Daniel Kahneman: Thinking Fast and Slow, Deep Learning, and AI
技术与编程心理与人性AI 与机器学习生物与进化音乐与艺术
📋 章节目录
暂无章节信息
🔑 关键词
humandongoingdoesnexperiencemachineselfsystemsexperimentsinterestingimportantlearningthinkingproblemspsychologyintelligencestoryapptalkgot
💬 精彩语录
"things about human nature. IA Well, I mean, what it brings out is the loyalty among soldiers. I mean,"
关于人性的事情。 IA 嗯，我的意思是，它带来的是士兵之间的忠诚。我是说，
— Daniel Kahneman （07:39.760）
"captures something that's important and something that's real and others are just running experiments."
捕捉到一些重要的东西和一些真实的东西，而其他人只是在进行实验。
— Daniel Kahneman （56:25.840）
"that donated money is used to maximum effectiveness. When you get Cash App from the App Store or Google"
使捐赠的资金得到最大程度的利用。当您从 App Store 或 Google 获取 Cash App 时
— Daniel Kahneman （02:11.360）
"so that you treat them not as people anymore, but as animals. And the same way that you can slaughter"
这样你就不再把他们当作人，而是当作动物了。就像你可以屠杀一样
— Daniel Kahneman （04:23.200）
"system two is that there is mental effort involved, and there is a limited capacity for mental effort,"
系统二是涉及脑力劳动，而脑力劳动的能力是有限的，
— Daniel Kahneman （10:04.560）
🎙️ 完整对话（732 条）
Lex Fridman (00:00.000)
The following is a conversation with Daniel Kahneman, winner of the Nobel Prize in Economics
以下是与诺贝尔经济学奖获得者丹尼尔·卡尼曼的对话
Lex Fridman (00:05.680)
for his integration of economic science with the psychology of human behavior,
表彰他将经济科学与人类行为心理学相结合，
Lex Fridman (00:10.080)
judgment, and decision making. He's the author of the popular book Thinking Fast and Slow that
判断和决策。他是畅销书《思考快与慢》的作者
Lex Fridman (00:16.240)
summarizes in an accessible way his research of several decades, often in collaboration with
以通俗易懂的方式总结了他几十年来的研究，通常是与
Lex Fridman (00:22.160)
Amos Tversky on cognitive biases, prospect theory, and happiness. The central thesis of this work
阿莫斯·特沃斯基关于认知偏差、前景理论和幸福。本著作的中心论点
Daniel Kahneman (00:29.600)
is the dichotomy between two modes of thought. What he calls system one is fast, instinctive,
是两种思维模式之间的二分法。他所说的系统一是快速的、本能的，
Lex Fridman (00:35.520)
and emotional. System two is slower, more deliberative, and more logical. The book
和情感。系统二更慢、更审慎、更合乎逻辑。这本书
Daniel Kahneman (00:41.440)
delineates cognitive biases associated with each of these two types of thinking.
描述了与这两种思维相关的认知偏差。
Daniel Kahneman (00:46.960)
His study of the human mind and its peculiar and fascinating limitations are both instructive and
他对人类思维及其独特而迷人的局限性的研究既具有启发性又具有启发性。
Daniel Kahneman (00:53.040)
inspiring for those of us seeking to engineer intelligent systems. This is the Artificial
对于我们这些寻求设计智能系统的人来说是鼓舞人心的。这就是人工
Daniel Kahneman (00:59.200)
Intelligence Podcast. If you enjoy it, subscribe on YouTube, give it five stars on Apple Podcast,
情报播客。如果您喜欢它，请在 YouTube 上订阅，在 Apple Podcast 上给它五颗星，
Daniel Kahneman (01:05.120)
follow on Spotify, support it on Patreon, or simply connect with me on Twitter at
在 Spotify 上关注，在 Patreon 上支持它，或者直接在 Twitter 上与我联系：
Daniel Kahneman (01:10.000)
Lex Friedman spelled F R I D M A N. I recently started doing ads at the end of the introduction.
Lex Friedman 拼写为 F R I D M A N。我最近开始在介绍末尾做广告。
Daniel Kahneman (01:16.800)
I'll do one or two minutes after introducing the episode and never any ads in the middle
我会在介绍剧集后播放一两分钟，中间不会出现任何广告
Daniel Kahneman (01:21.280)
that can break the flow of the conversation. I hope that works for you and doesn't hurt the
这可能会破坏谈话的流畅性。我希望这对你有用并且不会伤害
Daniel Kahneman (01:25.920)
listening experience. This show is presented by Cash App, the number one finance app in the App
聆听体验。本节目由App内排名第一的金融类App Cash App倾情呈现
Daniel Kahneman (01:32.160)
Store. I personally use Cash App to send money to friends, but you can also use it to buy, sell,
店铺。我个人使用 Cash App 向朋友汇款，但您也可以用它来购买、出售、
Lex Fridman (01:37.440)
and deposit Bitcoin in just seconds. Cash App also has a new investing feature. You can buy
只需几秒钟即可存入比特币。 Cash App还有一个新的投资功能。你可以购买
Daniel Kahneman (01:43.280)
fractions of a stock, say one dollar's worth, no matter what the stock price is. Broker services
股票的一小部分，比如价值一美元的股票，无论股票价格是多少。经纪服务
Daniel Kahneman (01:48.640)
are provided by Cash App Investing, a subsidiary of Square and member SIPC. I'm excited to be
由 Square 子公司、SIPC 成员 Cash App Investing 提供。我很高兴成为
Daniel Kahneman (01:55.280)
working with Cash App to support one of my favorite organizations called First, best known
Daniel Kahneman (02:00.480)
for their FIRST Robotics and Lego competitions. They educate and inspire hundreds of thousands
Daniel Kahneman (02:05.760)
of students in over 110 countries and have a perfect rating at Charity Navigator, which means
Daniel Kahneman (02:11.360)
that donated money is used to maximum effectiveness. When you get Cash App from the App Store or Google
Daniel Kahneman (02:17.120)
Play and use code LEXPODCAST, you'll get $10 and Cash App will also donate $10 to FIRST,
Daniel Kahneman (02:24.480)
which again is an organization that I've personally seen inspire girls and boys to dream
Daniel Kahneman (02:29.920)
of engineering a better world. And now here's my conversation with Daniel Kahneman.
Daniel Kahneman (02:36.800)
You tell a story of an SS soldier early in the war, World War II, in Nazi occupied France in
Daniel Kahneman (02:43.600)
Paris, where you grew up. He picked you up and hugged you and showed you a picture of a boy,
Daniel Kahneman (02:50.160)
Daniel Kahneman. Maybe not realizing that you were Jewish.
Lex Fridman (02:53.840)
Not maybe, certainly not.
Lex Fridman (02:56.400)
So I told you I'm from the Soviet Union that was significantly impacted by the war as well,
Lex Fridman (03:01.360)
and I'm Jewish as well. What do you think World War II taught us about human psychology broadly?
Daniel Kahneman (03:09.680)
Well, I think the only big surprise is the extermination policy, genocide,
Daniel Kahneman (03:17.520)
by the German people. That's when you look back on it, and I think that's a major surprise.
Daniel Kahneman (03:27.040)
It's a surprise because...
Lex Fridman (03:28.240)
It's a surprise that they could do it. It's a surprise that enough people
Daniel Kahneman (03:34.720)
willingly participated in that. This is a surprise. Now it's no longer a surprise,
Lex Fridman (03:41.520)
but it's changed many people's views, I think, about human beings. Certainly for me,
Daniel Kahneman (03:50.720)
the Ackman trial, that teaches you something because it's very clear that if it could happen
Lex Fridman (03:58.080)
in Germany, it could happen anywhere. It's not that the Germans were special.
Daniel Kahneman (04:04.080)
This could happen anywhere.
Lex Fridman (04:05.280)
So what do you think that is? Do you think we're all capable of evil? We're all capable of cruelty?
Daniel Kahneman (04:13.600)
I don't think in those terms. I think that what is certainly possible is you can dehumanize people
Lex Fridman (04:23.200)
so that you treat them not as people anymore, but as animals. And the same way that you can slaughter
Daniel Kahneman (04:32.480)
animals without feeling much of anything, it can be the same. And when you feel that,
Daniel Kahneman (04:41.120)
I think, the combination of dehumanizing the other side and having uncontrolled power over
Daniel Kahneman (04:49.360)
other people, I think that doesn't bring out the most generous aspect of human nature.
Lex Fridman (04:54.560)
So that Nazi soldier, he was a good man. And he was perfectly capable of killing a lot of people,
Lex Fridman (05:08.480)
and I'm sure he did.
Lex Fridman (05:10.080)
But what did the Jewish people mean to Nazis? So what the dismissal of Jewish as worthy of?
Daniel Kahneman (05:20.160)
IA Again, this is surprising that it was so extreme,
Lex Fridman (05:25.120)
but it's not one thing in human nature. I don't want to call it evil, but the distinction between
Daniel Kahneman (05:32.480)
the in group and the out group, that is very basic. So that's built in. The loyalty and
Daniel Kahneman (05:40.160)
affection towards in group and the willingness to dehumanize the out group, that is in human nature.
Daniel Kahneman (05:50.320)
That's what I think probably didn't need the Holocaust to teach us that. But the Holocaust is
Lex Fridman (05:57.920)
a very sharp lesson of what can happen to people and what people can do.
Daniel Kahneman (06:05.120)
SL. So the effect of the in group and the out group. IA It's clear. Those were people,
Daniel Kahneman (06:13.600)
you could shoot them. They were not human. There was no empathy, or very, very little empathy left.
Lex Fridman (06:23.680)
So occasionally, there might have been. And very quickly, by the way, the empathy disappeared,
Lex Fridman (06:32.720)
if there was initially. And the fact that everybody around you was doing it,
Daniel Kahneman (06:39.840)
that completely, the group doing it, and everybody shooting Jews, I think that makes it permissible.
