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

1. 0:00 – 8:18

   Dehumanization, in-groups vs out-groups, and the psychology of war

   1. DKDaniel Kahneman0:00

      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...... and what, what people can do.

   2. LFLex Fridman6:06

      So, the effect of the in group and the out group?

   3. DKDaniel Kahneman6:09

      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.

   4. LFLex Fridman7:06

      Mm-hmm.

   5. DKDaniel Kahneman7:08

      That's-

   6. LFLex Fridman7:08

      It's a question.

   7. DKDaniel Kahneman7:09

      It's an interesting question.

   8. LFLex Fridman7:10

      Are these artifacts of history or is it human nature?

   9. DKDaniel Kahneman7:14

      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.

   10. LFLex Fridman7:28

       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.

   11. DKDaniel Kahneman7:41

       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-

   12. LFLex Fridman8:04

       Yeah.

   13. DKDaniel Kahneman8:04

       ... 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-

2. 8:18 – 10:16

   System 1 vs System 2: effortless intuition and effortful reasoning

   1. LFLex Fridman8:18

      So, 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?

   2. DKDaniel Kahneman8:49

      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.

   3. LFLex Fridman8:58

      Yes.

   4. DKDaniel Kahneman8:58

      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.

   5. LFLex Fridman9:36

      You have to d- perform some algorithm in your head, some steps.

   6. DKDaniel Kahneman9:39

      Yes, and, and it takes time.

   7. LFLex Fridman9:41

      Yeah.

   8. DKDaniel Kahneman9:42

      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-

   9. LFLex Fridman9:53

      Right.

   10. DKDaniel Kahneman9:53

       ... 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.

3. 10:16 – 12:48

   Where the two-system idea comes from: evolution, language, and prediction

   1. LFLex Fridman10:16

      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?

   2. DKDaniel Kahneman11:02

      I mean, you know, evolutionary theorizing about this is cheap and-

   3. LFLex Fridman11:07

      Yeah. Fair enough.

   4. DKDaniel Kahneman11:07

      ... 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-

   5. LFLex Fridman11:22

      Mm-hmm.

   6. DKDaniel Kahneman11:23

      ... 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-

   7. LFLex Fridman11:40

      Mm-hmm.

   8. DKDaniel Kahneman11:40

      ... 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-

   9. LFLex Fridman12:21

      Right.

   10. DKDaniel Kahneman12:21

       ... it has language, it understands language. Indeed, it speaks for us. I mean-

   11. LFLex Fridman12:25

       Yeah.

   12. DKDaniel Kahneman12:26

       ... 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.

   13. LFLex Fridman12:37

       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-

4. 12:48 – 15:07

   Why we must trust System 1 (and when it fails)

   1. DKDaniel Kahneman12:48

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

   2. LFLex Fridman12:58

      It's useful.

   3. DKDaniel Kahneman12:58

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

   4. LFLex Fridman13:05

      Yeah, crossing the street, in New York.

   5. DKDaniel Kahneman13:06

      Crossing the street. I mean, many things depend on their being automatic.

   6. LFLex Fridman13:10

      Yeah.

   7. DKDaniel Kahneman13:11

      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...

   8. LFLex Fridman13:47

      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-

   9. DKDaniel Kahneman14:06

      Well-

   10. LFLex Fridman14:06

       ... happens, but it's... y- most of the time, it works, and so it's pretty -

   11. DKDaniel Kahneman14:09

       Most of the time, the anticipation of what's going to happen next is correct.

   12. LFLex Fridman14:14

       Yeah.

   13. DKDaniel Kahneman14:15

       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-

   14. LFLex Fridman14:32

       Mm-hmm.

   15. DKDaniel Kahneman14:32

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

   16. LFLex Fridman14:41

       Yes.

   17. DKDaniel Kahneman14:41

       ... 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.

5. 15:07 – 16:35

   Deep learning as “System 1”: pattern matching without causality or meaning

   1. LFLex Fridman15:07

      So, along this line of thinking, really, what we are, are machines that construct pretty effective system one. You could think of it that 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-

   2. DKDaniel Kahneman15:37

      Well-

   3. LFLex Fridman15:37

      ... instilling in robots?

   4. DKDaniel Kahneman15:39

      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.

   5. LFLex Fridman15:53

      Mm-hmm.

   6. DKDaniel Kahneman15:53

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

   7. LFLex Fridman16:02

      That's right.

   8. DKDaniel Kahneman16:03

      ... 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.

6. 16:35 – 21:30

   Speed of AI progress, sample efficiency, and the ‘mountain peaks’ metaphor

   1. LFLex Fridman16:35

      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?

