent real interactions. So, uh, until that is solved, uh, the... you know, what can be accomplished is marvelous and very exciting but limited.

Lex Fridman 16:35 That's actually really nice to think of, uh, current advances in machine learning as essentially system one advances. So how far can we get with just system one?

Daniel Kahneman 16:46 Well, um-

Lex Fridman 16:46 If we think of deep learning and artificial intelligence (laughs) systems as system one.

Daniel Kahneman 16:48 I mean, you know, it's very clear that DeepMind has already gone way, way beyond what people thought was possible. I think, I think the thing that has impressed me most about the developments in AI, is the speed. It's that things, at least in the context of deep learning, and maybe this is about to slow down, but things moved a lot faster than anticipated. The transition from solving, solving chess to solving Go, uh, was... I mean, that's bewildering how quickly it went. The move from AlphaGo to AlphaZero is sort of bewildering, the speed at which they accomplished that. Now, clearly, uh, they are... there re- so there are many problem that you can solve that way, but there are some problems for which you need something else. <|agent|><|en|> Well, reasoning and also... you know, the... uh, one of the real mysteries, uh, a psychologist Gar- Gary Marcus, who is also a critic of AI, um... I mean, he... what he points out, and I think he has a point, is that, uh, humans learn quickly.... uh, children don't need a million examples, they need two or three examples. So, clearly there is a fundamental difference. And what enables, uh, what enables a machine to, to learn quickly, what you have to build into the machine, because it's clear that you have to build some expectations or something in the machine to make it ready to learn quickly, uh, that's, that at the moment seems to be unsolved. I'm pretty sure that DeepMind is working on it but, um-

Lex Fridman 18:44 Yeah, they're-

Daniel Kahneman 18:45 ... if they have solved it, I, I haven't heard yet.

Lex Fridman 18:48 They're trying to actually, them and OpenAI are trying to, to start to get to use neural networks to reason. So assemble knowledge-

Daniel Kahneman 18:58 Yeah.

Lex Fridman 18:58 ... uh, of course causality is, temporal causality is out of reach to most everybody. You, you mentioned wha- the benefits of system one is essentially that it's fast, it allows us to function in the world.

Daniel Kahneman 19:11 Fast and skilled, yeah.

Lex Fridman 19:13 It's skill.

Daniel Kahneman 19:13 And it has a model of the world. You know, in a sense, I mean there was the earlier phase of, of, uh, AI, uh, attempted to model reasoning, and they were moderately successful, but you know, reasoning by itself doesn't get you m- much. Uh, deep learning has been much more successful in terms of, you know, what they can do. But now, it's an interesting question, whether it's approaching its limits. What do you think?

Lex Fridman 19:45 I think absolutely. So I, I just talked to Yann LeCun, he mentioned, you know... (laughs)

Daniel Kahneman 19:50 I know him.

Lex Fridman 19:51 So he thinks that, uh, the limits, we're not going to hit the limits with neural networks, that ultimately this kind of system one pattern matching will start to, start to look like system two, with, without significant transformation of the architecture. So I'm more with the, with the majority of the people who think that yes, neural networks will hit a limit in their capability.

Daniel Kahneman 20:17 Well he, on the one hand I have heard him tell Demis Hassabis essentially that, you know, what they have accomplished is not a big deal, that they have-

Lex Fridman 20:26 Mm-hmm.

Daniel Kahneman 20:26 ... just touched, that basically, you know, they can't do unsupervised learning-

Lex Fridman 20:32 Yeah.

Daniel Kahneman 20:32 ... in a, in an effective way. And, but you're telling me that he thinks that the current, within the current architecture, you can do causality and reasoning?

Lex Fridman 20:42 So he's very much a pragmatist in a sense that's saying that we're very far away, that there's still-

Daniel Kahneman 20:48 Yeah.

Lex Fridman 20:48 ... I think, uh, there's this idea that he says is, uh, we can only see one or two mountain peaks ahead, and there might be either a few more after or thousands more after.

Daniel Kahneman 21:00 Lots more.

Lex Fridman 21:00 Yeah. So that kind of idea.

Daniel Kahneman 21:02 I heard that metaphor, yeah.

Lex Fridman 21:04 (laughs) Right. But nevertheless, he doesn't see a, the final answer not fundamentally looking like one that we currently have. So neural networks being a huge part of that.

Daniel Kahneman 21:18 Yeah. I mean, that's very likely, because, because pattern matching is so much of what's going on. But...

Lex Fridman 21:26 And you can think of neural networks as processing information sequentially.

Daniel Kahneman 21:30 Yeah, I mean, uh, you know, there is, there is an important aspect to, for example, you get systems, uh, that translate and they do a very good job, but they really don't know what they're talking about. Uh, and, and, and for that I'm really quite surprised. For that, you would need, uh, you would need an AI that has sensation, an AI that is in touch with the world.

Lex Fridman 21:58 Yes.

Daniel Kahneman 21:58 Uh, and for that...

Lex Fridman 21:59 Has self-awareness and maybe even something resembles consciousness kind of ideas.

Daniel Kahneman 22:04 Certainly awareness of, you know, awareness of what's going on, so that the, the words have meaning or can get, uh, in touch with some perception or some action.

Lex Fridman 22:16 Yeah, so uh, that's a big thing for Yann as, uh, what he refers to as grounding to the physical space. So-

Daniel Kahneman 22:24 So that's what, we're talking about the same thing.

Lex Fridman 22:26 Yeah. So, but, so how, how you ground...

Daniel Kahneman 22:29 I mean the grounding, without grounding, then you get, you get a machine that doesn't know what it's talking about, because it is talking about the world ultimately.

Lex Fridman 22:40 The question, the open question is what it means to ground. I mean, we're very, uh, human-centric in our thinking, but what does it mean for a machine to understand what it means to be in this world? Does it need to have a body? Does it need to have a finiteness like we humans have? All of these elements, it's, it's a very...

Daniel Kahneman 23:00 Well...

Lex Fridman 23:00 (laughs) It's an open question.

Daniel Kahneman 23:02 Um, you know, I'm not sure about having a body, but having a perceptual system. Having a body would be very helpful too. I mean, if, if you think about human, mimicking human or...

Lex Fridman 23:13 Okay.

Daniel Kahneman 23:13 But having a perception, that seems to be essential, uh, so that you can build, you can accumulate knowledge about the world. So if a, you can im- you can imagine a human completely paralyzed and there's a lot that the human brain could learn, you know, with a paralyzed body. So, uh, if we got a machine that could do that, that would be a big deal.

Lex Fridman 23:39 And then the flip side of that, something you see in children and something in machine learning world is called active learning, maybe it is also in...

Daniel Kahneman 23:49 Yeah.

Lex Fridman 23:49 ... is, uh, being able to play with the world. Uh, how important for developing system one or s- or system two do you think it is to play with the world, to be able to interact with the world?

