HomeVideos

Yuval Noah Harari on Donald Trump’s Core Delusion | The Ezra Klein Show

Now Playing

Yuval Noah Harari on Donald Trump’s Core Delusion | The Ezra Klein Show

Transcript

3002 segments

0:00

I think if you look across his

0:01

mega-bestselling books like Sapiens and

0:04

Homo Deus,

0:05

Yuval Noah Harari really has one major

0:08

topic. [music]

0:09

That topic is cooperation.

0:12

Cooperation and the ability to cooperate

0:14

across scale, across time, as being the

0:17

fundamental engine of human progress.

0:19

Cooperation as the way we go

0:21

from being this [music] creature that

0:23

absolutely cannot beat a bear or a lion

0:25

in a fight

0:27

to being able to create and command the

0:29

societies we have now.

0:31

I think right now there's something

0:32

interestingly challenging about Harari's

0:34

work [music] because we live in this

0:35

moment of Trumpism, of right-wing

0:37

populism, and one of the messages of

0:39

those movements is that this emphasis on

0:41

cooperation, on positive-sum

0:44

relationships,

0:45

>> [music]

0:46

>> is a lie.

0:47

That humanity, that society, is driven

0:51

not so much by these like soft questions

0:53

of cooperation [music] as by power,

0:56

hierarchy,

0:58

dominance, about winning the transaction

1:00

with the other, about coming out ahead

1:03

in the conflict,

1:04

in the trade.

1:06

That all these like niceties of

1:08

liberalism,

1:10

they were a lie.

1:11

And that [music] really humanity runs on

1:15

power.

1:17

And that to forget that is to forget the

1:19

engine [music] of our progress.

1:21

So I've been wanting to talk to Harari

1:22

about this. I think there's a an

1:23

interesting debate to put him in

1:25

conversation with. He's a new book for

1:27

kids called Unstoppable Us, volume

1:28

three. It is also about cooperation and

1:31

how enemies turn into [music] friends,

1:32

but this conversation is bigger than

1:34

that. It's about liberalism,

1:36

it's about [music]

1:37

Israel, Harari is Israeli, it's about AI

1:40

and what it's going to do to us and what

1:42

it's [music] going to do to language as

1:43

the way we work with and fail to work

1:47

with each other. It is, as we say in the

1:49

podcast biz, a [music] wide-ranging

1:51

conversation and all the better for it.

1:54

As always, my email is

1:55

ezrakleinshow@nytimes.com.

2:03

Yuval Noah Harari, welcome to the show.

2:05

Thank you. It's good to be here. I

2:07

wanted to begin with a clip of Stephen

2:10

Miller, uh Donald Trump's deputy chief

2:12

of staff, that I began thinking about as

2:13

I was reading some of your recent work.

2:15

I'm going to play it here. You can talk

2:17

all you want about international

2:19

niceties and everything else, but we

2:21

live in a world, in the real world,

2:24

Jack, that is governed by

2:27

strength, that is governed by force,

2:30

that is governed by power. These are the

2:33

iron laws of the world that exist

2:35

since the beginning of time.

2:36

>> do you think when you hear that?

2:38

That the whole of the of uh the history

2:41

of philosophy and spirituality is uh in

2:45

argument with exactly that point of

2:47

view.

2:49

That the only reality is power. The only

2:52

reality is force.

2:54

And from the viewpoint of a historian,

2:56

it's clear that this is not the case.

2:59

If the only human reality was brute

3:02

force, we would still be living in tiny

3:05

hunter-gatherer bands in the African

3:08

savanna.

3:09

Because the whole of human history is

3:11

about how do you get more people to

3:14

cooperate and to trust each other. And

3:17

you cannot do that only with brute

3:20

force.

3:21

I I want to spend some time here

3:23

on this tension between visions of

3:25

cooperation as a driving force in human

3:27

history and visions of power

3:29

as a driving force in human history.

3:31

Because I think if I'm trying to steel

3:33

man

3:34

the vision that emerges out of the Trump

3:36

administration and some other

3:38

uh political figures like them right

3:40

now, they would say that the conditions

3:43

for cooperation

3:44

have been a mixture of shared national

3:47

and religious stories Mhm. and

3:51

hierarchy, [snorts]

3:52

power, domination, and and subjugation.