Daniel Kahneman (06:51.120)
Now, how much, whether it could happen in every culture, or whether the Germans were just
Daniel Kahneman (07:01.280)
particularly efficient and disciplined, so they could get away with it. It's an interesting
Daniel Kahneman (07:10.000)
question. SL. Are these artifacts of history or is it human nature? IA I think that's really human
Daniel Kahneman (07:15.360)
nature. You put some people in a position of power relative to other people, and then they become
Daniel Kahneman (07:24.480)
less human, they become different. SL. But in general, in war, outside of concentration camps
Daniel Kahneman (07:32.240)
in World War Two, it seems that war brings out darker sides of human nature, but also the beautiful
Daniel Kahneman (07:39.760)
things about human nature. IA Well, I mean, what it brings out is the loyalty among soldiers. I mean,
Daniel Kahneman (07:49.120)
it brings out the bonding, male bonding, I think is a very real thing that happens. And there is
Daniel Kahneman (07:57.920)
a certain thrill to friendship, and there is certainly a certain thrill to friendship under
Daniel Kahneman (08:03.840)
risk and to shared risk. And so people have very profound emotions, up to the point where it gets
Lex Fridman (08:12.400)
so traumatic that little is left. SL. So let's talk about psychology a little bit. In your book,
Daniel Kahneman (08:23.040)
Thinking Fast and Slow, you describe two modes of thought, system one, the fast and instinctive,
Lex Fridman (08:31.200)
and emotional one, and system two, the slower, deliberate, logical one. At the risk of asking
Daniel Kahneman (08:37.360)
Darwin to discuss theory of evolution, can you describe distinguishing characteristics for people
Daniel Kahneman (08:46.320)
who have not read your book of the two systems? IA Well, I mean, the word system is a bit
Daniel Kahneman (08:52.800)
misleading, but at the same time it's misleading, it's also very useful. But what I call system one,
Daniel Kahneman (09:01.440)
it's easier to think of it as a family of activities. And primarily, the way I describe it
Daniel Kahneman (09:09.120)
is there are different ways for ideas to come to mind. And some ideas come to mind automatically,
Lex Fridman (09:17.920)
and the standard example is two plus two, and then something happens to you. And in other cases,
Daniel Kahneman (09:26.480)
you've got to do something, you've got to work in order to produce the idea. And my example,
Daniel Kahneman (09:32.240)
I always give the same pair of numbers as 27 times 14, I think. SL. You have to perform some
Daniel Kahneman (09:38.000)
algorithm in your head, some steps. IA Yes, and it takes time. It's a very difference. Nothing
Daniel Kahneman (09:44.560)
comes to mind except something comes to mind, which is the algorithm, I mean, that you've got
Daniel Kahneman (09:50.640)
to perform. And then it's work, and it engages short term memory, it engages executive function,
Lex Fridman (09:58.000)
and it makes you incapable of doing other things at the same time. So the main characteristic of
Daniel Kahneman (10:04.560)
system two is that there is mental effort involved, and there is a limited capacity for mental effort,
Lex Fridman (10:10.960)
whereas system one is effortless, essentially. That's the major distinction.
Daniel Kahneman (10:15.600)
SL. So you talk about there, you know, it's really convenient to talk about two systems,
Lex Fridman (10:21.040)
but you also mentioned just now and in general that there's no distinct two systems in the brain
Daniel Kahneman (10:29.120)
from a neurobiological, even from a psychology perspective. But why does it seem to, from the
Daniel Kahneman (10:36.240)
experiments you've conducted, there does seem to be kind of emergent two modes of thinking? So
Daniel Kahneman (10:47.120)
at some point, these kinds of systems came into a brain architecture. Maybe mammals share it.
Daniel Kahneman (10:57.440)
Or do you not think of it at all in those terms that it's all a mush and these two things just
Daniel Kahneman (11:01.520)
emerge? RL. Evolutionary theorizing about this is cheap and easy. So it's the way I think about it
Daniel Kahneman (11:12.560)
is that it's very clear that animals have perceptual system, and that includes an ability
Daniel Kahneman (11:20.720)
to understand the world, at least to the extent that they can predict, they can't explain anything,
Lex Fridman (11:27.120)
but they can anticipate what's going to happen. And that's a key form of understanding the world.
Lex Fridman (11:34.720)
And my crude idea is that what I call system two, well, system two grew out of this.
Daniel Kahneman (11:45.200)
And, you know, there is language and there is the capacity of manipulating ideas and the capacity
Daniel Kahneman (11:51.840)
of imagining futures and of imagining counterfactual things that haven't happened
Lex Fridman (11:58.240)
and to do conditional thinking. And there are really a lot of abilities that without language
Lex Fridman (12:06.240)
and without the very large brain that we have compared to others would be impossible. Now,
Daniel Kahneman (12:13.760)
system one is more like what the animals are, but system one also can talk. I mean,
Daniel Kahneman (12:20.960)
it has language. It understands language. Indeed, it speaks for us. I mean, you know,
Daniel Kahneman (12:26.480)
I'm not choosing every word as a deliberate process. The words, I have some idea and then
Daniel Kahneman (12:32.800)
the words come out and that's automatic and effortless. And many of the experiments you've
Daniel Kahneman (12:39.040)
done is to show that, listen, system one exists and it does speak for us and we should be careful
Daniel Kahneman (12:44.480)
about the voice it provides. Well, I mean, you know, we have to trust it because it's
Daniel Kahneman (12:55.280)
the speed at which it acts. System two, if we're dependent on system two for survival,
Daniel Kahneman (13:01.760)
we wouldn't survive very long because it's very slow. Yeah. Crossing the street.
Daniel Kahneman (13:06.480)
Crossing the street. I mean, many things depend on their being automatic. One very important aspect
Daniel Kahneman (13:12.560)
of system one is that it's not instinctive. You use the word instinctive. It contains skills that
Daniel Kahneman (13:20.320)
clearly have been learned. So that skilled behavior like driving a car or speaking, in fact,
Daniel Kahneman (13:28.800)
skilled behavior has to be learned. And so it doesn't, you know, you don't come equipped with
Daniel Kahneman (13:35.920)
driving. You have to learn how to drive and you have to go through a period where driving is not
Daniel Kahneman (13:41.840)
automatic before it becomes automatic. So. Yeah. You construct, I mean, this is where you talk
Daniel Kahneman (13:48.880)
about heuristic and biases is you, to make it automatic, you create a pattern and then system
Daniel Kahneman (13:57.360)
one essentially matches a new experience against the previously seen pattern. And when that match
Daniel Kahneman (14:02.960)
is not a good one, that's when the cognitive, all the mess happens, but it's most of the time
Daniel Kahneman (14:08.160)
it works. And so it's pretty. Most of the time, the anticipation of what's going to happen next
Daniel Kahneman (14:13.840)
is correct. And most of the time the plan about what you have to do is correct. And so most of
Daniel Kahneman (14:22.000)
the time everything works just fine. What's interesting actually is that in some sense,
Daniel Kahneman (14:29.040)
system one is much better at what it does than system two is at what it does. That is there is
Daniel Kahneman (14:36.240)
that quality of effortlessly solving enormously complicated problems, which clearly exists so
Daniel Kahneman (14:44.480)
that the chess player, a very good chess player, all the moves that come to their mind are strong
Daniel Kahneman (14:52.160)
moves. So all the selection of strong moves happens unconsciously and automatically and
Lex Fridman (14:58.960)
very, very fast. And all that is in system one. So system two verifies.
Lex Fridman (15:07.280)
So along this line of thinking, really what we are are machines that construct
Daniel Kahneman (15:12.480)
a pretty effective system one. You could think of it that way. So we're not talking about humans,
Lex Fridman (15:19.360)
but if we think about building artificial intelligence systems, robots, do you think
Lex Fridman (15:26.400)
all the features and bugs that you have highlighted in human beings are useful
Lex Fridman (15:32.480)
for constructing AI systems? So both systems are useful for perhaps instilling in robots?
Lex Fridman (15:39.280)
What is happening these days is that actually what is happening in deep learning is more like
Daniel Kahneman (15:50.320)
a system one product than like a system two product. I mean, deep learning matches patterns
Lex Fridman (15:57.120)
and anticipate what's going to happen. So it's highly predictive. What deep learning
Daniel Kahneman (16:05.120)
doesn't have and many people think that this is the critical, it doesn't have the ability to
Daniel Kahneman (16:12.000)
reason. So there is no system two there. But I think very importantly, it doesn't have any
Daniel Kahneman (16:19.040)
causality or any way to represent meaning and to represent real interactions. So until that is
Lex Fridman (16:27.520)
solved, what can be accomplished is marvelous and very exciting, but limited.
Daniel Kahneman (16:35.600)
That's actually really nice to think of current advances in machine learning as essentially
Daniel Kahneman (16:40.560)
system one advances. So how far can we get with just system one? If we think of deep learning
Daniel Kahneman (16:46.960)
in artificial intelligence systems? I mean, you know, it's very clear that deep mind has already
Daniel Kahneman (16:52.320)
gone way beyond what people thought was possible. I think the thing that has impressed me most about
Daniel Kahneman (17:00.560)
the developments in AI is the speed. It's that things, at least in the context of deep learning,
Lex Fridman (17:07.840)
and maybe this is about to slow down, but things moved a lot faster than anticipated.
Daniel Kahneman (17:14.400)
The transition from solving chess to solving Go, that's bewildering how quickly it went.
Daniel Kahneman (17:25.600)
The move from Alpha Go to Alpha Zero is sort of bewildering the speed at which they accomplished
Daniel Kahneman (17:31.840)
that. Now, clearly, there are many problems that you can solve that way, but there are some problems
Lex Fridman (17:41.360)
for which you need something else. Something like reasoning.