   2. DKDaniel Kahneman16:46

      Well, um-

   3. LFLex Fridman16:46

      If we think of deep learning and artificial intelligence (laughs) systems as system one.

   4. DKDaniel Kahneman16:48

      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-

   5. LFLex Fridman18:44

      Yeah, they're-

   6. DKDaniel Kahneman18:45

      ... if they have solved it, I, I haven't heard yet.

   7. LFLex Fridman18:48

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

   8. DKDaniel Kahneman18:58

      Yeah.

   9. LFLex Fridman18:58

      ... 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.

   10. DKDaniel Kahneman19:11

       Fast and skilled, yeah.

   11. LFLex Fridman19:13

       It's skill.

   12. DKDaniel Kahneman19:13

       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?

   13. LFLex Fridman19:45

       I think absolutely. So I, I just talked to Yann LeCun, he mentioned, you know... (laughs)

   14. DKDaniel Kahneman19:50

       I know him.

   15. LFLex Fridman19:51

       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.

   16. DKDaniel Kahneman20:17

       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-

   17. LFLex Fridman20:26

       Mm-hmm.

   18. DKDaniel Kahneman20:26

       ... just touched, that basically, you know, they can't do unsupervised learning-

   19. LFLex Fridman20:32

       Yeah.

   20. DKDaniel Kahneman20:32

       ... 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?

   21. LFLex Fridman20:42

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

   22. DKDaniel Kahneman20:48

       Yeah.

   23. LFLex Fridman20:48

       ... 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.

   24. DKDaniel Kahneman21:00

       Lots more.

   25. LFLex Fridman21:00

       Yeah. So that kind of idea.

   26. DKDaniel Kahneman21:02

       I heard that metaphor, yeah.

   27. LFLex Fridman21:04

       (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.

   28. DKDaniel Kahneman21:18

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

   29. LFLex Fridman21:26

       And you can think of neural networks as processing information sequentially.

7. 21:30 – 25:39

   Grounding and embodiment: do machines need perception (or bodies) to understand?

   1. DKDaniel Kahneman21:30

      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.

   2. LFLex Fridman21:58

      Yes.

   3. DKDaniel Kahneman21:58

      Uh, and for that...

   4. LFLex Fridman21:59

      Has self-awareness and maybe even something resembles consciousness kind of ideas.

   5. DKDaniel Kahneman22:04

      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.

   6. LFLex Fridman22:16

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

   7. DKDaniel Kahneman22:24

      So that's what, we're talking about the same thing.

   8. LFLex Fridman22:26

      Yeah. So, but, so how, how you ground...

   9. DKDaniel Kahneman22:29

      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.

   10. LFLex Fridman22:40

       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...

   11. DKDaniel Kahneman23:00

       Well...

   12. LFLex Fridman23:00

       (laughs) It's an open question.

   13. DKDaniel Kahneman23:02

       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...

   14. LFLex Fridman23:13

       Okay.

   15. DKDaniel Kahneman23:13

       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.

   16. LFLex Fridman23:39

       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...

   17. DKDaniel Kahneman23:49

       Yeah.

   18. LFLex Fridman23:49

       ... 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?

   19. DKDaniel Kahneman24:01

       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.

   20. LFLex Fridman24:43

       Yeah.

   21. DKDaniel Kahneman24:43

       Uh, so we have a system that is ready to learn the world through action.

   22. LFLex Fridman24:49

       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-

   23. DKDaniel Kahneman25:07

       I-

   24. LFLex Fridman25:07

       ... mind of ours?

   25. DKDaniel Kahneman25:08

       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.

   26. LFLex Fridman25:24

       Yeah.

   27. DKDaniel Kahneman25:24

       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.

8. 25:39 – 29:56

   Autonomous driving and the pedestrian ‘dance’: anticipation vs understanding

   1. LFLex Fridman25:39

      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?

   2. DKDaniel Kahneman26:11

      I'm, you know, I'm fairly optimistic about that, actually-

   3. LFLex Fridman26:14

      Mm-hmm.

   4. DKDaniel Kahneman26:15

      ... 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.

   5. LFLex Fridman26:34

      Yes.

   6. DKDaniel Kahneman26:35

      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.

   7. LFLex Fridman27:19

      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 ...

   8. DKDaniel Kahneman27:36

      I mean-

   9. LFLex Fridman27:36

      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.

   10. DKDaniel Kahneman27:55

       You know, there is a dance, this part of the-

   11. LFLex Fridman27:58

       There's a dance.

   12. DKDaniel Kahneman27:58

       ... 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.

   13. LFLex Fridman28:09

       Yeah.