Daniel Kahneman 24:01 Well, certainly a lot, a lot of what you learn is you learn to anticipate, uh, the outcomes of your actions. I mean, you can see that how babies learn it, you know, with their hands, uh, they, how they learn...... uh, you know, to connect, uh, you know, the movements of their hands with something that clearly is something that happens in the brain, and, and, and the ability of the brain to learn new patterns. So, you know, it's the kind of thing that you get with artificial limbs, that you connect it, and then people learn to operate the artificial limb r- you know, really impressively quickly, at least from, from what I hear.

Lex Fridman 24:43 Yeah.

Daniel Kahneman 24:43 Uh, so we have a system that is ready to learn the world through action.

Lex Fridman 24:49 At the risk of going into way too mysterious of land, what do you think it takes to build a system like that? Wh- obviously, we're very far from understanding how the, the brain works, but how difficult is it to build this-

Daniel Kahneman 25:07 I-

Lex Fridman 25:07 ... mind of ours?

Daniel Kahneman 25:08 You know, I mean, I think that Yann LeCun's answer, that j- we don't know how many mountains there are, I think that's a very good answer. I think that, uh, you know, uh, if you, if you look at what Kurt- Ray Kurzweil is saying, that strikes me as off-the-wall.

Lex Fridman 25:24 Yeah.

Daniel Kahneman 25:24 But, uh, but I think people are much more realistic than that. We're actually ... Demis Hassabis is and Yann is, and so the people who are actually doing the work are fairly realistic, I think.

Lex Fridman 25:39 To maybe phrase it another way, from a perspective not of building it, but from understanding it, how complicated are human beings in, in the following sense: You know, I work with autonomous vehicles and pedestrians, so we tried to model pedestrians. How difficult is it to model a human being, their perception of the world, the two systems they operate under, sufficiently to be able to predict whether the pedestrian's gonna cross the road or not?

Daniel Kahneman 26:11 I'm, you know, I'm fairly optimistic about that, actually-

Lex Fridman 26:14 Mm-hmm.

Daniel Kahneman 26:15 ... because what we're talking about is, uh, a huge amount of information that every vehicle has and that feeds into one system, into one gigantic system. And so anything that any vehicle learns becomes part of what the whole system knows.

Lex Fridman 26:34 Yes.

Daniel Kahneman 26:35 And with, with a system multiplier like that, uh, there is a lot that you can do. So, human beings are very complicated, but ... and, and, you know, a system is going to make mistakes, but human makes mistakes. I think that s- they'll be able to ... I think they are able to anticipate pedestrians, otherwise, (laughs) a lot would happen. They're able to, uh, you know, they're able to get into a roundabout and into the, into traffic, so they must know both to expect or to anticipate how people will react when they're sneaking in, and there's a lot of learning that's involved in that.

Lex Fridman 27:19 Currently, the pedestrians are treated as things that cannot be hit, and they're not treated as agents with whom you interact in a game theoretic way. So ...

Daniel Kahneman 27:36 I mean-

Lex Fridman 27:36 It's not ... It's a totally open problem and every time somebody tries to solve it, it seems to be harder than we think. And nobody's really tried to seriously solve the problem of that dance, because, uh - I'm not sure if you've thought about the problem of pedestrians - but you're really putting your life in the hands of the driver.

Daniel Kahneman 27:55 You know, there is a dance, this part of the-

Lex Fridman 27:58 There's a dance.

Daniel Kahneman 27:58 ... dance that would be quite complicated. But for example, when I cross the street and there is a vehicle approaching, I look the driver in the eye, and I think many people do that.

Lex Fridman 28:09 Yeah.

Daniel Kahneman 28:09 And, you know, that's a signal, uh, that, that I'm sending, and I would be sending that machine to an autonomous vehicle, and it had better understand it because it means I'm crossing.

Lex Fridman 28:22 So, and there's another thing you do that actually ... So I'll tell you what you do, 'cause we watched, I watched, uh, hundreds of hours of video on this, is when you step in the street, you do that before you step in the street. And when you step in the street, you actually look away.

Daniel Kahneman 28:37 Look away.

Lex Fridman 28:37 Yeah.

Daniel Kahneman 28:38 Yeah.

Lex Fridman 28:38 Uh, now, what, what does that- (laughs) what, what that's saying is, I mean, you're trusting that the car, who hasn't s- slowed down yet, will slow down.

Daniel Kahneman 28:48 Yeah. And you're telling him-

Lex Fridman 28:50 Yeah.

Daniel Kahneman 28:50 ... "I'm committed."

Lex Fridman 28:51 Yeah.

Daniel Kahneman 28:51 I mean, this is like in a game of chicken, so I'm committed.

Lex Fridman 28:55 Yeah.

Daniel Kahneman 28:55 And if I'm committed, I'm looking away. So, there is ... you, you just have to stop.

Lex Fridman 29:01 So the question is whether a machine that observes that needs to understand mortality.

Daniel Kahneman 29:07 Here, I'm not sure that it's got to understand so much as it's got to anticipate. So, and here ... but, you know, you're surprising me because here I would think that maybe you can anticipate without understanding-

Lex Fridman 29:27 Mm-hmm.

Daniel Kahneman 29:27 ... because I think this is clearly what's happening playing Go or in playing chess.

Lex Fridman 29:32 Yes.

Daniel Kahneman 29:32 There's a lot of anticipation and there is zero understanding.

Lex Fridman 29:35 Exactly.

Daniel Kahneman 29:36 So, uh, th- I thought that you didn't need a model of the human-

Lex Fridman 29:43 Yes.

Daniel Kahneman 29:44 ... and a model of the human mind to avoid hitting pedestrians. But you are suggesting that actually-

Lex Fridman 29:51 That you do, yeah.

Daniel Kahneman 29:52 You do.

Lex Fridman 29:52 As ... and-

Daniel Kahneman 29:53 Then it's, then it's a lot harder, I thought.

Lex Fridman 29:56 So this is ... (laughs) And I have a follow-up question to see where your intuition lies,

Lex Fridman 30:01 is it seems that almost every robot-human collaboration system is a lot harder than people realize. So, do you think it's possible for robots and humans to collaborate successfully?Uh, we, we talked a little bit about semi-autonomous vehicles, like in the Tesla, autopilot, but just in tasks in general... i- if you think, we talked about current neural networks being kind of system one. Do you think, uh, those same systems can borrow humans for system two type tasks and collaborate successfully?

Daniel Kahneman 30:41 Well, I think that in any system where humans and, and the machine interact, uh, the human will be superfluous within a fairly short time. Uh, that is if, if the machine is advanced enough so that it can really help the human, then it may not need the human for a long time. Now, it would be very interesting if, if there are problems that for some reason the machine doesn't, cannot solve, but that people could solve, then you would have to build into the machine an ability to recognize that it is in that kind of problematic situation-

Lex Fridman 31:21 Mm.