Daniel Kahneman (17:45.760)
Well, reasoning and also, you know, one of the real mysteries, psychologist Gary Marcus, who is
Daniel Kahneman (17:54.160)
also a critic of AI. I mean, what he points out, and I think he has a point, is that humans learn
Daniel Kahneman (18:05.920)
quickly. Children don't need a million examples, they need two or three examples. So, clearly,
Daniel Kahneman (18:16.000)
there is a fundamental difference. And what enables a machine to learn quickly, what you have
Daniel Kahneman (18:25.280)
to build into the machine, because it's clear that you have to build some expectations or
Daniel Kahneman (18:30.400)
or something in the machine to make it ready to learn quickly. That at the moment seems to be
Daniel Kahneman (18:38.320)
unsolved. I'm pretty sure that DeepMind is working on it, but if they have solved it, I haven't heard
Daniel Kahneman (18:47.680)
yet. They're trying to actually, them and OpenAI are trying to start to get to use neural networks
Daniel Kahneman (18:54.640)
to reason. So, assemble knowledge. Of course, causality is, temporal causality, is out of
Daniel Kahneman (19:02.960)
reach to most everybody. You mentioned the benefits of System 1 is essentially that it's
Daniel Kahneman (19:09.200)
fast, allows us to function in the world.
Lex Fridman (19:10.960)
Fast and skilled, yeah.
Daniel Kahneman (19:13.040)
It's skill.
Lex Fridman (19:13.680)
And it has a model of the world. You know, in a sense, I mean, there was the early phase of
Daniel Kahneman (19:19.920)
AI attempted to model reasoning. And they were moderately successful, but, you know, reasoning
Daniel Kahneman (19:29.440)
by itself doesn't get you much. Deep learning has been much more successful in terms of, you know,
Lex Fridman (19:37.440)
what they can do. But now, it's an interesting question, whether it's approaching its limits.
Lex Fridman (19:43.920)
What do you think?
Daniel Kahneman (19:44.640)
I think absolutely. So, I just talked to Gian LeCun. He mentioned, you know, so he thinks
Daniel Kahneman (19:51.840)
that the limits, we're not going to hit the limits with neural networks, that ultimately,
Daniel Kahneman (19:57.840)
this kind of System 1 pattern matching will start to look like System 2 without significant
Daniel Kahneman (1:00:06.660)
I'm working with. Otherwise, I mean, there is that kind of collaboration, which is like
Daniel Kahneman (1:00:13.280)
an exchange, a commercial exchange of giving this, you give me that. But the real ones
Daniel Kahneman (1:00:21.880)
are interpersonal. They're between people who like each other and who like making each
Daniel Kahneman (1:00:28.080)
other think and who like the way that the other person responds to your thoughts. You
Lex Fridman (1:00:34.400)
have to be lucky.
Lex Fridman (1:00:37.080)
But I already noticed that even just me showing up here, you've quickly started to digging
Daniel Kahneman (1:00:43.760)
in on a particular problem I'm working on and already new information started to emerge.
Lex Fridman (1:00:49.840)
Is that a process, just the process of curiosity of talking to people about problems and seeing?
Daniel Kahneman (1:00:56.420)
I'm curious about anything to do with AI and robotics. And I knew you were dealing with
Daniel Kahneman (1:01:03.400)
that. So I was curious.
Daniel Kahneman (1:01:05.240)
Just follow your curiosity. Jumping around on the psychology front, the dramatic sounding
Daniel Kahneman (1:01:13.100)
terminology of replication crisis, but really just the, at times, this effect that at times
Lex Fridman (1:01:24.960)
studies do not, are not fully generalizable. They don't.
Daniel Kahneman (1:01:29.240)
You are being polite. It's worse than that.
Daniel Kahneman (1:01:33.040)
Is it? So I'm actually not fully familiar to the degree how bad it is, right? So what
Lex Fridman (1:01:39.360)
do you think is the source? Where do you think?
Daniel Kahneman (1:01:41.520)
I think I know what's going on actually. I mean, I have a theory about what's going on
Lex Fridman (1:01:47.520)
and what's going on is that there is, first of all, a very important distinction between
Daniel Kahneman (1:01:55.460)
two types of experiments. And one type is within subject. So it's the same person has
Daniel Kahneman (1:02:03.120)
two experimental conditions. And the other type is between subjects where some people
Daniel Kahneman (1:02:09.200)
are this condition, other people are that condition. They're different worlds. And between
Daniel Kahneman (1:02:14.160)
subject experiments are much harder to predict and much harder to anticipate. And the reason,
Lex Fridman (1:02:25.560)
and they're also more expensive because you need more people. And it's just, so between
Daniel Kahneman (1:02:31.880)
subject experiments is where the problem is. It's not so much in within subject experiments,
Daniel Kahneman (1:02:38.600)
it's really between. And there is a very good reason why the intuitions of researchers about
Daniel Kahneman (1:02:46.920)
between subject experiments are wrong. And that's because when you are a researcher,
Daniel Kahneman (1:02:54.180)
you're in a within subject situation. That is you are imagining the two conditions and
Daniel Kahneman (1:03:00.560)
you see the causality and you feel it. But in the between subject condition, they live
Daniel Kahneman (1:03:09.680)
in one condition and the other one is just nowhere. So our intuitions are very weak about
Daniel Kahneman (1:03:18.440)
between subject experiments. And that I think is something that people haven't realized.
Lex Fridman (1:03:26.520)
And in addition, because of that, we have no idea about the power of manipulations of
Daniel Kahneman (1:03:34.800)
experimental manipulations because the same manipulation is much more powerful when you
Daniel Kahneman (1:03:42.420)
are in the two conditions than when you live in only one condition. And so the experimenters
Daniel Kahneman (1:03:48.880)
have very poor intuitions about between subject experiments. And there is something else which
Daniel Kahneman (1:03:56.760)
is very important, I think, which is that almost all psychological hypotheses are true.
Daniel Kahneman (1:04:04.080)
That is in the sense that, you know, directionally, if you have a hypothesis that A really causes
Daniel Kahneman (1:04:13.200)
B, that it's not true that A causes the opposite of B. Maybe A just has very little effect,
Lex Fridman (1:04:21.000)
but hypotheses are true mostly, except mostly they're very weak. They're much weaker than
Daniel Kahneman (1:04:28.840)
you think when you are having images. So the reason I'm excited about that is that I recently
Daniel Kahneman (1:04:38.000)
heard about some friends of mine who they essentially funded 53 studies of behavioral
Daniel Kahneman (1:04:50.560)
change by 20 different teams of people with a very precise objective of changing the number
Daniel Kahneman (1:04:59.420)
of times that people go to the gym. And the success rate was zero. Not one of the 53 studies
Daniel Kahneman (1:05:12.600)
worked. Now, what's interesting about that is those are the best people in the field
Lex Fridman (1:05:18.160)
and they have no idea what's going on. So they're not calibrated. They think that it's
Daniel Kahneman (1:05:24.440)
going to be powerful because they can imagine it, but actually it's just weak because you
Daniel Kahneman (1:05:30.760)
are focusing on your manipulation and it feels powerful to you. There's a thing that I've
Daniel Kahneman (1:05:37.880)
written about that's called the focusing illusion. That is that when you think about something,
Daniel Kahneman (1:05:43.480)
it looks very important, more important than it really is.
Daniel Kahneman (1:05:48.400)
More important than it really is. But if you don't see that effect, the 53 studies, doesn't
Lex Fridman (1:05:53.800)
that mean you just report that? So what was, I guess, the solution to that?
Daniel Kahneman (1:05:59.320)
Well, I mean, the solution is for people to trust their intuitions less or to try out
Daniel Kahneman (1:06:07.600)
their intuitions before. I mean, experiments have to be pre registered and by the time
Daniel Kahneman (1:06:14.760)
you run an experiment, you have to be committed to it and you have to run the experiment seriously
Daniel Kahneman (1:06:20.960)
enough and in a public. And so this is happening. The interesting thing is what happens before
Lex Fridman (1:06:32.800)
and how do people prepare themselves and how they run pilot experiments. It's going to
Daniel Kahneman (1:06:37.920)
train the way psychology is done and it's already happening.
Lex Fridman (1:06:41.360)
Do you have a hope for, this might connect to the study sample size.
Daniel Kahneman (1:06:48.520)
Yeah.
Lex Fridman (1:06:49.520)
Do you have a hope for the internet?
Daniel Kahneman (1:06:51.320)
Well, I mean, you know, this is really happening. MTurk, everybody's running experiments on
Lex Fridman (1:06:59.040)
MTurk and it's very cheap and very effective.
Lex Fridman (1:07:03.640)
Do you think that changes psychology essentially? Because you're thinking you cannot run 10,000
Lex Fridman (1:07:09.200)
subjects.
Daniel Kahneman (1:07:10.200)
Eventually it will. I mean, I, you know, I can't put my finger on how exactly, but it's,
Daniel Kahneman (1:07:18.480)
that's been true in psychology with whenever an important new method came in, it changes
Daniel Kahneman (1:07:24.880)
the field. So, and MTurk is really a method because it makes it very much easier to do
Lex Fridman (1:07:33.160)
something, to do some things.
Daniel Kahneman (1:07:35.520)
Is there a undergrad students who'll ask me, you know, how big a neural network should
Daniel Kahneman (1:07:40.680)
be for a particular problem? So let me ask you an equivalent question. How big, how many
Lex Fridman (1:07:49.080)
subjects does the study have for it to have a conclusive result?