   14. DKDaniel Kahneman28:09

       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.

   15. LFLex Fridman28:22

       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.

   16. DKDaniel Kahneman28:37

       Look away.

   17. LFLex Fridman28:37

       Yeah.

   18. DKDaniel Kahneman28:38

       Yeah.

   19. LFLex Fridman28:38

       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.

   20. DKDaniel Kahneman28:48

       Yeah. And you're telling him-

   21. LFLex Fridman28:50

       Yeah.

   22. DKDaniel Kahneman28:50

       ... "I'm committed."

   23. LFLex Fridman28:51

       Yeah.

   24. DKDaniel Kahneman28:51

       I mean, this is like in a game of chicken, so I'm committed.

   25. LFLex Fridman28:55

       Yeah.

   26. DKDaniel Kahneman28:55

       And if I'm committed, I'm looking away. So, there is ... you, you just have to stop.

   27. LFLex Fridman29:01

       So the question is whether a machine that observes that needs to understand mortality.

   28. DKDaniel Kahneman29:07

       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-

   29. LFLex Fridman29:27

       Mm-hmm.

   30. DKDaniel Kahneman29:27

       ... because I think this is clearly what's happening playing Go or in playing chess.
9. 29:56 – 37:20

   Human–AI collaboration: will humans quickly become unnecessary?

   1. LFLex Fridman29:56

      So this is ... (laughs) And I have a follow-up question to see where your intuition lies, 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?

   2. DKDaniel Kahneman30:41

      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-

   3. LFLex Fridman31:21

      Mm.

   4. DKDaniel Kahneman31:22

      ... 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...

   5. LFLex Fridman31:41

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

   6. DKDaniel Kahneman31:50

      To solve it.

   7. LFLex Fridman31:51

      ... to solve all those problems.

   8. DKDaniel Kahneman31:53

      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.

   9. LFLex Fridman32:12

      Yeah.

   10. DKDaniel Kahneman32:12

       And AlphaZero certainly doesn't need people.

   11. LFLex Fridman32:16

       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-

   12. DKDaniel Kahneman32:24

       Let me-

   13. LFLex Fridman32:24

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

   14. DKDaniel Kahneman32:30

       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-

   15. LFLex Fridman32:41

       Yes.

   16. DKDaniel Kahneman32:42

       ... to solve the problem.

   17. LFLex Fridman32:42

       And, and that's surprising to people.

   18. DKDaniel Kahneman32:43

       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.

   19. LFLex Fridman33:03

       Right.

   20. DKDaniel Kahneman33:04

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

   21. LFLex Fridman33:16

       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-

   22. DKDaniel Kahneman33:41

       Yeah.

   23. LFLex Fridman33:42

       ... ba- basically on how hard it is for them to do the task.

   24. DKDaniel Kahneman33:45

       Yeah.

   25. LFLex Fridman33:46

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

   26. DKDaniel Kahneman33:54

       Well-

   27. LFLex Fridman33:54

       ... 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.

   28. DKDaniel Kahneman34:04

       Yeah.

   29. LFLex Fridman34:04

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

   30. DKDaniel Kahneman34:15

       Well...
10. 37:20 – 40:05

    Explainability, trust, and the role of stories in human judgment

    1. DKDaniel Kahneman37:20

       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-

    2. LFLex Fridman37:45

       Right.

    3. DKDaniel Kahneman37:45

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

    4. LFLex Fridman37:56

       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-

    5. DKDaniel Kahneman38:10

       No. But-

    6. LFLex Fridman38:11

       ... themselves?

    7. DKDaniel Kahneman38:12

       ... uh, I mean, there is a very interesting, uh, aspect of that. Humans think they can explain themselves.

    8. LFLex Fridman38:21

       Right.

    9. DKDaniel Kahneman38:22

       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.

    10. LFLex Fridman38:58

        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.

    11. DKDaniel Kahneman39:21

        Yeah.

    12. LFLex Fridman39:22

        (laughs)

    13. DKDaniel Kahneman39:22

        Absolutely.

    14. LFLex Fridman39:23

        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.

    15. DKDaniel Kahneman39:32

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

    16. LFLex Fridman39:38

        Right. Sure.

    17. DKDaniel Kahneman39:38

        I mean, so... but, but the objective-

    18. LFLex Fridman39:42

        Brilliant.

    19. DKDaniel Kahneman39:42

        ... 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.

11. 40:05 – 51:59

    Two selves: experienced vs remembering, and why time disappears in memory

    1. LFLex Fridman40:05

       That's quite brilliant, actually. Uh, but so on the, 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?

    2. DKDaniel Kahneman40:26

       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.