Daniel Kahneman 31:22 ... and, and to call the human. That, that cannot be easy without understanding. That is, it's, it must be very difficult to, to program a recognition that you are in a problematic situation without understanding the problem, but...

Lex Fridman 31:41 That's very true. In order to understand the full scope of situations that are problematic, you almost need to be smart enough-

Daniel Kahneman 31:50 To solve it.

Lex Fridman 31:51 ... to solve all those problems.

Daniel Kahneman 31:53 Yeah. It's not clear to me how much the machine will need the human. I think the example of chess is very instructive. I mean, there was a time at which Kasparov was saying that human-machine combinations will beat everybody. Uh, even Stockfish doesn't need people.

Lex Fridman 32:12 Yeah.

Daniel Kahneman 32:12 And AlphaZero certainly doesn't need people.

Lex Fridman 32:16 The question is, just like you said, how many problems are like chess and how many problems are the ones where are not like chess, where-

Daniel Kahneman 32:24 Let me-

Lex Fridman 32:24 ... well, every problem probably in the end is like chess. The question is, how long is that transition period?

Daniel Kahneman 32:30 I mean, you know, that's, that's a question I would ask you in terms of... I mean, autonomous vehicle, just driving, is probably a lot more complicated than Go to solve the-

Lex Fridman 32:41 Yes.

Daniel Kahneman 32:42 ... to solve the problem.

Lex Fridman 32:42 And, and that's surprising to people.

Daniel Kahneman 32:43 Because it's open. No. I mean, I, you know, it wouldn't, that's not surprising to me because the, because the, there is a hierarchical aspect to this, which is recognizing a situation, and then within the situation, bringing, bringing up the relevant knowledge.

Lex Fridman 33:03 Right.

Daniel Kahneman 33:04 And, uh, and for that hierarchical type of system to work, uh, you need a more complicated system than we currently have.

Lex Fridman 33:16 A lot of people think because as human beings, this is probably the, the cognitive biases, they think of driving as pretty simple because they think of their own experience. This is actually a, a b- big problem for AI researchers or people thinking about AI because they evaluate how hard a particular problem is based on very limited knowledge-

Daniel Kahneman 33:41 Yeah.

Lex Fridman 33:42 ... ba- basically on how hard it is for them to do the task.

Daniel Kahneman 33:45 Yeah.

Lex Fridman 33:46 And then they take for granted... I me- maybe you can speak to that because most people tell me driving is trivial, and-

Daniel Kahneman 33:54 Well-

Lex Fridman 33:54 ... and humans, in fact, are terrible at driving is what people tell me. And I see humans, and humans are actually incredible at driving, and driving is really terribly difficult.

Daniel Kahneman 34:04 Yeah.

Lex Fridman 34:04 Uh, so do you... (laughs) is that just another element of the effects that you've described in your work on the psychology side?

Daniel Kahneman 34:15 Well...

Lex Fridman 34:15 And how can we-

Daniel Kahneman 34:15 No. I mean, I haven't really, you know, I've, I, I would say that my research has contributed nothing to understanding the ecology and to understanding the structure of situations-

Lex Fridman 34:26 Yeah.

Daniel Kahneman 34:26 ... and the complexity of problems. Uh, so all, all we know is very clear, that, that Go, it's endlessly complicated, but it's very constrained. So, uh, and, and in the real world, there are far fewer constraints, and, and many more potential surprises. So, uh...

Lex Fridman 34:50 So that's obviously, because it's not always obvious to people, right? So when you think about...

Daniel Kahneman 34:55 Well, I mean, you know, people thought that reasoning was hard-

Lex Fridman 35:00 Right.

Daniel Kahneman 35:00 ... and perceiving was easy, but, you know, they quickly learned that actually modeling vision was tremendously complicated and modeling, even proving theorems was relatively straightforward.

Lex Fridman 35:15 To push back on that a little bit, on the quickly part, they haven't... (laughs) it took several decades to learn that, and most people still haven't learned that. I mean, our intuition, of course, AI researchers have, but you, you drift a little bit outside the specific AI field, the, the intuition is still perceptually to solving the task, right?

Daniel Kahneman 35:36 Oh, yeah. No, I mean, that's true. Intuitions, the intuitions of the public haven't changed r- radically, and they are, they are, as you said, they're evaluating the complexity of problems by how difficult it is for, for them to solve the problems. And that's got very little to do with the complexities of solving them in AI.

Lex Fridman 35:59 How do you think, from the perspective of AI researcher, do we deal with the intuitions of the public? So in trying to think, I mean, arguably, the combination of, uh, hype, investment, and the public intuition is what led to the AI winters. I'm sure that same could be applied to tech or...... that the intuition of the public leads to media hype, leads to companies investing in the tech, and then the tech doesn't make the companies money, and then there's a crash. Is there a way to educate people sort of to fight the, let's call it System 1 thinking?

Daniel Kahneman 36:44 In general, no. I mean-

Lex Fridman 36:46 (laughs)

Daniel Kahneman 36:47 ... I, I, I think that's the simple answer. Uh, and it's going to take a long time before the understanding of, uh, what those systems can do becomes, you know, part... becomes public knowledge. Uh, and, and then... and the expectations, you know, there are several aspects that are going to be very compl- uh, that, uh, uh, the... (pause of three seconds) The fact that you have a device that cannot explain itself is a major, major difficulty, and, uh, and we're already seeing that. I mean, this is, this is really something that is happening. So it's happening in the judicial system. So you have, uh, you have systems that are clearly better at predicting parole violations than-

Lex Fridman 37:45 Right.

Daniel Kahneman 37:45 ... uh, than judges, but, uh, but they can't explain their reasoning. And so, uh, people don't want to trust them.

Lex Fridman 37:56 We, uh, seem to, in System 1 even, use cues to make judgments about our environment. So this explainability point, do you think humans can explain stuff-

Daniel Kahneman 38:10 No. But-

Lex Fridman 38:11 ... themselves?

Daniel Kahneman 38:12 ... uh, I mean, there is a very interesting, uh, aspect of that. Humans think they can explain themselves.

Lex Fridman 38:21 Right.

Daniel Kahneman 38:22 So when you say something, and I ask you, "Why do you believe that?" then reasons will occur to you, and you will... but actually, my own belief is that in most cases, the reasons have very little to do with why you believe what you believe. So that the reasons are a story that, that comes to your mind when you need to explain yourself. But, um, but, but people traffic in those explanations. I mean, the human interaction depends on those shared fictions and, and the stories that people tell themselves.

Lex Fridman 38:58 You just made me actually realize, and we'll talk about stories in a second, uh, that, not to be cynical about it, but perhaps there's a whole movement of people trying to do explainable AI. And really, we don't necessarily need to explain, AI doesn't need to explain itself. It just needs to tell a convincing story.

Daniel Kahneman 39:21 Yeah.

Lex Fridman 39:22 (laughs)

Daniel Kahneman 39:22 Absolutely.