Daniel Kahneman (1:07:53.560)
Well, it depends on the strength of the effect. So if you're studying visual perception or
Daniel Kahneman (1:08:00.760)
the perception of color, many of the classic results in visual, in color perception were
Daniel Kahneman (1:08:08.600)
done on three or four people. And I think one of them was colorblind, but partly colorblind,
Lex Fridman (1:08:14.600)
but on vision, you know, it's highly reliable. Many people don't need a lot of replications
Daniel Kahneman (1:08:24.820)
for some type of neurological experiment. When you're studying weaker phenomena and
Daniel Kahneman (1:08:35.800)
especially when you're studying them between subjects, then you need a lot more subjects
Daniel Kahneman (1:08:41.120)
than people have been running. And that is, that's one of the things that are happening
Daniel Kahneman (1:08:47.000)
in psychology now is that the power, the statistical power of experiments is increasing rapidly.
Lex Fridman (1:08:54.220)
Does the between subject, as the number of subjects goes to infinity approach?
Daniel Kahneman (1:08:59.200)
Well, I mean, you know, it goes to infinity is exaggerated, but people, the standard number
Daniel Kahneman (1:09:06.440)
of subjects for an experiment in psychology were 30 or 40. And for a weak effect, that's
Daniel Kahneman (1:09:15.040)
simply not enough. And you may need a couple of hundred. I mean, it's that sort of order
Lex Fridman (1:09:25.720)
of magnitude.
Lex Fridman (1:09:28.760)
What are the major disagreements in theories and effects that you've observed throughout
Daniel Kahneman (1:09:35.840)
your career that still stand today? You've worked on several fields, but what still is
Lex Fridman (1:09:42.520)
out there as a major disagreement that pops into your mind?
Daniel Kahneman (1:09:47.320)
I've had one extreme experience of, you know, controversy with somebody who really doesn't
Daniel Kahneman (1:09:54.840)
like the work that Amos Tversky and I did. And he's been after us for 30 years or more,
Lex Fridman (1:10:01.720)
at least.
Lex Fridman (1:10:02.720)
Do you want to talk about it?
Daniel Kahneman (1:10:03.720)
Well, I mean, his name is Gerd Gigerenzer. He's a well known German psychologist. And
Daniel Kahneman (1:10:10.400)
that's the one controversy, which I, it's been unpleasant. And no, I don't particularly
Lex Fridman (1:10:18.960)
want to talk about it.
Lex Fridman (1:10:21.040)
But is there is there open questions, even in your own mind, every once in a while? You
Daniel Kahneman (1:10:25.680)
know, we talked about semi autonomous vehicles. In my own mind, I see what the data says,
Lex Fridman (1:10:31.640)
but I also constantly torn. Do you have things where you or your studies have found something,
Lex Fridman (1:10:38.200)
but you're also intellectually torn about what it means? And there's maybe disagreements
Daniel Kahneman (1:10:44.800)
within your own mind about particular things.
Daniel Kahneman (1:10:47.560)
I mean, it's, you know, one of the things that are interesting is how difficult it is
Daniel Kahneman (1:10:52.280)
for people to change their mind. Essentially, you know, once they are committed, people
Daniel Kahneman (1:11:00.440)
just don't change their mind about anything that matters. And that is surprisingly, but
Daniel Kahneman (1:11:05.600)
it's true about scientists. So the controversy that I described, you know, that's been going
Daniel Kahneman (1:11:12.240)
on like 30 years and it's never going to be resolved. And you build a system and you live
Daniel Kahneman (1:11:19.000)
within that system and other other systems of ideas look foreign to you and there is
Daniel Kahneman (1:11:27.000)
very little contact and very little mutual influence. That happens a fair amount.
Lex Fridman (1:11:33.400)
Do you have a hopeful advice or message on that? Thinking about science, thinking about
Daniel Kahneman (1:11:41.000)
politics, thinking about things that have impact on this world, how can we change our
Lex Fridman (1:11:47.840)
mind?
Daniel Kahneman (1:11:49.760)
I think that, I mean, on things that matter, which are political or really political or
Daniel Kahneman (1:11:56.920)
religious and people just don't, don't change their mind. And by and large, and there's
Daniel Kahneman (1:12:04.360)
very little that you can do about it. The, what does happen is that if leaders change
Daniel Kahneman (1:12:13.360)
their minds. So for example, the public, the American public doesn't really believe in
Daniel Kahneman (1:12:19.840)
climate change, doesn't take it very seriously. But if some religious leaders decided this
Daniel Kahneman (1:12:26.920)
is a major threat to humanity, that would have a big effect. So that we have the opinions
Daniel Kahneman (1:12:34.600)
that we have, not because we know why we have them, but because we trust some people and
Daniel Kahneman (1:12:39.840)
we don't trust other people. And so it's much less about evidence than it is about stories.
Lex Fridman (1:12:49.120)
So the way, one way to change your mind isn't at the individual level, is that the leaders
Daniel Kahneman (1:12:55.040)
of the communities you look up with, the stories change and therefore your mind changes with
Daniel Kahneman (1:12:59.640)
them. So there's a guy named Alan Turing, came up with a Turing test. What do you think
Daniel Kahneman (1:13:08.400)
is a good test of intelligence? Perhaps we're drifting in a topic that we're maybe philosophizing
Daniel Kahneman (1:13:18.760)
about, but what do you think is a good test for intelligence, for an artificial intelligence
Lex Fridman (1:13:22.240)
system?
Daniel Kahneman (1:13:23.240)
Well, the standard definition of artificial general intelligence is that it can do anything
Daniel Kahneman (1:13:32.760)
that people can do and it can do them better. What we are seeing is that in many domains,
Daniel Kahneman (1:13:39.540)
you have domain specific devices or programs or software, and they beat people easily in
Daniel Kahneman (1:13:51.360)
a specified way. What we are very far from is that general ability, general purpose intelligence.
Daniel Kahneman (1:14:04.080)
In machine learning, people are approaching something more general. I mean, for Alpha
Daniel Kahneman (1:14:08.800)
Zero was much more general than Alpha Go, but it's still extraordinarily narrow and
Daniel Kahneman (1:14:18.840)
specific in what it can do. So we're quite far from something that can, in every domain,
Daniel Kahneman (1:14:28.160)
think like a human except better.
Lex Fridman (1:14:30.960)
What aspect, so the Turing test has been criticized, it's natural language conversation that is
Daniel Kahneman (1:14:36.560)
too simplistic. It's easy to quote unquote pass under constraints specified. What aspect
Daniel Kahneman (1:14:44.080)
of conversation would impress you if you heard it? Is it humor? What would impress the heck
Lex Fridman (1:14:52.120)
out of you if you saw it in conversation?
Daniel Kahneman (1:14:55.680)
Yeah, I mean, certainly wit would be impressive and humor would be more impressive than just
Daniel Kahneman (1:15:06.120)
factual conversation, which I think is easy. And allusions would be interesting and metaphors
Daniel Kahneman (1:15:17.080)
would be interesting. I mean, but new metaphors, not practiced metaphors. So there is a lot
Daniel Kahneman (1:15:25.640)
that would be sort of impressive that is completely natural in conversation, but that you really
Lex Fridman (1:15:33.160)
wouldn't expect.
Daniel Kahneman (1:15:34.160)
Does the possibility of creating a human level intelligence or superhuman level intelligence
Lex Fridman (1:15:40.440)
system excite you, scare you? How does it make you feel?
Daniel Kahneman (1:15:47.440)
I find the whole thing fascinating. Absolutely fascinating.
Lex Fridman (1:15:51.520)
So exciting.
Daniel Kahneman (1:15:52.520)
I think. And exciting. It's also terrifying, you know, but I'm not going to be around
Daniel Kahneman (1:16:00.360)
to see it. And so I'm curious about what is happening now, but I also know that predictions
Daniel Kahneman (1:16:09.200)
about it are silly. We really have no idea what it will look like 30 years from now.
Lex Fridman (1:16:16.160)
No idea.
Daniel Kahneman (1:16:18.360)
Speaking of silly, bordering on the profound, let me ask the question of, in your view,
Lex Fridman (1:16:26.480)
what is the meaning of it all? The meaning of life? He's a descendant of great apes that
Daniel Kahneman (1:16:32.400)
we are. Why, what drives us as a civilization, as a human being, as a force behind everything
Lex Fridman (1:16:40.680)
that you've observed and studied? Is there any answer or is it all just a beautiful mess?
Daniel Kahneman (1:16:49.920)
There is no answer that I can understand and I'm not, and I'm not actively looking for
Lex Fridman (1:16:58.760)
one.
Lex Fridman (1:16:59.760)
Do you think an answer exists?
Daniel Kahneman (1:17:02.160)
No. There is no answer that we can understand. I'm not qualified to speak about what we cannot
Daniel Kahneman (1:17:08.200)
understand, but there is, I know that we cannot understand reality, you know. I mean, there
Daniel Kahneman (1:17:17.400)
are a lot of things that we can do. I mean, you know, gravity waves, I mean, that's a
Daniel Kahneman (1:17:22.720)
big moment for humanity. And when you imagine that ape, you know, being able to go back
Lex Fridman (1:17:29.800)
to the Big Bang, that's, that's, but...
Lex Fridman (1:17:34.200)
But the why.
Lex Fridman (1:17:35.200)
Yeah, the why.
Daniel Kahneman (1:17:36.200)
It's bigger than us.
Lex Fridman (1:17:37.200)
The why is hopeless, really.
Daniel Kahneman (1:17:40.200)
Danny, thank you so much. It was an honor. Thank you for speaking today.
Lex Fridman (1:17:43.640)
Thank you.
Daniel Kahneman (1:17:44.640)
Thanks for listening to this conversation. And thank you to our presenting sponsor, Cash
Daniel Kahneman (1:17:49.480)
App. Download it, use code LexPodcast, you'll get $10 and $10 will go to FIRST, a STEM education
Daniel Kahneman (1:17:56.720)
nonprofit that inspires hundreds of thousands of young minds to become future leaders and
Daniel Kahneman (1:18:01.880)
innovators. If you enjoy this podcast, subscribe on YouTube, give it five stars on Apple Podcast,
Daniel Kahneman (1:18:08.280)
follow on Spotify, support it on Patreon, or simply connect with me on Twitter.