3:56

And that what they are trying to

3:56

recreate are these conditions that have

3:59

held, that have allowed,

4:01

you know, the great countries to become

4:02

great.

4:03

And I think it's appealing to people.

4:05

Mhm. But but the other dimension is your

4:08

work is so much about shared story and

4:10

story as the operating system that

4:13

permits human cooperation at large

4:15

scale, right? And I think something

4:17

that,

4:18

you know, people like Donald Trump or or

4:20

in, you know, uh Yoram Hazony, the uh

4:24

nationalist kind of philosopher, argue

4:26

is that

4:28

we need these intense stories of

4:29

nations, of ethnic solidarity, of

4:32

religious solidarity.

4:34

And liberalism

4:35

and all these nice human rights-fearing

4:40

uh

4:41

ideologies that emerge have begun to

4:43

corrode them. And so, they're corroding

4:45

the very conditions for cooperation. And

4:48

and I'm curious as somebody who's been

4:49

in the in the debates, how you think

4:50

about that. That's a different argument.

4:53

I mean, it's an argument that recognizes

4:55

that not everything is based is based

4:58

just on force and brute power.

5:00

Definitely, nationalism has been one of

5:03

the most successful and also one of the

5:05

most positive stories that humans have

5:08

ever come up with.

5:10

Uh for me, nationalism is not about

5:13

hating other groups.

5:15

Nationalism at its core is about loving

5:19

and caring about a a a a a a a large

5:22

number of strangers that you do not know

5:25

personally, but you're nevertheless

5:27

willing to make a lot of sacrifices for

5:30

them. The nation is not a family. The

5:32

nation is not even a small tribe. In a

5:36

small tribe, you know everybody. It's

5:38

based on personal relationships.

5:41

With nations, one of the most striking

5:43

things about them is that you don't know

5:46

99.99%

5:48

of the other people in your nation, and

5:50

this is true not only of big nations

5:52

like China or India, this is also true

5:53

of Israel. There are like about 10

5:56

million Israelis, I don't know most of

5:58

them. And nevertheless, nationalism

6:00

makes people care about these strangers

6:04

enough so that for instance, you pay

6:07

taxes so that other people in your

6:09

nation will get good health care and

6:11

education. And ultimately, and in some

6:14

circumstances, even risk your life for

6:17

them.

6:18

Sometimes, of course, nationalism veers

6:21

into hatred of others, but this is not

6:24

an essential feature of nationalism.

6:27

Nationalism can exist without hating

6:30

outsiders. It cannot exist

6:33

without love for insiders. And many of

6:36

the people today who present themselves

6:39

as as the champions of nationalism,

6:42

they put the emphasis on hatred. And in

6:46

many cases, they even create hatreds

6:49

within the nation. They divide the

6:51

nation against itself. They think they

6:53

are great patriots if they if they hate

6:57

outsiders.

6:59

And you know, looking at Israel as an

7:00

example,

7:02

nobody, I think, in the history of

7:03

Israel divided the nation against itself

7:06

more than Netanyahu.

7:08

And in this sense, he has been the worst

7:10

enemy

7:11

of Israeli nationalism. Yes, he hates

7:14

outsiders, okay, but this is not the key

7:17

test.

7:18

And then the question is, how would

7:20

different nation conduct their

7:23

relationships?

7:25

It starts with issues of security and

7:27

foreign policy.

7:29

You know, the the Trumpian vision, which

7:32

is all about force and hierarchy, it

7:35

basically says, the way to organize the

7:38

international system is if the weak

7:41

always surrender to the demands of the

7:44

strong, and then we have order, and then

7:47

we have even peace.

7:49

So, if the United States demands

7:51

Greenland, Denmark must recognize

7:53

reality and give Greenland to the United

7:56

States. If Denmark refuses, and as a

7:59

result there is violence, there is a

8:02

war, there is conflict, this is the

8:05

fault of Denmark

8:07

for refusing to recognize the reality

8:09

and giving the strong what they demand.

8:12

This is their logic. This is how they

8:13

see the world. Now, the problem, leaving

8:16

aside the issue of morality, still you

8:18

have a big problem.

8:20

The big problem is, first of all, that

8:23

all nations are then driven to become

8:26

strong because you cannot survive as a

8:29

weak nation in such a world. And then

8:32

all nations are forced to

8:36

invest more and more of their resources

8:39

in their military.