Lex Fridman 39:23 It doesn't neces- the story doesn't necessarily need to, uh, reflect the truth as... it might... it just needs to be convincing. There's something to that.

Daniel Kahneman 39:32 Uh, it can... you can say exactly the same thing in a way that's... sounds cynical or doesn't sound cynical.

Lex Fridman 39:38 Right. Sure.

Daniel Kahneman 39:38 I mean, so... but, but the objective-

Lex Fridman 39:42 Brilliant.

Daniel Kahneman 39:42 ... of having an explanation is, is to tell a story that will be acceptable to people. And, uh, and, and for it to be acceptable and to be robustly acceptable, it has to have some element (laughs) of truth. But, but the objective is for people to accept it.

Lex Fridman 40:05 That's quite brilliant, actually. Uh, but so on the,

Lex Fridman 40:08 uh, on the stories that we tell, sorry to ask me the... ask you the question that most people know the answer to. But, uh, you talk about two selves in terms of how life is lived, the experienced self and the remembering self. Can you describe the distinction between the two?

Daniel Kahneman 40:26 Well, sure. I mean, the... there is an aspect of, uh, of life that occasionally... you know, most of the time, we just live, and we have experiences, and they're better, and they are worse, and it goes on over time. And mostly, we forget everything that happens, or we forget most of what happens. Then occasionally, you... when something ends or at different points, uh, you evaluate the past, and you form a memory, and the memory is schematic. It's not that you can roll a film of an interaction. You construct, in effect, the elements of a story about an, about an episode. So there is the experience, and there is the story that is created about the experience, and that's what I call the remembering. So I, uh, I had the image of two selves. So there is a self that lives, and there is a self that evaluates life. Now, the paradox and the deep paradox in that is that, um, we have one system or one self that does the living, but the other system, uh, the remembering self, is all we get to keep. And basically, decision-making and, and everything that we do is governed by our memories, not by what actually happened. It's, it's governed by, by the story that we told ourselves or by the story that we're keeping. So that's, that's the distinction.

Lex Fridman 42:04 I mean, there's a lot of brilliant ideas about the pursuit of happiness that come out of that. Wh- what are the properties of happiness which emerge from, uh-

Daniel Kahneman 42:12 Well, I mean, the-

Lex Fridman 42:13 ... the remembering self?

Daniel Kahneman 42:14 There are, there are properties of how we construct stories that are really important. So, uh, that I studied a few. But, but...... a couple are really very striking, and one is that in stories, time doesn't matter.

Lex Fridman 42:33 Hmm.

Daniel Kahneman 42:33 There's a sequence of events or they'll highlight or not, the, and- and how long it took, you know, they lived happily ever after, or, and three years later, something, it... Time really doesn't matter, and in stories, events matter, but time doesn't. That- that leads to a very interesting set of problems, because time is all we got to live. I mean, you know, time is the currency of life, uh, and yet time is not represented basically in evaluative memories. So that- that creates a lot of, uh, paradoxes that I've thought about.

Lex Fridman 43:18 Yeah, they are fascinating. But if you were to give, uh, advice on how one lives a happy life-

Daniel Kahneman 43:28 Well-

Lex Fridman 43:28 ... based on such properties, what- what's the optimal...

Daniel Kahneman 43:32 Well, you know, I gave up... I abandoned happiness research because I couldn't solve that problem.

Lex Fridman 43:38 Yeah.

Daniel Kahneman 43:38 I couldn't, I couldn't see, uh, and in the first place, it's very clear that if you do talk in terms of those two selves, then that what makes the remembering self happy and what makes the experiencing self happy are different things. And I- I asked the question, uh, of suppose you're planning a vacation and you're just told that at the end of the vacation you'll get an amnesic drug so you remember nothing, and they'll also destroy all your photos so there'll be nothing. Would you still go to the same vacation? And- and it's... It turns out we go to vacations in large part to construct memories, not to have experiences, but to construct memories, and it turns out that the vacation that you would want for yourself if you knew what you would not remember is probably not the same vacation that you will want for yourself if you will remember. So, uh, I have no solution to these problems. But clearly, those are big issues-

Lex Fridman 44:47 And you've talked about actually-

Daniel Kahneman 44:48 ... difficult issues.

Lex Fridman 44:49 You've talked about sort of how many minutes or hours you spend about the vacation, it's an interesting way to think about it, because that's how you really experience the vacation outside the being in it. But there's also a modern... I don't know if you think about this or interact with it, there's a modern way to, uh, magnify the remembering self, which is by posting on Instagram, on Twitter, on social networks. A lot of people live life for the picture that you take, that you post somewhere. And now thousands of people share it and potentially- potentially millions, and then you can relive it even much more than just those minutes. Do you think about that-

Daniel Kahneman 45:32 I-

Lex Fridman 45:32 ... magnification much?

Daniel Kahneman 45:34 You know, I'm too old for social networks. I, you know, I- I've never seen Instagram, so-

Lex Fridman 45:41 (laughs)

Daniel Kahneman 45:42 ... I cannot really speak intelligently about those things. I'm just too old.

Lex Fridman 45:46 But it's interesting to watch the exact effects you described?

Daniel Kahneman 45:49 I- I think it will make a very big difference. I mean, and it will make... It will also make a difference, and that I don't know, whether, uh... It's clear that in some ways the devices that serve us, uh, supplant function. So you don't have to remember phone numbers, you don't have... You really don't have to know facts. I mean, the number of conversations I'm involved with where somebody says, "Well, let's look it up."

Lex Fridman 46:21 Yeah.

Daniel Kahneman 46:21 Uh, so it's- it's a... In a way, it's made conversations... Well, it's- it means that it's much less important to know things. You know, it used to be very important to know things. This is changing. So the requirements of that- that we have for ourselves and for other people are changing because of all those supports and because... And I have no idea what Instagram does-

Lex Fridman 46:53 (laughs)

Daniel Kahneman 46:53 ... but it's, uh-

Lex Fridman 46:55 Well, I'll tell you-

Daniel Kahneman 46:55 ... I wish I knew.

Lex Fridman 46:56 I mean, I- I wish I could just have the... my remembering self could enjoy this conversation, but I'll get to enjoy it even more by having wat- by watching it and then talking to others, it'll be about 100,000 people, scary as it is to say, (laughs) "Well, listen or watch this." Right? It changes things. It changes the experience of the world, that you seek out experiences which could be shared in that way. It's in... And- and I haven't seen... It's- it's the same effects that you described, and I don't think the psychology of that magnification has been described yet, 'cause it's a new world.

Daniel Kahneman 47:33 I mean, you know, the sharing... There was a per- there was a time when people read books and, uh-

Lex Fridman 47:41 (laughs)

Daniel Kahneman 47:42 ... and- and you could assume that your friends had read the same books that you read.

Lex Fridman 47:49 Mm-hmm.