Lex Fridman (1:18:13.880)
And now, let me leave you with some words of wisdom from Daniel Kahneman.
Daniel Kahneman (1:18:19.160)
Intelligence is not only the ability to reason, it is also the ability to find relevant material
Lex Fridman (1:18:24.780)
and memory and to deploy attention when needed.
Daniel Kahneman (1:18:29.320)
Thank you for listening and hope to see you next time.
Daniel Kahneman (20:06.720)
transformation of the architecture. So, I'm more with the majority of the people who think that,
Daniel Kahneman (20:12.480)
yes, neural networks will hit a limit in their capability.
Daniel Kahneman (20:16.400)
He, on the one hand, I have heard him tell them it's a sub, it's essentially that, you know,
Lex Fridman (20:22.960)
what they have accomplished is not a big deal, that they have just touched, that basically,
Daniel Kahneman (20:28.080)
you know, they can't do unsupervised learning in an effective way. But you're telling me that he
Lex Fridman (20:35.520)
thinks that the current, within the current architecture, you can do causality and reasoning?
Daniel Kahneman (20:41.520)
So, he's very much a pragmatist in a sense that's saying that we're very far away,
Daniel Kahneman (20:47.200)
that there's still, I think there's this idea that he says is, we can only see
Lex Fridman (20:54.240)
one or two mountain peaks ahead and there might be either a few more after or
Daniel Kahneman (20:59.280)
thousands more after. Yeah, so that kind of idea.
Lex Fridman (21:01.920)
I heard that metaphor.
Daniel Kahneman (21:03.120)
Yeah, right. But nevertheless, it doesn't see the final answer not fundamentally looking like one
Lex Fridman (21:13.520)
that we currently have. So, neural networks being a huge part of that.
Daniel Kahneman (21:18.720)
Yeah, I mean, that's very likely because pattern matching is so much of what's going on.
Lex Fridman (21:26.400)
And you can think of neural networks as processing information sequentially.
Daniel Kahneman (21:30.640)
Yeah, I mean, you know, there is an important aspect to, for example, you get systems that
Daniel Kahneman (21:39.680)
translate and they do a very good job, but they really don't know what they're talking about.
Lex Fridman (21:45.760)
And for that, I'm really quite surprised. For that, you would need an AI that has sensation,
Lex Fridman (21:55.920)
an AI that is in touch with the world.
Daniel Kahneman (21:58.000)
Yes, self awareness and maybe even something resembles consciousness kind of ideas.
Daniel Kahneman (22:04.480)
Certainly awareness of, you know, awareness of what's going on so that the words have meaning
Daniel Kahneman (22:10.640)
or can get, are in touch with some perception or some action.
Daniel Kahneman (22:16.400)
Yeah, so that's a big thing for Jan and as what he refers to as grounding to the physical space.
Lex Fridman (22:23.920)
So that's what we're talking about the same thing.
Lex Fridman (22:26.160)
Yeah, so how do you ground?
Daniel Kahneman (22:29.360)
I mean, the grounding, without grounding, then you get a machine that doesn't know what
Lex Fridman (22:35.200)
it's talking about because it is talking about the world ultimately.
Daniel Kahneman (22:40.240)
The question, the open question is what it means to ground. I mean, we're very
Daniel Kahneman (22:44.880)
human centric in our thinking, but what does it mean for a machine to understand what it means
Daniel Kahneman (22:50.240)
to be in this world? Does it need to have a body? Does it need to have a finiteness like we humans
Lex Fridman (22:57.280)
have all of these elements? It's a very, it's an open question.
Daniel Kahneman (23:02.240)
You know, I'm not sure about having a body, but having a perceptual system,
Daniel Kahneman (23:05.920)
having a body would be very helpful too. I mean, if you think about human, mimicking human,
Daniel Kahneman (23:12.480)
you know, but having a perception that seems to be essential so that you can build,
Daniel Kahneman (23:20.080)
you can accumulate knowledge about the world. So if you can imagine a human completely paralyzed,
Lex Fridman (23:28.240)
and there's a lot that the human brain could learn, you know, with a paralyzed body.
Lex Fridman (23:33.520)
So if we got a machine that could do that, that would be a big deal.
Daniel Kahneman (23:38.640)
TK And then the flip side of that, something you see in children and something in machine
Daniel Kahneman (23:44.960)
learning world is called active learning. Maybe it is also in, is being able to play with the world.
Lex Fridman (23:52.640)
How important for developing System 1 or System 2 do you think it is to play with the world?
Lex Fridman (23:59.760)
To be able to interact with the world?
Daniel Kahneman (24:00.960)
MG A lot of what you learn is you learn to anticipate the outcomes of your actions. I mean,
Daniel Kahneman (24:08.960)
you can see that how babies learn it, you know, with their hands, how they learn, you know,
Daniel Kahneman (24:15.600)
to connect, you know, the movements of their hands with something that clearly is something
Daniel Kahneman (24:20.640)
that happens in the brain and the ability of the brain to learn new patterns. So, you know,
Daniel Kahneman (24:28.320)
it's the kind of thing that you get with artificial limbs, that you connect it and then people learn
Daniel Kahneman (24:34.880)
to operate the artificial limb, you know, really impressively quickly, at least from what I hear.
Lex Fridman (24:44.000)
So we have a system that is ready to learn the world through action.
Lex Fridman (24:49.040)
TK At the risk of going into way too mysterious of land,
Lex Fridman (24:52.640)
what do you think it takes to build a system like that? Obviously, we're very far from understanding
Lex Fridman (25:00.000)
how the brain works, but how difficult is it to build this mind of ours?
Daniel Kahneman (25:08.000)
MG You know, I mean, I think that Jan LeCun's answer that we don't know how many mountains
Daniel Kahneman (25:13.200)
there are, I think that's a very good answer. I think that, you know, if you look at what Ray
Daniel Kahneman (25:20.080)
Kurzweil is saying, that strikes me as off the wall. But I think people are much more realistic
Daniel Kahneman (25:28.800)
than that, where actually Demis Hassabis is and Jan is, and so the people are actually doing the
Daniel Kahneman (25:35.520)
work fairly realistic, I think. TK To maybe phrase it another way,
Lex Fridman (25:41.440)
from a perspective not of building it, but from understanding it,
Lex Fridman (25:44.960)
how complicated are human beings in the following sense? You know, I work with autonomous vehicles
Lex Fridman (25:52.240)
and pedestrians, so we tried to model pedestrians. How difficult is it to model a human being,
Daniel Kahneman (26:00.480)
their perception of the world, the two systems they operate under, sufficiently to be able to
Lex Fridman (26:06.080)
predict whether the pedestrian is going to cross the road or not?
Daniel Kahneman (26:09.280)
MG I'm, you know, I'm fairly optimistic about that, actually, because what we're talking about
Daniel Kahneman (26:18.000)
is a huge amount of information that every vehicle has, and that feeds into one system,
Daniel Kahneman (26:26.800)
into one gigantic system. And so anything that any vehicle learns becomes part of what the whole
Daniel Kahneman (26:33.440)
system knows. And with a system multiplier like that, there is a lot that you can do.
Lex Fridman (26:41.040)
So human beings are very complicated, and the system is going to make mistakes, but human
Daniel Kahneman (26:48.560)
makes mistakes. I think that they'll be able to, I think they are able to anticipate pedestrians,
Daniel Kahneman (26:56.400)
otherwise a lot would happen. They're able to, you know, they're able to get into a roundabout
Lex Fridman (27:04.640)
and into traffic, so they must know both to expect or to anticipate how people will react
Daniel Kahneman (27:14.000)
when they're sneaking in. And there's a lot of learning that's involved in that.
Daniel Kahneman (27:18.800)
RL Currently, the pedestrians are treated as things that cannot be hit, and they're not
Daniel Kahneman (27:28.080)
treated as agents with whom you interact in a game theoretic way. So, I mean, it's not,
Daniel Kahneman (27:37.040)
it's a totally open problem, and every time somebody tries to solve it, it seems to be harder
Daniel Kahneman (27:41.520)
than we think. And nobody's really tried to seriously solve the problem of that dance,
Daniel Kahneman (27:46.640)
because I'm not sure if you've thought about the problem of pedestrians, but you're really
Daniel Kahneman (27:52.080)
putting your life in the hands of the driver.
Daniel Kahneman (27:54.960)
RL You know, there is a dance, there's part of the dance that would be quite complicated,
Lex Fridman (28:00.320)
but for example, when I cross the street and there is a vehicle approaching, I look the driver
Daniel Kahneman (28:05.920)
in the eye, and I think many people do that. And, you know, that's a signal that I'm sending,
Lex Fridman (28:13.360)
and I would be sending that machine to an autonomous vehicle, and it had better understand
Lex Fridman (28:18.480)
it, because it means I'm crossing.
Daniel Kahneman (28:20.720)
RL So, and there's another thing you do, that actually, so I'll tell you what you do,
Daniel Kahneman (28:26.240)
because we watched, I've watched hundreds of hours of video on this, is when you step
Daniel Kahneman (28:31.440)
in the street, you do that before you step in the street, and when you step in the street,
Lex Fridman (28:35.440)
you actually look away.
Daniel Kahneman (28:36.400)
RL Look away.
Daniel Kahneman (28:36.960)
RL Yeah. Now, what is that? What that's saying is, I mean, you're trusting that the car who
Daniel Kahneman (28:45.360)
hasn't slowed down yet will slow down.
Daniel Kahneman (28:48.000)
RL Yeah. And you're telling him, I'm committed. I mean, this is like in a game of chicken,
Lex Fridman (28:53.680)
so I'm committed, and if I'm committed, I'm looking away. So, there is, you just have
Lex Fridman (28:59.840)
to stop.
Daniel Kahneman (29:00.320)
RL So, the question is whether a machine that observes that needs to understand mortality.