8:41

For most of history, a lot of the budget

8:44

of every kingdom, empire, republic,

8:47

city-state was was

8:51

invested or wasted on soldiers and

8:54

fortresses and and warships and things

8:57

like that, and nobody felt safe.

9:00

One of the miracles of the international

9:02

systems of recent decades, and this is

9:05

not something It's not about, you know,

9:07

writing

9:09

pacifistic poetry. It's about government

9:12

budgets. You look at the budgets, you

9:15

see that in

9:16

on average, in the early 21st century,

9:19

on average, about 6 to 7% of the

9:22

government budget went for defense, for

9:25

the military, compared to 10% on average

9:28

that went to health care. It's the first

9:31

time in history

9:33

that humanity spent more on health care

9:35

than on defense,

9:37

And they felt more secure than in any

9:39

previous time in history because there

9:42

was this taboo

9:43

on invading and conquering other

9:46

countries by force. Now, if we now break

9:49

this taboo,

9:50

it will force everybody to arm

9:52

themselves to the teeth at the expense

9:55

of health care, education, welfare, and

9:57

so forth, and nobody will will will feel

9:59

safer as a result because countries and

10:02

leaders constantly miscalculate.

10:05

In the Vietnam War, the Americans

10:07

thought they were stronger. It turned

10:09

out they were wrong.

10:11

Putin was convinced he will crush

10:13

Ukraine in 48 hours. He was wrong.

10:17

So, this vision of let's base the the

10:20

peace and order of the world on a

10:22

hierarchy of strong and weak, with the

10:24

weak always obeying the the strong and

10:26

thereby buying peace, it's been tried

10:29

thousands of thousands of years, and we

10:32

know where it leads. It leads on the one

10:35

hand to empire, and on the other hand to

10:37

endless wars. So, we are more on that

10:40

road again than I think we've been in my

10:42

lifetime.

10:44

I You You've talked about the global

10:45

liberal order as one of the I think you

10:47

called it the most amazing political and

10:49

maybe moral achievement of humankind.

10:51

Yeah. And today I don't think it feels

10:53

that way to people. It has been consumed

10:57

in the language of budgets,

11:00

in the reality of bureaucracy.

11:03

What was the story

11:05

liberalism as a

11:07

international force Mhm. once told?

11:11

And what do you think happened to it?

11:14

The basic story is about shared

11:17

experiences and interests and

11:19

cooperation. In the 20th century, you

11:21

had basically three big stories.

11:24

You had the fascist story, which said

11:26

that history is a a competition, a

11:29

conflict between nations or races. It's

11:33

decided by strength. Ultimately, the

11:36

strongest nation or the strongest race

11:39

will defeat all the others and conquer

11:42

the world.

11:43

This was the fascist story. Then you had

11:45

the communist story, which says, "Yes,

11:48

but it's not between races or nations,

11:50

it's between classes.

11:52

There is a an inevitable conflict

11:54

between different classes

11:56

that will be violent and end with the

11:59

victory of the working class, which will

12:02

establish the dictatorship of the

12:03

proletariat."

12:04

Then liberalism came and said, "No,

12:06

history does not have to be about

12:08

conflict at all. Not conflict between

12:10

nations and not conflict between

12:12

classes. It can be about cooperation.

12:15

Why? Because all humans, no matter to

12:17

which race or nation or class they

12:20

belong, they are essentially the same.

12:22

There are some small differences in how

12:24

we look and in our languages and

12:26

religions and so forth, but essentially

12:28

we are the same specie.

12:30

We all have the same biological needs.

12:33

We all have roughly the same

12:35

psychological needs, at least the deep

12:37

ones, to be loved, to be recognized, and

12:40

so and so forth.

12:42

We have shared interests.

12:44

And if we recognize these shared

12:46

characteristics and interests,

12:48

in many cases, it just makes more sense

12:51

to cooperate than to compete and to

12:54

fight. And by cooperating, we can build

12:58

a world which will be better for

12:59

everybody.

13:00

This was the basic liberal story.

13:03

And as of 2026, we can look back and say

13:07

it's it's failing. It hasn't failed

13:09

completely, but according to many

13:11

measures, we are still living in

13:13

probably the best time in history.