Daniel Kahneman 47:50 So there was-

Lex Fridman 47:52 Kind of invisible sharing that you're-

Daniel Kahneman 47:54 There was a lot of sharing going on, and there was a lot of assumed common knowledge.

Lex Fridman 48:00 Yeah.

Daniel Kahneman 48:00 And, you know, that was built in. I mean, it was obvious that you had read The New York Times, it was obvious that you had read the reviews. I mean, uh, so a lot was taken for granted that was shared, uh, and, you know, when there were... When there were three television channels, it was obvious that you'd seen one of them, probably the same as... Uh, so sharing ex- sharing-

Lex Fridman 48:27 Has always been there.

Daniel Kahneman 48:28 Always... It was always there, it was just different.

Lex Fridman 48:32 At the risk of, uh, inviting mockery from you, let me say that... (laughs) that I'm also a fan of Sartre and Camus and existentialist philosophers. And, um, I'm joking, of course, about mockery. But from the perspective of the two selves, what do you think of the existentialist philosophy of life? So, trying to really emphasize the experiencing self as the proper way to... uh, or the best way to live life.

Daniel Kahneman 49:06 I don't know enough philosophy to answer that, but it's not, uh... You know, the emphasis on, on experience is also the emphasis in Buddhism.

Lex Fridman 49:17 Yeah, right. That's right.

Daniel Kahneman 49:18 So, uh, that's... You just have got to, to experience things and, and, and not to evaluate, and not to pass judgment, and not to score, not to keep score. So, uh-

Lex Fridman 49:33 But if... When you look at the, the grand picture of experience, do you think there's something to that? That one, one of the ways to achieve contentment, and maybe even happiness, is letting go of any of the things... uh, any of the procedures of the remembering self?

Daniel Kahneman 49:52 Well, yeah. I mean, I think, you know, if one could imagine a life in which people don't score themselves-

Lex Fridman 49:59 Mm-hmm.

Daniel Kahneman 50:00 ... uh, it, it feels as if that would be a better life. As if the self-scoring and, you know, "How am I doing?" uh, kind of question, uh, is not, is not a very happy thing to have. But I got out of that field because I couldn't solve that problem.

Lex Fridman 50:23 You couldn't solve the problem.

Daniel Kahneman 50:23 And, and that was because my intuition was that the experiencing self, that's reality. But then it turns out that what people want for themselves is not experiences. They want memories and they want a good story about their life. And so, you cannot have a theory of happiness that doesn't correspond to what people want for themselves. And when I, when I realized that this, this was where things were going, I really sort of left the field of research.

Lex Fridman 50:55 Do you think there's something instructive about this emphasis of reliving memories in building AI systems? So, currently, artificial intelligence systems are more like experiencing self, in that they react to the environment, there's some pattern-formation, like, uh, learning and so on, but you really don't construct memories. Uh, except in reinforcement learning every once in a while that you replay over and over.

Daniel Kahneman 51:25 Yeah. But, but, you know, that would... in principle, would not be... you know-

Lex Fridman 51:30 Do you think that's useful? Do you think it's a feature or a bug of human beings that we, uh... e- that we look back?

Daniel Kahneman 51:37 Oh, I think that's definitely a feature. That's not a bug. I mean, you, you have to look back in order to look forward. So, uh, without, without looking back, you couldn't, you couldn't really intelligently look forward.

Lex Fridman 51:53 You're looking for the echoes of the same kind of experience in order to predict how... what the future holds?

Daniel Kahneman 51:58 Yeah.

Lex Fridman 51:59 Uh, though Viktor Frankl, in his book, Man's Search for Meaning, I'm not sure if you've read, but describes his experience at the consecration-... uh, concentration camps during World War II as a way to describe that finding, identifying a purpose in life, a positive purpose in life, can save one from suffering. First of all, do you connect with the philosophy that he describes there, and...?

Daniel Kahneman 52:28 Not really. I mean, the... So, I can, I can really see that somebody who has that feeling of purpose and meaning and so on, that that could sustain you. Uh, I, in general, don't have that feeling. And I'm pretty sure that if I were in a concentration camp, I'd, I'd give up and die, you know? So, he talks... He is, he is a survivor-

Lex Fridman 52:56 Yeah.

Daniel Kahneman 52:56 ... and, you know, he survived with that. And I'm, and I'm not sure how essential to survival this sense-

Lex Fridman 53:04 Purpose is, yeah.

Daniel Kahneman 53:06 ... is. But I do know, when I think about myself, that I would have given up at, "Oh, yeah, this isn't going anywhere." Uh, and there is, there is a sort of character that, that, that manages to survive in conditions like that. And then, because they survive, they tell stories, and it sounds as if they survived because of what they were doing. We have no idea. They survived because of the kind of people that they are, and they're the kind of people who survives and will tell themselves stories of a particular kind. So, I'm not, uh... I-

Lex Fridman 53:41 So, d- you don't think seeking purpose is a significant driver in our behavior?

Daniel Kahneman 53:46 Oh, I mean, it's, it's a very interesting question. Because when you ask people whether it's very important to have meaning in their life, they say, "Oh, yes, that's the most important thing." But when you ask people, "What kind of a day did you have?" and, and, you know, "What were the experiences that you remember?" you don't get much meaning. You get social experiences. Then, uh... And, and some people say that, for example, in, in, in child... you know, in taking care of children, the fact that they are your children and you're taking care of them, uh, makes a very big difference. I think that's entirely true, uh, but it's more because...... of a story that we're telling ourselves, which is a very different story when we're taking care of our children or when we're taking care of other things.

Lex Fridman 54:44 Jumping around a little bit,

Lex Fridman 54:46 in doing a lot of experiments, let me ask a question: Most of the work I do, for example, is in- in the w- in the real world, but m- most of the clean, good science that you can do is in the lab, so that distinction... D- do you think we can understand the fundamentals of human behavior through controlled experiments in the lab? If we talk about pupil diameter, for example, it's much easier to do when you can control lighting conditions, right?

Daniel Kahneman 55:19 Yeah, of course.

Lex Fridman 55:20 Uh, so when we look at driving, lighting variation destroys-

Daniel Kahneman 55:26 Yeah. I- yeah.

Lex Fridman 55:27 ... almost completely your ability to use pupil diameter. But in the lab, for, uh, as I mentioned, semi-autonomous or autonomous vehicles and driving simulators, we can't- we don't capture true, honest, uh, human behavior in that particular domain. So, in your... What's your intuition? How much of human behavior can we study in this controlled environment of the lab?

Daniel Kahneman 55:53 A lot, but you'd have to verify it, you know, that you're- your conclusions are- are basically limited to the situation, to the experimental situation. Then you have to jump the- the big inductive leap to the real world, uh, so... And- and that's the flare, that's where the difference, I think, between the good psychologists and others that are mediocre is in the sense that- that your experiment captures something that's important-

Lex Fridman 56:28 Right.