Daniel Kahneman (29:06.880)
RL Here, I'm not sure that it's got to understand so much as it's got to anticipate. So, and
Daniel Kahneman (29:17.120)
here, but you know, you're surprising me, because here I would think that maybe you
Daniel Kahneman (29:24.400)
can anticipate without understanding, because I think this is clearly what's happening in
Daniel Kahneman (29:30.560)
playing go or in playing chess. There's a lot of anticipation, and there is zero understanding.
Lex Fridman (29:35.600)
RL Exactly.
Daniel Kahneman (29:36.240)
RL So, I thought that you didn't need a model of the human and a model of the human mind
Lex Fridman (29:46.400)
to avoid hitting pedestrians, but you are suggesting that actually…
Daniel Kahneman (29:50.880)
RL There you go, yeah.
Lex Fridman (29:51.840)
RL You do. Then it's a lot harder, I thought.
Daniel Kahneman (29:56.720)
RL And I have a follow up question to see where your intuition lies. It seems that almost
Daniel Kahneman (30:02.560)
every robot human collaboration system is a lot harder than people realize. So, do you
Daniel Kahneman (30:10.800)
think it's possible for robots and humans to collaborate successfully? We talked a little
Daniel Kahneman (30:17.200)
bit about semi autonomous vehicles, like in the Tesla autopilot, but just in tasks in
Daniel Kahneman (30:23.360)
general. If you think we talked about current neural networks being kind of system one,
Lex Fridman (30:30.160)
do you think those same systems can borrow humans for system two type tasks and collaborate
Lex Fridman (30:40.240)
successfully?
Daniel Kahneman (30:40.880)
RL Well, I think that in any system where humans and the machine interact, the human
Daniel Kahneman (30:49.520)
will be superfluous within a fairly short time. That is, if the machine is advanced
Daniel Kahneman (30:55.760)
enough so that it can really help the human, then it may not need the human for a long
Daniel Kahneman (31:01.600)
time. Now, it would be very interesting if there are problems that for some reason the
Daniel Kahneman (31:08.320)
machine cannot solve, but that people could solve. Then you would have to build into the
Daniel Kahneman (31:14.240)
machine an ability to recognize that it is in that kind of problematic situation and
Daniel Kahneman (31:22.080)
to call the human. That cannot be easy without understanding. That is, it must be very difficult
Daniel Kahneman (31:30.880)
to program a recognition that you are in a problematic situation without understanding
Lex Fridman (31:38.400)
the problem.
Daniel Kahneman (31:39.440)
SL. That's very true. In order to understand the full scope of situations that are problematic,
Lex Fridman (31:47.360)
you almost need to be smart enough to solve all those problems.
Daniel Kahneman (31:51.680)
RL It's not clear to me how much the machine will need the human. I think the example of
Daniel Kahneman (32:01.120)
chess is very instructive. I mean, there was a time at which Kasparov was saying that human
Daniel Kahneman (32:06.160)
machine combinations will beat everybody. Even stockfish doesn't need people and Alpha
Lex Fridman (32:13.440)
Zero certainly doesn't need people.
Daniel Kahneman (32:15.280)
SL. The question is, just like you said, how many problems are like chess and how many
Daniel Kahneman (32:20.880)
problems are not like chess? Every problem probably in the end is like chess. The question
Lex Fridman (32:27.760)
is, how long is that transition period?
Daniel Kahneman (32:29.760)
RL That's a question I would ask you. Autonomous vehicle, just driving, is probably a lot more
Daniel Kahneman (32:38.880)
complicated than Go to solve that problem. Because it's open. That's not surprising to
Daniel Kahneman (32:47.840)
me because there is a hierarchical aspect to this, which is recognizing a situation
Lex Fridman (32:58.960)
and then within the situation bringing up the relevant knowledge. For that hierarchical
Daniel Kahneman (33:09.280)
type of system to work, you need a more complicated system than we currently have.
Daniel Kahneman (33:15.760)
SL. A lot of people think, because as human beings, this is probably the cognitive biases,
Daniel Kahneman (33:22.720)
they think of driving as pretty simple because they think of their own experience. This is
Daniel Kahneman (33:28.720)
actually a big problem for AI researchers or people thinking about AI because they evaluate
Lex Fridman (33:36.400)
how hard a particular problem is based on very limited knowledge, based on how hard
Daniel Kahneman (33:43.280)
it is for them to do the task. And then they take for granted, maybe you can speak to that
Daniel Kahneman (33:49.120)
because most people tell me driving is trivial and humans in fact are terrible at driving
Daniel Kahneman (33:56.720)
is what people tell me. And I see humans and humans are actually incredible at driving
Lex Fridman (34:02.040)
and driving is really terribly difficult. Is that just another element of the effects
Lex Fridman (34:08.520)
that you've described in your work on the psychology side?
Daniel Kahneman (34:13.680)
No, I mean, I haven't really, I would say that my research has contributed nothing to
Daniel Kahneman (34:22.000)
understanding the ecology and to understanding the structure of situations and the complexity
Daniel Kahneman (34:27.800)
of problems. So all we know is very clear that that goal, it's endlessly complicated,
Lex Fridman (34:38.720)
but it's very constrained. And in the real world, there are far fewer constraints and
Lex Fridman (34:46.840)
many more potential surprises.
Daniel Kahneman (34:49.320)
SL. So that's obvious because it's not always obvious to people, right? So when you think
Lex Fridman (34:54.720)
about…
Daniel Kahneman (34:55.720)
Well, I mean, you know, people thought that reasoning was hard and perceiving was easy,
Lex Fridman (35:02.880)
but you know, they quickly learned that actually modeling vision was tremendously complicated
Lex Fridman (35:09.920)
and modeling, even proving theorems was relatively straightforward.
Daniel Kahneman (35:15.960)
To push back on that a little bit on the quickly part, it took several decades to learn that
Lex Fridman (35:22.800)
and most people still haven't learned that. I mean, our intuition, of course, AI researchers
Daniel Kahneman (35:28.400)
have, but you drift a little bit outside the specific AI field, the intuition is still
Daniel Kahneman (35:34.760)
perceptible to solve that.
Daniel Kahneman (35:36.320)
No, I mean, that's true. Intuitions, the intuitions of the public haven't changed
Daniel Kahneman (35:41.280)
radically. And they are, as you said, they're evaluating the complexity of problems by how
Daniel Kahneman (35:48.760)
difficult it is for them to solve the problems. And that's got very little to do with the
Daniel Kahneman (35:55.720)
complexities of solving them in AI.
Daniel Kahneman (35:58.360)
SL. How do you think from the perspective of an AI researcher, do we deal with the intuitions
Daniel Kahneman (36:06.120)
of the public? So in trying to think, arguably, the combination of hype investment and the
Daniel Kahneman (36:15.080)
public intuition is what led to the AI winters. I'm sure that same could be applied to tech
Daniel Kahneman (36:21.160)
or that the intuition of the public leads to media hype, leads to companies investing
Daniel Kahneman (36:29.700)
in the tech, and then the tech doesn't make the company's money. And then there's a crash.
Lex Fridman (36:36.700)
Is there a way to educate people to fight the, let's call it system one thinking?
Daniel Kahneman (36:43.280)
In general, no. I think that's the simple answer. And it's going to take a long time
Daniel Kahneman (36:54.600)
before the understanding of what those systems can do becomes public knowledge. And then
Daniel Kahneman (37:09.240)
the expectations, there are several aspects that are going to be very complicated. The
Daniel Kahneman (37:20.920)
fact that you have a device that cannot explain itself is a major, major difficulty. And we're
Daniel Kahneman (37:29.720)
already seeing that. I mean, this is really something that is happening. So it's happening
Daniel Kahneman (37:35.520)
in the judicial system. So you have system that are clearly better at predicting parole
Daniel Kahneman (37:43.600)
violations than judges, but they can't explain their reasoning. And so people don't want
Daniel Kahneman (37:54.220)
to trust them.
Daniel Kahneman (37:56.040)
We seem to in system one, even use cues to make judgements about our environment. So
Lex Fridman (38:05.400)
this explainability point, do you think humans can explain stuff?
Daniel Kahneman (38:11.040)
No, but I mean, there is a very interesting aspect of that. Humans think they can explain
Daniel Kahneman (38:20.400)
themselves. So when you say something and I ask you, why do you believe that? Then reasons
Daniel Kahneman (38:28.160)
will occur to you. But actually, my own belief is that in most cases, the reasons have very
Daniel Kahneman (38:35.880)
little to do with why you believe what you believe. So that the reasons are a story that
Daniel Kahneman (38:41.880)
comes to your mind when you need to explain yourself. But people traffic in those explanations
Daniel Kahneman (38:50.200)
I mean, the human interaction depends on those shared fictions and, and the stories that
Lex Fridman (38:56.680)
people tell themselves.
Daniel Kahneman (38:58.580)
You just made me actually realize and we'll talk about stories in a second. That not to
Daniel Kahneman (39:05.960)
be cynical about it, but perhaps there's a whole movement of people trying to do explainable
Daniel Kahneman (39:11.520)
AI. And really, we don't necessarily need to explain AI doesn't need to explain itself.
Lex Fridman (39:19.360)
It just needs to tell a convincing story.
Daniel Kahneman (39:21.880)
Yeah, absolutely.
Daniel Kahneman (39:23.560)
It doesn't necessarily, the story doesn't necessarily need to reflect the truth as it
Daniel Kahneman (39:29.160)
might, it just needs to be convincing. There's something to that.
Daniel Kahneman (39:32.800)
You can say exactly the same thing in a way that sounds cynical or doesn't sound cynical.
Daniel Kahneman (39:38.840)
Right.
Lex Fridman (39:39.840)
But the objective of having an explanation is to tell a story that will be acceptable
Daniel Kahneman (39:48.000)
to people. And, and, and for it to be acceptable and to be robustly acceptable, it has to have
Lex Fridman (39:56.360)
some elements of truth. But, but the objective is for people to accept it.