    3. LFLex Fridman42:04

       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-

    4. DKDaniel Kahneman42:12

       Well, I mean, the-

    5. LFLex Fridman42:13

       ... the remembering self?

    6. DKDaniel Kahneman42:14

       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.

    7. LFLex Fridman42:33

       Hmm.

    8. DKDaniel Kahneman42:33

       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.

    9. LFLex Fridman43:18

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

    10. DKDaniel Kahneman43:28

        Well-

    11. LFLex Fridman43:28

        ... based on such properties, what- what's the optimal...

    12. DKDaniel Kahneman43:32

        Well, you know, I gave up... I abandoned happiness research because I couldn't solve that problem.

    13. LFLex Fridman43:38

        Yeah.

    14. DKDaniel Kahneman43:38

        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-

    15. LFLex Fridman44:47

        And you've talked about actually-

    16. DKDaniel Kahneman44:48

        ... difficult issues.

    17. LFLex Fridman44:49

        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-

    18. DKDaniel Kahneman45:32

        I-

    19. LFLex Fridman45:32

        ... magnification much?

    20. DKDaniel Kahneman45:34

        You know, I'm too old for social networks. I, you know, I- I've never seen Instagram, so-

    21. LFLex Fridman45:41

        (laughs)

    22. DKDaniel Kahneman45:42

        ... I cannot really speak intelligently about those things. I'm just too old.

    23. LFLex Fridman45:46

        But it's interesting to watch the exact effects you described?

    24. DKDaniel Kahneman45:49

        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."

    25. LFLex Fridman46:21

        Yeah.

    26. DKDaniel Kahneman46:21

        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-

    27. LFLex Fridman46:53

        (laughs)

    28. DKDaniel Kahneman46:53

        ... but it's, uh-

    29. LFLex Fridman46:55

        Well, I'll tell you-

    30. DKDaniel Kahneman46:55

        ... I wish I knew.
12. 51:59 – 1:01:07

    Meaning, purpose, and why people rarely change their minds

    1. LFLex Fridman51:59

       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...?

    2. DKDaniel Kahneman52:28

       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-

    3. LFLex Fridman52:56

       Yeah.

    4. DKDaniel Kahneman52:56

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

    5. LFLex Fridman53:04

       Purpose is, yeah.

    6. DKDaniel Kahneman53:06

       ... 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-

    7. LFLex Fridman53:41

       So, d- you don't think seeking purpose is a significant driver in our behavior?

    8. DKDaniel Kahneman53:46

       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.

    9. LFLex Fridman54:44

       Jumping around a little bit, 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?

    10. DKDaniel Kahneman55:19

        Yeah, of course.

    11. LFLex Fridman55:20

        Uh, so when we look at driving, lighting variation destroys-

    12. DKDaniel Kahneman55:26

        Yeah. I- yeah.

    13. LFLex Fridman55:27

        ... 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?

    14. DKDaniel Kahneman55:53

        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-

    15. LFLex Fridman56:28

        Right.

    16. DKDaniel Kahneman56:28

        ... and something that's real. And others are just running experiments.

    17. LFLex Fridman56:33

        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?

    18. DKDaniel Kahneman56:46

        Yeah.

    19. LFLex Fridman56:46

        You basically use your intuition to build up...

    20. DKDaniel Kahneman56:49

        Yeah, but I mean, you know, it's- it's very skilled intuition.

    21. LFLex Fridman56:53

        Right. Absolutely, absolutely.

    22. DKDaniel Kahneman56:53

        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.

    23. LFLex Fridman57:08

        Mm-hmm.

    24. DKDaniel Kahneman57:08

        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.

    25. LFLex Fridman57:31

        This is something- there's something here.

    26. DKDaniel Kahneman57:32

        And that's- that's part of the skill.

    27. LFLex Fridman57:34

        Is that something you can reduce to words in describing a process in- in the form of advice to others?

    28. DKDaniel Kahneman57:42

        No. No.

    29. LFLex Fridman57:44

        Follow your heart, essentially? (laughs)

    30. DKDaniel Kahneman57:46

        I mean, you know, it's- it's like trying to explain what it's like to drive. It's not-
13. 1:01:07 – 1:12:59

    Replication crisis and weak effects: why psychology studies fail and what improves them

    1. LFLex Fridman1:01:07

       Jumping around on, on the psychology 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-

    2. DKDaniel Kahneman1:01:29

       You're being polite. Uh, it's worse than that, but... (laughs)

    3. LFLex Fridman1:01:33

       Is it?

    4. DKDaniel Kahneman1:01:33

       Yeah.