Daniel Kahneman 56:28 ... and something that's real. And others are just running experiments.

Lex Fridman 56:33 So, what is that? Like, the birth of an idea to its development in your mind, to something that leads to an experiment. Is that similar to maybe, like, what Einstein or a good physicist do as your intuition?

Daniel Kahneman 56:46 Yeah.

Lex Fridman 56:46 You basically use your intuition to build up...

Daniel Kahneman 56:49 Yeah, but I mean, you know, it's- it's very skilled intuition.

Lex Fridman 56:53 Right. Absolutely, absolutely.

Daniel Kahneman 56:53 I mean, I- I just had that experience actually. I had an idea that, uh, turned out to be a very good idea, uh, a couple of days ago. And- and you- and you have a sense of that building up, so I'm working with a collaborator.

Lex Fridman 57:08 Mm-hmm.

Daniel Kahneman 57:08 And he- he essentially was saying, you know, "What- what are you doing? You know, what's- what's going on?" And I was- I really... I- I couldn't exactly explain it, but I knew this is going somewhere. But, you know, I've been around that game for a very long time, and so I can... You- you develop that anticipation that, yes, this- this is worth following up with.

Lex Fridman 57:31 This is something- there's something here.

Daniel Kahneman 57:32 And that's- that's part of the skill.

Lex Fridman 57:34 Is that something you can reduce to words in describing a process in- in the form of advice to others?

Daniel Kahneman 57:42 No. No.

Lex Fridman 57:44 Follow your heart, essentially? (laughs)

Daniel Kahneman 57:46 I mean, you know, it's- it's like trying to explain what it's like to drive. It's not-

Lex Fridman 57:51 Yeah.

Daniel Kahneman 57:51 You've got to break it apart, and it's not, uh...

Lex Fridman 57:54 And then you lose the essence of it.

Daniel Kahneman 57:55 And then you lose the experience, though.

Lex Fridman 57:58 You mention collaboration. You've written about your collaboration with Amos Tversky, that... This is you writing, "The 12 or 13 years in which most of our work was joint were years of interpersonal and intellectual bliss. Everything was interesting, almost everything was funny, and there was a current joy of seeing an idea take shape. So many times in those years, we shared the magical experience of one of us saying something which the other one would understand more deeply than the speaker had done. Contrary to the old laws of information theory, it was common for us to find that more information was received than had been sent. I have almost never had the experience with anyone else. If you have not had it, you don't know how marvelous collaboration can be." So, let me ask a per- perhaps a silly question. Uh, how does one find and create such a collaboration? That may be asking, like, how does one find love, but-

Daniel Kahneman 59:01 You have... Yeah.

Lex Fridman 59:01 (laughs)

Daniel Kahneman 59:01 You have to be- you have to be lucky, um, and- and I think you have to have the character for that, because I've had many collaborations. I mean, none were as exciting as with Amos B., but I've had, and I'm having, just very... So, it's a skill. I think I'm good at it. Uh, not everybody's good at it. And then it's the luck of finding people who are also good at it.

Lex Fridman 59:32 Is there advice in the form- for- for a young scientist who also seeks to violate this law of information theory?

Daniel Kahneman 59:48 I really think it's so much luck is involved, and, you know, in- in those really serious collaborations, at least in my experience, are a very personal experience. And- and I have to like the person I'm working with. Otherwise, you know, I mean, there is that kind of a collaboration which is like, uh, an exchange, a commercial exchange of, uh, "I'm giving this, you give me that," but the- the real ones are interpersonal, they're between people who like each other, and- and who like making each other think, and who like the way that the other person responds to your thoughts. Uh, you have to be lucky.

Lex Fridman 1:00:36 Yeah, and I mean... But I already noticed the pa- even just me showing up here-... you've, uh, you've quickly started to digging in a particular problem I'm working on, and already new information started to emerge. If y- is that a process y- y- just a process of curiosity-

Daniel Kahneman 1:00:53 Yeah.

Lex Fridman 1:00:53 ... of talking to people about problems and seeing?

Daniel Kahneman 1:00:56 I'm curious about anything to do with AI and robotics and s- you know, and, uh, so, and I knew you were dealing with that, so I was curious.

Lex Fridman 1:01:05 Just follow your curiosity?

Daniel Kahneman 1:01:06 Yeah.

Lex Fridman 1:01:07 Jumping around on, on the psychology

Lex Fridman 1:01:09 front, the, uh, dramatic-sounding terminology of replication crisis, but really just the, at times, th- this effect that at times studies do not, are not fully generalizable. They don't-

Daniel Kahneman 1:01:29 You're being polite. Uh, it's worse than that, but... (laughs)

Lex Fridman 1:01:33 Is it?

Daniel Kahneman 1:01:33 Yeah.

Lex Fridman 1:01:33 So I'm actually not fully familiar-

Daniel Kahneman 1:01:36 Well, I mean-

Lex Fridman 1:01:36 ... to the degree how bad it is, right? So, what do you think is the source? Where do you think?

Daniel Kahneman 1:01:41 I think I know what's going on, actually. I mean, I have a theory about what's going on. And what's going on is that there is, first of all, a very important distinction between two types of experiments. And one type is within-subject, so it's the same person-

Lex Fridman 1:02:02 Right.

Daniel Kahneman 1:02:02 ... as two experimental conditions. And the other type is between-subjects, where some people are this condition, other people are that condition. They're different worlds. And between-subject experiments are much harder to predict and much harder to anticipate. And the reason, uh, and they're also more expensive because you need more people, and it's, it's just... So between-subject experiments is where the problem is.

Lex Fridman 1:02:35 Okay.

Daniel Kahneman 1:02:35 Uh, it's not so much in within-subject experiments. It's really between. And there is a very good reason why the intuitions of researchers about between-subject experiments are wrong, and that's because when you are a researcher, you are in a within-subject situation. That is, you are imagining the two conditions, and you see the causality, and you feel it.

Lex Fridman 1:03:03 Mm-hmm.

Daniel Kahneman 1:03:03 And, but in the between-subjects condition, they don't... They, uh, they see, they live in one condition, and the other one is just nowhere. So, our intuitions are very weak about between-subject experiments. And that, I think, is something that people haven't realized. And, and in addition, because of that, we have no idea about the power of, uh, manipulations, of experimental manipulations, because the same manipulation is much more powerful when, when you are in the two conditions-

Lex Fridman 1:03:43 Mm-hmm.

Daniel Kahneman 1:03:44 ... than when you live in only one condition. And so, the experimenters have very poor intuitions about between-subject experiments. And, and there is something else, which is very important, I think, uh, which is that almost all psychological hypotheses are true. That is, in the sense that, you know, directionally, if you have a hypothesis that A really causes B, that, that it's not true that A causes the opposite of B. Maybe A just has very little effect. But hypotheses are true-

Lex Fridman 1:04:23 Yeah.