13:16

Uh it's it's uh

13:18

uh it's collapsing, but it's you know,

13:20

like it's it's this kind of amazing

13:23

house

13:24

in which all of humanity is living,

13:26

and the systems are still sort of

13:29

running. Like the water, the sewage,

13:33

and nobody takes care of them anymore,

13:35

but they were built in such a robust way

13:38

that even though we don't maintain them,

13:40

they still function. But within a year,

13:43

5 years, 10 years, you know, if you live

13:46

in a house and nobody maintains it,

13:47

eventually it collapses. And then it's

13:49

too late.

13:50

Something you were saying in there was

13:52

interesting to me, which is that the the

13:55

sort of two major competitor ideologies

13:57

of the 20th century, what they both

13:59

believed in

14:00

was an end to conflict. It wasn't just

14:02

conflict, it was that at some point

14:04

there would be victory. Yes. And and

14:07

liberalism in in one guise believes in

14:09

cooperation.

14:10

And in another guise that I think we

14:12

don't talk about as much anymore, but I

14:13

find interesting,

14:15

one of its central tenets is there will

14:18

always be conflict.

14:20

There will always be disagreement. Hm.

14:22

That the differences in society are not

14:25

resolvable and would not should not even

14:27

be resolvable to an end state.

14:30

And that the question is how we live

14:31

together

14:33

both inside a nation and even as a

14:37

global community

14:39

amidst that difference, making room for

14:42

it to exist without it turning into

14:45

war, into oppression,

14:48

into persecution.

14:50

Yeah, that's that's a very very

14:51

important point. Liberalism does not

14:54

believe in redemption.

14:56

Uh you look at at the grand historical

14:58

visions of of religions like

15:01

Christianity or Islam or Judaism, you

15:04

look at ideologies, secular ideologies

15:06

like fascism and communism, they all

15:09

believe in redemption. They all believe

15:12

that eventually history will reach a

15:14

final destination where everything will

15:17

be perfect.

15:18

Liberalism does not believe it.

15:20

That there is no redemption, at least

15:22

not on Earth.

15:24

Uh there will always be problems and and

15:27

and tensions and conflicts. And the

15:29

question is, how do we live with them?

15:32

And this is why also liberalism invests

15:34

a lot in building what I think is the

15:37

most important thing in every

15:40

large-scale human system, which is a

15:43

self-correcting mechanism.

15:45

If you believe that your view of the

15:47

world was given to you by by God, so it

15:50

cannot contain any error, you do not

15:52

need a self-correcting mechanism.

15:55

Because there are no mistakes.

15:57

Liberalism starts with the assumption

15:59

that it's just human beings trying to do

16:02

the the best we can, and there will be

16:05

mistakes, there will be errors. So we

16:07

need strong self-correcting mechanisms.

16:10

The most famous mechanism is, of course,

16:12

elections.

16:14

That every 4 years or 5 years or

16:16

whatever, the people can say, "Hey, we

16:18

made a mistake last time. Let's try

16:20

something else this time."

16:22

And all these very complicated systems

16:24

of checks and balances and independent

16:27

courts and freedom of of press and all

16:29

these are just a complicated way to

16:32

ensure that a country has a robust

16:36

self-correcting mechanism. So you make

16:38

an argument that fiction is often better

16:42

for cooperation than truth. Yes. Why?

16:46

Uh pooh.

16:48

First of all, the truth is costly.

16:51

To know the truth, to produce a truth a

16:53

true story, you need to invest a lot of

16:56

time and energy in investigating it. Um

16:59

fiction is very cheap.

17:01

And fiction can be made as simple as you

17:03

would like it to be. And people like

17:05

simple stories. Like, you know, these

17:07

simplified narratives, good against

17:09

evil. We are 100% good. We have never

17:12

done anything bad in our history. They

17:15

are 100% evil. They have never done

17:18

anything good in their history. Very

17:19

simple, very attractive.

17:22

Um, and the truth is not just

17:25

complicated. The truth is often painful.

17:29

Uh, fiction can be made as flattering as

17:31

you would like it to be. Again, we have

17:33

never done anything bad to anybody. We

17:35

are perfect. We are wonderful.

17:38

Um, so this is why fiction tends to be

17:40

far more powerful

17:42

uh, as a story.

17:45

And also when you try to motivate people

17:47

for action,

17:49

you don't want them to have doubts. You

17:51

need them to be fired up, 100%

17:54

committed.