Daniel Kahneman (40:04.480)
It's quite brilliant, actually. But so on the, on the stories that we tell, sorry to
Daniel Kahneman (40:11.720)
ask me, ask you the question that most people know the answer to, but you talk about two
Daniel Kahneman (40:18.000)
selves in terms of how life is lived, the experienced self and remembering self. Can
Lex Fridman (40:24.780)
you describe the distinction between the two?
Daniel Kahneman (40:26.920)
Well, sure. I mean, the, there is an aspect of, of life that occasionally, you know, most
Daniel Kahneman (40:33.680)
of the time we just live and we have experiences and they're better and they're worse and it
Daniel Kahneman (40:38.520)
goes on over time. And mostly we forget everything that happens or we forget most of what happens.
Daniel Kahneman (40:45.760)
Then occasionally you, when something ends or at different points, you evaluate the past
Lex Fridman (40:56.280)
and you form a memory and the memory is schematic. It's not that you can roll a film of an interaction.
Daniel Kahneman (41:03.560)
You construct, in effect, the elements of a story about an, about an episode. So there
Daniel Kahneman (41:12.960)
is the experience and there is the story that is created about the experience. And that's
Lex Fridman (41:18.360)
what I call the remembering. So I had the image of two selves. So there is a self that
Daniel Kahneman (41:24.320)
lives and there is a self that evaluates life. Now the paradox and the deep paradox in that
Daniel Kahneman (41:32.200)
is that we have one system or one self that does the living, but the other system, the
Daniel Kahneman (41:41.960)
remembering self is all we get to keep. And basically decision making and, and everything
Daniel Kahneman (41:49.180)
that we do is governed by our memories, not by what actually happened. It's, it's governed
Daniel Kahneman (41:55.000)
by, by the story that we told ourselves or by the story that we're keeping. So that's,
Lex Fridman (42:02.280)
that's the distinction.
Daniel Kahneman (42:03.280)
I mean, there's a lot of brilliant ideas about the pursuit of happiness that come out of
Lex Fridman (42:08.000)
that. What are the properties of happiness which emerge from a remembering self?
Daniel Kahneman (42:14.160)
There are, there are properties of how we construct stories that are really important.
Lex Fridman (42:19.160)
So that I studied a few, but, but a couple are really very striking. And one is that
Daniel Kahneman (42:29.720)
in stories, time doesn't matter. There's a sequence of events or there are highlights
Daniel Kahneman (42:37.080)
or not. And, and how long it took, you know, they lived happily ever after or three years
Daniel Kahneman (42:45.240)
later or something. It, time really doesn't matter. And in stories, events matter, but
Daniel Kahneman (42:53.480)
time doesn't. That, that leads to a very interesting set of problems because time is all we got
Daniel Kahneman (43:03.740)
to live. I mean, you know, time is the currency of life. And yet time is not represented basically
Daniel Kahneman (43:11.040)
in evaluated memories. So that, that creates a lot of paradoxes that I've thought about.
Daniel Kahneman (43:18.520)
Yeah. They're fascinating. But if you were to give advice on how one lives a happy life
Lex Fridman (43:27.520)
based on such properties, what's the optimal?
Daniel Kahneman (43:33.120)
You know, I gave up, I abandoned happiness research because I couldn't solve that problem.
Daniel Kahneman (43:38.880)
I couldn't, I couldn't see. And in the first place, it's very clear that if you do talk
Daniel Kahneman (43:46.160)
in terms of those two selves, then that what makes the remembering self happy and what
Daniel Kahneman (43:51.520)
makes the experiencing self happy are different things. And I, I asked the question of, suppose
Daniel Kahneman (43:59.320)
you're planning a vacation and you're just told that at the end of the vacation, you'll
Daniel Kahneman (44:04.160)
get an amnesic drug, so you remember nothing. And they'll also destroy all your photos.
Lex Fridman (44:10.160)
So there'll be nothing. Would you still go to the same vacation? And, and it's, it turns
Daniel Kahneman (44:20.640)
out we go to vacations in large part to construct memories, not to have experiences, but to
Daniel Kahneman (44:26.600)
construct memories. And it turns out that the vacation that you would want for yourself,
Daniel Kahneman (44:32.520)
if you knew, you will not remember is probably not the same vacation that you will want for
Daniel Kahneman (44:38.080)
yourself if you will remember. So I have no solution to these problems, but clearly those
Lex Fridman (44:46.240)
are big issues.
Lex Fridman (44:47.240)
And you've talked about, you've talked about sort of how many minutes or hours you spend
Daniel Kahneman (44:53.060)
about the vacation. It's an interesting way to think about it because that's how you really
Daniel Kahneman (44:58.120)
experience the vacation outside the being in it. But there's also a modern, I don't
Daniel Kahneman (45:03.640)
know if you think about this or interact with it. There's a modern way to, um, magnify the
Daniel Kahneman (45:11.440)
remembering self, which is by posting on Instagram, on Twitter, on social networks. A lot of people
Daniel Kahneman (45:17.820)
live life for the picture that you take, that you post somewhere. And now thousands of people
Daniel Kahneman (45:24.680)
share and potentially potentially millions. And then you can relive it even much more
Lex Fridman (45:29.040)
than just those minutes. Do you think about that magnification much?
Daniel Kahneman (45:34.280)
You know, I'm too old for social networks. I, you know, I've never seen Instagram, so
Lex Fridman (45:41.960)
I cannot really speak intelligently about those things. I'm just too old.
Lex Fridman (45:46.640)
But it's interesting to watch the exact effects you've described.
Daniel Kahneman (45:49.840)
Make a very big difference. I mean, and it will make, it will also make a difference.
Lex Fridman (45:55.560)
And that I don't know whether, uh, it's clear that in some ways the devices that serve us
Daniel Kahneman (46:06.040)
are supplant functions. So you don't have to remember phone numbers. You don't have,
Daniel Kahneman (46:12.960)
you really don't have to know facts. I mean, the number of conversations I'm involved with,
Daniel Kahneman (46:19.080)
somebody says, well, let's look it up. Uh, so it's, it's in a way it's made conversations.
Daniel Kahneman (46:27.640)
Well it's, it means that it's much less important to know things. You know, it used to be very
Daniel Kahneman (46:33.360)
important to know things. This is changing. So the requirements of that, that we have
Daniel Kahneman (46:43.200)
for ourselves and for other people are changing because of all those supports and because,
Lex Fridman (46:50.560)
and I have no idea what Instagram does, but it's, uh, well, I'll tell you, I wish I could
Daniel Kahneman (46:57.600)
just have the, my remembering self could enjoy this conversation, but I'll get to enjoy it
Daniel Kahneman (47:03.600)
even more by having watched, by watching it and then talking to others. It'll be about
Lex Fridman (47:08.520)
a hundred thousand people as scary as this to say, well, listen or watch this, right?
Daniel Kahneman (47:14.880)
It changes things. It changes the experience of the world that you seek out experiences
Daniel Kahneman (47:20.320)
which could be shared in that way. It's in, and I haven't seen, it's, it's the same effects
Daniel Kahneman (47:25.920)
that you described. And I don't think the psychology of that magnification has been
Daniel Kahneman (47:30.760)
described yet because it's a new world.
Lex Fridman (47:33.240)
But the sharing, there was a, there was a time when people read books and, uh, and,
Lex Fridman (47:43.240)
and you could assume that your friends had read the same books that you read. So there
Daniel Kahneman (47:51.140)
was kind of invisible sharing. There was a lot of sharing going on and there was a lot
Daniel Kahneman (47:57.760)
of assumed common knowledge and, you know, that was built in. I mean, it was obvious
Daniel Kahneman (48:03.780)
that you had read the New York Times. It was obvious that you had read the reviews. I mean,
Lex Fridman (48:09.520)
so a lot was taken for granted that was shared. And, you know, when there were, when there
Daniel Kahneman (48:17.040)
were three television channels, it was obvious that you'd seen one of them probably the same.
Lex Fridman (48:26.000)
So sharing, sharing always was always there. It was just different.
Daniel Kahneman (48:32.400)
At the risk of, uh, inviting mockery from you, let me say that I'm also a fan of Sartre
Lex Fridman (48:40.920)
and Camus and existentialist philosophers. And, um, I'm joking of course about mockery,
Lex Fridman (48:47.560)
but from the perspective of the two selves, what do you think of the existentialist philosophy
Daniel Kahneman (48:54.180)
of life? So trying to really emphasize the experiencing self as the proper way to, or
Lex Fridman (49:03.680)
the best way to live life.
Daniel Kahneman (49:05.960)
I don't know enough philosophy to answer that, but it's not, uh, you know, the emphasis on,
Lex Fridman (49:13.600)
on experience is also the emphasis in Buddhism.
Daniel Kahneman (49:16.760)
Yeah, right. That's right.
Daniel Kahneman (49:18.040)
So, uh, that's, you just have got to, to experience things and, and, and not to evaluate and not
Daniel Kahneman (49:27.280)
to pass judgment and not to score, not to keep score. So, uh,
Daniel Kahneman (49:33.560)
If, when you look at the grand picture of experience, you think there's something to
Daniel Kahneman (49:37.760)
that, that one, one of the ways to achieve contentment and maybe even happiness is letting
Lex Fridman (49:44.480)
go of any of the things, any of the procedures of the remembering self.
Daniel Kahneman (49:51.800)
Well, yeah, I mean, I think, you know, if one could imagine a life in which people don't
Daniel Kahneman (49:58.080)
score themselves, uh, it, it feels as if that would be a better life as if the self scoring
Lex Fridman (50:05.960)
and you know, how am I doing a kind of question, uh, is not, is not a very happy thing to have.