Daniel Kahneman 1:04:23 ... mostly, except, mostly, they're very weak. They're much weaker than you think when you are having images of... So, uh, the reason I'm excited about that is that I recently heard about, uh, some, uh, some friends of mine, uh, who, uh, they e- essentially funded 53 studies of behavioral change-

Lex Fridman 1:04:51 Yep.

Daniel Kahneman 1:04:52 ... by 20 different teams of people with a very precise objective of changing the number of time that people go to the gym, but-

Lex Fridman 1:05:01 Mm-hmm.

Daniel Kahneman 1:05:01 ... you know, so... And, and the success rate was zero.

Lex Fridman 1:05:10 They're not successful?

Daniel Kahneman 1:05:10 Not one of the 53 studies worked.

Lex Fridman 1:05:13 (laughs)

Daniel Kahneman 1:05:13 Now, what's interesting about that is those are the best people in the field, and they have no idea what's going on. So they are not calibrated. They think that it's going to be powerful because they can imagine it, but actually it's just weak because the-

Lex Fridman 1:05:30 It's weak.

Daniel Kahneman 1:05:30 ... you are focusing on, on your manipulation, and it feels powerful to you.

Lex Fridman 1:05:36 Mm-hmm.

Daniel Kahneman 1:05:37 There's a thing that I've written about that's called the focusing illusion.

Lex Fridman 1:05:40 Mm-hmm.

Daniel Kahneman 1:05:40 That is that when you think about something, it looks very important-

Lex Fridman 1:05:46 But-

Daniel Kahneman 1:05:46 ... more important than it really is.

Lex Fridman 1:05:48 ... more important than it really is, but if you don't see that affects the 53 studies, doesn't that mean you just report that? So what, what's, I guess, the solution to that?

Daniel Kahneman 1:05:59 Well, I mean, the, the solution is for people to trust their intuitions less or to try out their intuitions before... I mean, experiments have to be pre-registered, and by the time you run an experiment, you have to be committed to it, and you have to run the experiment seriously enough and, uh, in a public... And so this is happening. The interesting thing is, uh, what, what happens before and how do people prepare themselves and how they run pilot experiments. It's going to train the way psychology is done, and it's already happening.

Lex Fridman 1:06:42 Do you have a hope for-This might connect to, uh, the- the s- study sample size.

Daniel Kahneman 1:06:49 Yeah.

Lex Fridman 1:06:49 Uh, do you have a hope for the internet or digitalization?

Daniel Kahneman 1:06:52 Well, I mean, you know, this is really happening. MTurk.

Lex Fridman 1:06:56 Yeah.

Daniel Kahneman 1:06:56 Uh, everybody is running experiments on MTurk, and- and it's very cheap and very effective, so...

Lex Fridman 1:07:04 Do you think that changes psychology essentially because you're- think- you can now-

Daniel Kahneman 1:07:08 I mean, all the-

Lex Fridman 1:07:08 ... run it on 10,000 subjects?

Daniel Kahneman 1:07:09 Eventually, it will.

Lex Fridman 1:07:11 Yeah.

Daniel Kahneman 1:07:11 I mean, I, you know, I can't put my finger on how exactly, but it's- that's been true in psychology with whenever an important new method came in, it changes the field. So... and- and MTurk is really a method because it-

Lex Fridman 1:07:30 Exactly.

Daniel Kahneman 1:07:30 ... it makes it very much easier to do something, to do some things.

Lex Fridman 1:07:35 Is there... uh, undergrad students will ask me, you know, how big a neural network should be for a particular problem, so let me ask you an equivalent- equivalent question (laughs) . Uh, how big sh- how many subjects does a study have for it to have a conclusive result in general?

Daniel Kahneman 1:07:53 Well, it depends on the strength of the effect. So if you're studying visual perception-

Lex Fridman 1:07:59 Mm-hmm.

Daniel Kahneman 1:08:00 ... or the perception of color, many of the- of the classic results in- in visual- in color perception were done on three or four people, and I think, and one of them was colorblind, but-

Lex Fridman 1:08:12 (laughs)

Daniel Kahneman 1:08:12 ... or- or partly colorblind, but on vision, you know, you- it's highly reliable.

Lex Fridman 1:08:20 The effect is strong.

Daniel Kahneman 1:08:20 Many, uh, people don't need a lot of replications for some type of- of neurological, uh, experiment, neuro. Uh, when you're studying weaker phenomena, and especially when you're studying them between subjects, then you need a lot more subjects than people have been running, and that is- that's one of the thing that are happening in psychology now is that the power- the statistical power of experiments is- is increasing rapidly.

Lex Fridman 1:08:54 Does the between subject as the number of subjects goes to infinity approach?

Daniel Kahneman 1:08:59 Well, I mean, you know-

Lex Fridman 1:09:00 (laughs)

Daniel Kahneman 1:09:01 ... goes to infinity is exaggerated, but people... the standard number of subjects for an experiment in psychology were 30 or 40, and for a weak effect, that's simply not enough, and you may need a couple of hundred. I mean, it's that- that sort of, uh, order of magnitude.

Lex Fridman 1:09:28 What are the major disagreements in theories and effects that you've observed throughout your career-

Daniel Kahneman 1:09:36 (laughs)

Lex Fridman 1:09:36 ... that still stand today?

Daniel Kahneman 1:09:38 Well, I mean, I-

Lex Fridman 1:09:38 You've worked on several fields (laughs) .

Daniel Kahneman 1:09:40 Yeah. I- I-

Lex Fridman 1:09:41 But what still is out there as- as- as major disagreement-

Daniel Kahneman 1:09:44 And then-

Lex Fridman 1:09:45 ... that pops into your mind?

Daniel Kahneman 1:09:46 ... and... I've had one extreme experience of c- you know, controversy with somebody who really doesn't like the work that Amis Tversky and I did, and- and he's been after us for 30 years or more. At least 30 years.

Lex Fridman 1:10:02 Do you want to talk about it?

Daniel Kahneman 1:10:04 Well, I mean, his name is Gert Gigerenzer. He's a well-known German psychologist, and that's the one controversy I have which I... it's been unpleasant, and- and no, I don't particularly want to talk about it.

Lex Fridman 1:10:20 (laughs) But is there... is there open questions even in your own mind? Every once in a while, you know, uh, we talked about semi-autonomous vehicles. In my own mind, I see what the data says, but I also am constantly torn. Do you have things where you- your studies have found something, but you're also intellectually torn about what it means, and there's maybe-

Daniel Kahneman 1:10:43 Well-

Lex Fridman 1:10:43 ... there've been maybe disagreements with how you... within your own mind (laughs) about particular things.

Daniel Kahneman 1:10:47 I mean, it's... you know, one of the things that are interesting is how difficult it is for people to change their mind. Essentially, uh, you know, once they're committed, people just don't change their mind about anything that matters, and that is surprisingly, but it's true about scientists. So the controversy that I described, uh, you know, that's been going on like 30 years, and it's never going to be resolved, uh, and you build a system, and you live within that system, and other- other systems of ideas look foreign to you, and- n- um, and there is very little contact and very little mutual influence. That happens a fair amount.