    5. LFLex Fridman1:01:33

       So I'm actually not fully familiar-

    6. DKDaniel Kahneman1:01:36

       Well, I mean-

    7. LFLex Fridman1:01:36

       ... to the degree how bad it is, right? So, what do you think is the source? Where do you think?

    8. DKDaniel Kahneman1:01:41

       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-

    9. LFLex Fridman1:02:02

       Right.

    10. DKDaniel Kahneman1:02:02

        ... 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.

    11. LFLex Fridman1:02:35

        Okay.

    12. DKDaniel Kahneman1:02:35

        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.

    13. LFLex Fridman1:03:03

        Mm-hmm.

    14. DKDaniel Kahneman1:03:03

        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-

    15. LFLex Fridman1:03:43

        Mm-hmm.

    16. DKDaniel Kahneman1:03:44

        ... 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-

    17. LFLex Fridman1:04:23

        Yeah.

    18. DKDaniel Kahneman1:04:23

        ... 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-

    19. LFLex Fridman1:04:51

        Yep.

    20. DKDaniel Kahneman1:04:52

        ... by 20 different teams of people with a very precise objective of changing the number of time that people go to the gym, but-

    21. LFLex Fridman1:05:01

        Mm-hmm.

    22. DKDaniel Kahneman1:05:01

        ... you know, so... And, and the success rate was zero.

    23. LFLex Fridman1:05:10

        They're not successful?

    24. DKDaniel Kahneman1:05:10

        Not one of the 53 studies worked.

    25. LFLex Fridman1:05:13

        (laughs)

    26. DKDaniel Kahneman1:05:13

        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-

    27. LFLex Fridman1:05:30

        It's weak.

    28. DKDaniel Kahneman1:05:30

        ... you are focusing on, on your manipulation, and it feels powerful to you.

    29. LFLex Fridman1:05:36

        Mm-hmm.

    30. DKDaniel Kahneman1:05:37

        There's a thing that I've written about that's called the focusing illusion.
14. 1:12:59 – 1:18:35

    Testing intelligence: beyond the Turing test toward wit, metaphor, and generality

    1. LFLex Fridman1:12:59

       ... changes with them. So, there's a guy named Alan Turing, came up with the Turing test.

    2. DKDaniel Kahneman1:13:05

       Yeah.

    3. LFLex Fridman1:13:07

       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?

    4. DKDaniel Kahneman1:13:23

       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.

    5. LFLex Fridman1:13:35

       Yes.

    6. DKDaniel Kahneman1:13:35

       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.

    7. LFLex Fridman1:14:22

       So, a test-

    8. DKDaniel Kahneman1:14:22

       So, we're quite far from, from something that can, in every domain, think like a human, except better.

    9. LFLex Fridman1:14:30

       What aspects... So, the, the Turing test has been criticized, this natural language conversation-

    10. DKDaniel Kahneman1:14:36

        Yeah.

    11. LFLex Fridman1:14:36

        ... 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?

    12. DKDaniel Kahneman1:14:55

        Yeah. I mean, certainly wit would-

    13. LFLex Fridman1:14:57

        Wit.

    14. DKDaniel Kahneman1:14:57

        ... 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.

    15. LFLex Fridman1:15:34

        Does the possibility of creating an, a human-level intelligence or superhuman-level intelligence system excite you? Scare you?

    16. DKDaniel Kahneman1:15:44

        Well, I mean, you know, I'm, uh-

    17. LFLex Fridman1:15:44

        How does it make you feel?

    18. DKDaniel Kahneman1:15:45

        I find the whole thing fascinating, absolutely fascinating.

    19. LFLex Fridman1:15:51

        So, exciting?

    20. DKDaniel Kahneman1:15:52

        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.

    21. LFLex Fridman1:16:11

        (laughs)

    22. DKDaniel Kahneman1:16:11

        Uh, we really have no idea what it will look like 30 years from now. No idea.

    23. LFLex Fridman1:16:18

        Speaking of silly, bordering on the profound, let me ask the question of, in your view, what is the meaning of it all?

    24. DKDaniel Kahneman1:16:27

        (laughs)

    25. LFLex Fridman1:16:28

        The meaning of life?

    26. DKDaniel Kahneman1:16:30

        Uh-

    27. LFLex Fridman1:16:30

        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-

    28. DKDaniel Kahneman1:16:46

        Uh-

    29. LFLex Fridman1:16:46

        ... just a beautiful mess?

    30. DKDaniel Kahneman1:16:49

        There is no answer that, that I can understand. Uh, and I'm not, and I'm not actively looking for one, um, bec-

Episode duration: 1:18:40