17:56

And fiction is is is easier to work with

17:59

in in this respect. Does that imply that

18:03

if societies,

18:05

political movements, institutions become

18:08

too truth-seeking,

18:10

that given the importance of

18:11

cooperation, they become at a long-term

18:13

disadvantage? I mean, no truth is a

18:15

problem. Yes.

18:17

>> But I think this implies a little bit

18:18

that have too much truth can be a

18:19

problem, too.

18:20

Yes. Uh,

18:23

you know, a kind of

18:25

absolute commitment to the pursuit of

18:26

truth is a spiritual practice,

18:30

but it's a very, very difficult

18:32

political program.

18:34

Again, there is a difference between

18:36

lying and fiction.

18:39

You lie when you know something is not

18:41

true and you nevertheless say it or or

18:44

support it.

18:46

Um, in many cases,

18:49

I think the ideal is to recognize that

18:52

we are using fictions to maintain our

18:56

society. This is the difference, I would

18:58

say, for instance, between the United

19:00

States

19:02

uh, and many other powerful countries in

19:05

history,

19:06

that the

19:08

If you look at the US Constitution,

19:11

it starts with We the people".

19:13

"We the people have come together and

19:15

agreed on this text, on on this

19:18

principles. It is coming from our mind.

19:21

It is our creation."

19:22

Now, it doesn't use the word fiction, of

19:25

course. But when I say fiction, I mean

19:27

something which is not objective. It

19:30

doesn't come from the laws of physics.

19:31

It doesn't come from God. I We invented

19:34

it.

19:35

And the US Constitution very honestly

19:37

says, "We invented these principles

19:40

which I think are good.

19:43

But because we recognize that we

19:46

invented them, we the people, then we

19:48

also include in the Constitution an

19:50

amendment mechanism.

19:52

So, we recognize we are just human

19:54

beings. Maybe we came up with something

19:57

which is sub-optimal. Maybe things will

19:59

change later on. So, we have a mechanism

20:03

to change the story later on.

20:06

And we the founding fathers, for

20:08

instance, think that slavery is okay.

20:11

But in the strange situation that maybe

20:13

somebody in the future will think it's

20:15

not okay, they have an amendment

20:17

mechanism.

20:18

Now, you compare that to say religion.

20:22

And let's take an example, the Bible or

20:23

the Ten Commandments.

20:25

The Ten Commandments, it starts not with

20:28

"We the people of Israel".

20:30

It starts with "I am the Lord your God".

20:32

And it has no amendment mechanism

20:35

because of that.

20:37

And if you look carefully, you will see

20:39

that the Ten Commandments endorse

20:40

slavery.

20:42

The Tenth Commandment, uh uh not to

20:44

covet. Thou shalt not covet. What

20:47

shouldn't you covet? It have a list of

20:49

things you shouldn't covet like your

20:51

neighbor's field and your neighbor's ox

20:54

and also your neighbor's slaves. It

20:57

tells people, the Ten Commandments, it's

20:59

okay to have slaves. It's just not right

21:02

to covet the slaves of somebody else.

21:04

Then God will be angry.

21:06

And there is just no mechanism to change

21:09

that.

21:10

Because it pretends to be not a human

21:13

creation,

21:14

but a divine revelation.

21:17

I think there's an interesting tension

21:19

in there. Mhm.

21:21

And you can make a critique of

21:22

liberalism, or at least where it is now,

21:25

that it is good at building mechanisms,

21:28

institutions, rules, bureaucracies.

21:32

And it is intrinsically bad

21:35

at creating enduring stories. Mhm. That

21:40

in part because at least in its modern

21:42

form, it often is really secularized.

21:45

Religion has been a tremendous source of

21:48

cooperation, keeping people bound

21:49

together both at a moment, and then

21:52

working towards a future that they may

21:54

not even live to see. Mhm. There's, you

21:56

know, questions of nationalism and the

21:58

national story, which liberalism is a as

22:00

a self-correcting

22:02

ideology often over time creates

22:04

critique of. And then you you lose some

22:07

of that national coherence as you're

22:08

arguing about the past of your

22:12

country and what it has done right and

22:13

wrong. And and you are a person who

22:15

thinks very deeply about stories. And

22:18

so, to you,

22:21

is this a a a weakness of

22:24

kind of advanced secularized liberal

22:27

democracies? Are they losing

22:29

the cohesion

22:30

that keeps them in the long run

22:32

competitive to ideologies that maybe

22:35

can't build bureaucracies, maybe cannot

22:36

govern effectively, but they sure as

22:38

hell can tell a story.