Lex Fridman (50:18.040)
But I got out of that field because I couldn't solve that problem and, and that was because
Daniel Kahneman (50:25.360)
my intuition was that the experiencing self, that's reality.
Lex Fridman (50:31.500)
But then it turns out that what people want for themselves is not experiences. They want
Daniel Kahneman (50:36.560)
memories and they want a good story about their life. And so you cannot have a theory
Daniel Kahneman (50:41.600)
of happiness that doesn't correspond to what people want for themselves. And when I, when
Daniel Kahneman (50:47.880)
I realized that this, this was where things were going, I really sort of left the field
Lex Fridman (50:53.760)
of research.
Lex Fridman (50:54.760)
Do you think there's something instructive about this emphasis of reliving memories in
Daniel Kahneman (51:01.100)
building AI systems. So currently artificial intelligence systems are more like experiencing
Daniel Kahneman (51:09.200)
self in that they react to the environment. There's some pattern formation like a learning
Lex Fridman (51:16.280)
so on, but you really don't construct memories, uh, except in reinforcement learning every
Daniel Kahneman (51:23.120)
once in a while that you replay over and over.
Lex Fridman (51:25.720)
Yeah, but you know, that would in principle would not be.
Lex Fridman (51:30.280)
Do you think that's useful? Do you think it's a feature or a bug of human beings that we,
Lex Fridman (51:36.000)
that we look back?
Daniel Kahneman (51:37.000)
Oh, I think that's definitely a feature. That's not a bug. I mean, you, you have to look back
Daniel Kahneman (51:43.360)
in order to look forward. So, uh, without, without looking back, you couldn't, you couldn't
Daniel Kahneman (51:50.440)
really intelligently look forward.
Daniel Kahneman (51:53.080)
You're looking for the echoes of the same kind of experience in order to predict what
Daniel Kahneman (51:57.080)
the future holds.
Lex Fridman (51:58.080)
Yeah.
Lex Fridman (51:59.080)
So though Victor Frankel in his book, man's search for meaning, I'm not sure if you've
Daniel Kahneman (52:05.320)
read, describes his experience at the consecration concentration camps during world war two as
Daniel Kahneman (52:10.720)
a way to describe that finding identifying a purpose in life, a positive purpose in life
Daniel Kahneman (52:18.480)
can save one from suffering. First of all, do you connect with the philosophy that he
Lex Fridman (52:23.840)
describes there?
Daniel Kahneman (52:28.420)
Not really. I mean, the, so I can, I can really see that somebody who has that feeling of
Daniel Kahneman (52:37.040)
purpose and meaning and so on, that, that could sustain you. Uh, I in general don't
Daniel Kahneman (52:44.640)
have that feeling and I'm pretty sure that if I were in a concentration camp, I'd give
Daniel Kahneman (52:50.800)
up and die, you know? So he talks, he is, he is a survivor.
Lex Fridman (52:56.240)
Yeah.
Daniel Kahneman (52:57.240)
And, you know, he survived with that. And I'm, and I'm not sure how essential to survival
Daniel Kahneman (53:04.000)
this sense is, but I do know when I think about myself that I would have given up. Oh,
Daniel Kahneman (53:12.220)
this isn't going anywhere. And there is, there is a sort of character that, that, that manages
Daniel Kahneman (53:20.140)
to survive in conditions like that. And then because they survive, they tell stories and
Daniel Kahneman (53:26.120)
it sounds as if they survive because of what they were doing. We have no idea. They survived
Daniel Kahneman (53:31.840)
because the kind of people that they are and the other kind of people who survives and
Daniel Kahneman (53:36.240)
would tell themselves stories of a particular kind. So I'm not, uh,
Lex Fridman (53:41.800)
So you don't think seeking purpose is a significant driver in our being?
Daniel Kahneman (53:46.840)
Oh, I mean, it's, it's a very interesting question because when you ask people whether
Daniel Kahneman (53:52.400)
it's very important to have meaning in their life, they say, oh yes, that's the most important
Daniel Kahneman (53:56.240)
thing. But when you ask people, what kind of a day did you have? And, and you know,
Lex Fridman (54:03.880)
what were the experiences that you remember? You don't get much meaning. You get social
Daniel Kahneman (54:10.320)
experiences. Then, uh, and, and some people say that, for example, in, in, in child, you
Daniel Kahneman (54:21.480)
know, in taking care of children, the fact that they are your children and you're taking
Daniel Kahneman (54:25.720)
care of them, uh, makes a very big difference. I think that's entirely true. Uh, but it's
Daniel Kahneman (54:34.040)
more because of a story that we're telling ourselves, which is a very different story
Daniel Kahneman (54:40.560)
when we're taking care of our children or when we're taking care of other things.
Daniel Kahneman (54:45.140)
Jumping around a little bit in doing a lot of experiments, let me ask a question. Most
Daniel Kahneman (54:50.880)
of the work I do, for example, is in the, in the real world, but most of the clean good
Daniel Kahneman (54:56.840)
science that you can do is in the lab. So that distinction, do you think we can understand
Daniel Kahneman (55:04.480)
the fundamentals of human behavior through controlled experiments in the lab? If we talk
Daniel Kahneman (55:12.680)
about pupil diameter, for example, it's much easier to do when you can control lighting
Daniel Kahneman (55:18.920)
conditions, right? So when we look at driving, lighting variation destroys almost completely
Daniel Kahneman (55:27.680)
your ability to use pupil diameter. But in the lab for, as I mentioned, semi autonomous
Daniel Kahneman (55:34.740)
or autonomous vehicles in driving simulators, we can't, we don't capture true, honest, uh,
Daniel Kahneman (55:43.080)
human behavior in that particular domain. So what's your intuition? How much of human
Daniel Kahneman (55:49.000)
behavior can we study in this controlled environment of the lab? A lot, but you'd have to verify
Daniel Kahneman (55:56.160)
it, you know, that your, your conclusions are basically limited to the situation, to
Daniel Kahneman (56:03.240)
the experimental situation. Then you have to jump the big inductive leap to the real
Daniel Kahneman (56:09.000)
world. Uh, so, and, and that's the flare. That's where the difference, I think, between
Daniel Kahneman (56:17.920)
the good psychologists and others that are mediocre is in the sense of that your experiment
Daniel Kahneman (56:25.840)
captures something that's important and something that's real and others are just running experiments.
Lex Fridman (56:33.520)
So what is that? Like the birth of an idea to his development in your mind to something
Daniel Kahneman (56:39.000)
that leads to an experiment. Is that similar to maybe like what Einstein or a good physicist
Daniel Kahneman (56:44.840)
do is your intuition. You basically use your intuition to build up.
Daniel Kahneman (56:48.840)
Yeah, but I mean, you know, it's, it's very skilled intuition. I mean, I just had that
Daniel Kahneman (56:54.280)
experience actually. I had an idea that turns out to be very good idea a couple of days
Daniel Kahneman (57:00.840)
ago and, and you, and you have a sense of that building up. So I'm working with a collaborator
Lex Fridman (57:08.400)
and he essentially was saying, you know, what, what are you doing? What's, what's going on?
Lex Fridman (57:14.280)
And I was, I really, I couldn't exactly explain it, but I knew this is going somewhere, but
Daniel Kahneman (57:21.000)
you know, I've been around that game for a very long time. And so I can, you, you develop
Daniel Kahneman (57:26.920)
that anticipation that yes, this, this is worth following up. That's part of the skill.
Daniel Kahneman (57:34.640)
Is that something you can reduce to words in describing a process in the form of advice
Lex Fridman (57:41.560)
to others?
Daniel Kahneman (57:42.560)
No.
Lex Fridman (57:43.560)
Follow your heart, essentially.
Daniel Kahneman (57:45.560)
I mean, you know, it's, it's like trying to explain what it's like to drive. It's not,
Lex Fridman (57:51.680)
you've got to break it apart and it's not.
Lex Fridman (57:54.140)
And then you lose.
Lex Fridman (57:55.140)
And then you lose the experience.
Daniel Kahneman (57:58.080)
You mentioned collaboration. You've written about your collaboration with Amos Tversky
Daniel Kahneman (58:05.140)
that this is you writing, the 12 or 13 years in which most of our work was joint were years
Daniel Kahneman (58:10.780)
of interpersonal and intellectual bliss. Everything was interesting. Almost everything
Daniel Kahneman (58:16.720)
was funny. And there was a current joy of seeing an idea take shape. So many times in
Daniel Kahneman (58:22.080)
those years, we shared the magical experience of one of us saying something, which the other
Daniel Kahneman (58:27.320)
one would understand more deeply than the speaker had done. Contrary to the old laws
Daniel Kahneman (58:32.520)
of information theory, it was common for us to find that more information was received
Daniel Kahneman (58:38.000)
than had been sent. I have almost never had the experience with anyone else. If you have
Daniel Kahneman (58:43.860)
not had it, you don't know how marvelous collaboration can be.
Lex Fridman (58:49.120)
So let me ask a perhaps a silly question. How does one find and create such a collaboration?
Lex Fridman (58:58.840)
That may be asking like, how does one find love?
Daniel Kahneman (59:01.120)
Yeah, you have to be lucky. And I think you have to have the character for that because
Daniel Kahneman (59:10.600)
I've had many collaborations. I mean, none were as exciting as with Amos, but I've had
Lex Fridman (59:17.600)
and I'm having just very. So it's a skill. I think I'm good at it. Not everybody is good
Daniel Kahneman (59:27.040)
at it. And then it's the luck of finding people who are also good at it.
Daniel Kahneman (59:32.100)
Is there advice in a form for a young scientist who also seeks to violate this law of information
Lex Fridman (59:39.420)
theory?
Daniel Kahneman (59:48.520)
I really think it's so much luck is involved. And in those really serious collaborations,
Daniel Kahneman (59:59.560)
at least in my experience, are a very personal experience. And I have to like the person
🔗 相关节目