Lex Fridman 1:11:33 Do you have a hopeful advice or message on that? Uh, we... uh, thinking about science, thinking about politics, thinking about things that have impact on this world, how can we change our mind?

Daniel Kahneman 1:11:49 I think that... I mean, on things that matter, you know, which are political or reli- political or religious, and people just don't- don't change their mind, and by and large, and there's very little that you can do about it. Uh, the... what does happen is that if leaders change their mind, so for example, the public- the American public doesn't really believe in climate change, doesn't take it very seriously, but if some religious leaders decided this is a major threat to humanity, that would have a big effect, so that we- we have the opinions that we have not because we know why we have them but because we trust some people and we don't trust other people, and, uh... so it's much less about-... evidence than it is about stories.

Lex Fridman 1:12:49 So, uh, th- the way, one way to change your mind isn't at the individual level, is that the leaders-

Daniel Kahneman 1:12:55 It's at the collective level.

Lex Fridman 1:12:55 ... in the communities you look up with, the stories change, and therefore your mind-

Daniel Kahneman 1:12:59 Yeah.

Lex Fridman 1:12:59 ... changes with them.

Lex Fridman 1:13:01 So, there's a guy named Alan Turing, came up with the Turing test.

Daniel Kahneman 1:13:05 Yeah.

Lex Fridman 1:13:07 Uh, what, what do you think is a good test of intelligence? Perhaps we're drifting in a topic that we're, um, maybe philosophizing about, but what do you think is a good test for intelligence for an artificial intelligence system?

Daniel Kahneman 1:13:23 Well, the standard definition of, you know, of artificial general intelligence is that it can do anything that people can do, and it can do them better.

Lex Fridman 1:13:35 Yes.

Daniel Kahneman 1:13:35 And what, what we are seeing is that in many domains, you have domain-specific, uh, um, you know, devices or programs or software, and they beat people easily in specified way. What we are very far from is, uh, the general ability, uh, general purpose intelligence. So, we, w- in, in machine learning, people are approaching something more general. I mean, for AlphaZero was, was much more general than, than AlphaGo, and, but it's still extraordinarily narrow and specific in what it can do.

Lex Fridman 1:14:22 So, a test-

Daniel Kahneman 1:14:22 So, we're quite far from, from something that can, in every domain, think like a human, except better.

Lex Fridman 1:14:30 What aspects... So, the, the Turing test has been criticized, this natural language conversation-

Daniel Kahneman 1:14:36 Yeah.

Lex Fridman 1:14:36 ... that it's too simplistic. Uh, it's, it's easy to, quote-unquote, "pass" under, under constraints specified. Uh, what aspect of conversation would impress you if you heard it? Is it humor? Is it... uh, (laughs) wha- what, what would impress the heck outta you if, uh, if you saw it in conversation?

Daniel Kahneman 1:14:55 Yeah. I mean, certainly wit would-

Lex Fridman 1:14:57 Wit.

Daniel Kahneman 1:14:57 ... yeah, wit would be impressive. Uh, um, and, and humor would be more impressive than just factual conversation, which I think is, is easy. And allusions would be interesting, and metaphors would be interesting, I mean, but new metaphors, not practiced metaphors. So, there is a lot that's, you know, would be sort of impressive if... and that, uh, it's completely natural in conversation, but that you really wouldn't expect.

Lex Fridman 1:15:34 Does the possibility of creating an, a human-level intelligence or superhuman-level intelligence system excite you? Scare you?

Daniel Kahneman 1:15:44 Well, I mean, you know, I'm, uh-

Lex Fridman 1:15:44 How does it make you feel?

Daniel Kahneman 1:15:45 I find the whole thing fascinating, absolutely fascinating.

Lex Fridman 1:15:51 So, exciting?

Daniel Kahneman 1:15:52 I think, and exciting. It's also terrifying, you know? But, uh, but I'm not going to be around to see it. And, uh, so I'm curious about what is happening now, but I also know that, that predictions about it are silly.

Lex Fridman 1:16:11 (laughs)

Daniel Kahneman 1:16:11 Uh, we really have no idea what it will look like 30 years from now. No idea.

Lex Fridman 1:16:18 Speaking of silly, bordering on the profound, let me ask the question of, in your view, what is the meaning of it all?

Daniel Kahneman 1:16:27 (laughs)

Lex Fridman 1:16:28 The meaning of life?

Daniel Kahneman 1:16:30 Uh-

Lex Fridman 1:16:30 These, uh, descendant of great apes that we are, why... what drives us as a civilization, as a human being, as a force behind everything that you've observed and studied? Is there any answer, or is it all-

Daniel Kahneman 1:16:46 Uh-

Lex Fridman 1:16:46 ... just a beautiful mess?

Daniel Kahneman 1:16:49 There is no answer that, that I can understand. Uh, and I'm not, and I'm not actively looking for one, um, bec-

Lex Fridman 1:17:00 Do you think an answer exists?

Daniel Kahneman 1:17:02 No. There is no answer that we can understand. I'm not qualified to speak about what we cannot understand.

Lex Fridman 1:17:09 (laughs)

Daniel Kahneman 1:17:09 But there is, I know, that we cannot understand reality, you know? And, I mean, there are a lot of thing that we can do. I mean, you know, m- m- gravity waves. I mean, that's, that's a big moment for humanity. And-

Lex Fridman 1:17:25 Yeah.

Daniel Kahneman 1:17:25 ... when you imagine that ape, you know, being able to, to go back to The Big Bang.

Lex Fridman 1:17:30 (laughs)

Daniel Kahneman 1:17:31 That's, that... But, but-

Lex Fridman 1:17:34 But the why-

Daniel Kahneman 1:17:35 Yeah, the why.

Lex Fridman 1:17:36 ... is bigger than us.

Daniel Kahneman 1:17:37 (laughs) The why is hopeless, really.

Lex Fridman 1:17:40 Danny, thank you so much. It was an honor. Thank you for speaking today.

Daniel Kahneman 1:17:43 (laughs) Thank you.

Lex Fridman 1:17:45 Thanks for listening to this conversation, and thank you to our presenting sponsor, Cash App. Download it, use code: LEXPODCAST. You'll get $10, and $10 will go to FIRST, a STEM education nonprofit that inspires hundreds of thousands of young minds to become future leaders and innovators. If you enjoyed this podcast, subscribe on YouTube, give it five stars on Apple Podcasts, follow on Spotify, support it on Patreon, or simply connect with me on Twitter. And now, let me leave you with some words of wisdom from Daniel Kahneman. "Intelligence is not only the ability to reason, it is also the ability to find relevant material in memory and to deploy attention when needed." Thank you for listening, and hope to see you next time.