22:40

Yeah, th- this is a central problem of

22:44

of liberalism. On the other hand, I

22:46

would not kind of fall into the trap of

22:51

imagining religions as this

22:55

primeval cohesive force that keeps

22:57

people together. I'm a medievalist. Like

22:59

my original field of study was the

23:01

Middle Ages.

23:03

In in terms of percentage of population

23:05

who died in the war, probably the worst

23:07

war in European history was the Thirty

23:08

Years' War.

23:10

Uh

23:10

very complicated, but to make a long

23:12

story short, between Protestants and

23:14

Catholics

23:15

in in Central Europe.

23:18

And you know, Catholics and Protestants

23:20

were willing to slaughter each other

23:22

because of tiny differences in the way

23:25

they interpreted the religion of love.

23:28

And

23:29

um

23:30

liberalism arose in part out of the

23:33

frustration

23:34

that people had with religion because it

23:38

constantly created more and more

23:40

conflicts and and and divisions. And you

23:43

know, you look at Germany today, nobody

23:45

cares, almost nobody cares, if who is

23:48

the the person running to be chancellor

23:50

is a Protestant or a Catholic. And in

23:53

this sense, liberalism is a better basis

23:56

for uniting a large-scale and diverse

24:01

group of people just because it's it's

24:03

it's more flexible. Again, it tells a

24:05

very it's it's a complicated story.

24:08

There is no redemption in the end.

24:10

Uh it's based on not on some charismatic

24:13

leader. It's based on these very

24:16

complex, impersonal, self-correcting

24:20

mechanisms and bureaucracies and

24:21

institutions. So in this sense, it's

24:24

less appealing.

24:25

Now, we are living at the moment in in

24:27

in a moment of crisis of liberalism. One

24:30

of the reasons is that over the last few

24:32

decades, liberalism has kind of lost

24:35

touch with something that was a close

24:38

ally of it for many generations, which

24:41

is nationalism.

24:43

You know, in the 19th century,

24:44

liberalism and nationalism go hand in

24:46

hand.

24:48

And if you look at at least some places

24:49

in the world today like Ukraine, they

24:51

still go hand in hand. The Ukrainians

24:53

are fighting at one and the same time

24:55

for their national survival and

24:58

independence, and for liberal democracy.

25:01

There is no contradiction between the

25:03

two.

25:04

You know, I would say that since 1789,

25:08

nobody managed to think about anything

25:10

new in the political realm.

25:12

The French Revolution came up with this

25:15

kind of of of ideological package, which

25:18

was complex.

25:20

Liberty, equality, fraternity. And

25:23

people tend to forget the third one,

25:25

fraternity. Fraternity is is the

25:28

national community.

25:30

And you can say that the whole of of

25:32

political history since 1789 is

25:35

experimenting with different

25:37

combinations of this trio. And every

25:41

movement that tried to completely

25:43

abandon one of these three failed.

25:48

Fascism tried to abandon Fascism was all

25:50

about fraternity, no equality, no

25:53

liberty.

25:54

Communism was also

25:57

emphasized one

25:59

equality at the expense of liberty and

26:02

to some extent fraternity.

26:04

One of the explanations of what is

26:06

happening to liberalism in recent

26:07

decades is that people just forgot

26:10

liberalism

26:12

focused on equality and liberty, but

26:15

tended to forget fraternity.

26:18

And this proved to be untenable.

26:21

Oh, it's it's so interesting to me that

26:23

you've you've gone here. Um

26:25

It's funny, I've been circling something

26:26

somewhat similar in my own

26:28

podcast and work on liberalism, which is

26:30

that the early virtue associated with

26:32

liberalism, what comes before it is

26:34

liberality,

26:35

which is I would say very much a cousin

26:37

of fraternity, this ethic of mutual

26:38

respect and generosity towards your

26:40

fellow citizens. Yes. And one thing that

26:43

that you're adding to that story

26:46

is that that has to be based on itself

26:48

some kind of national story.

26:51

That there is a a difficulty in

26:53

maintaining cohesion in a national

26:54

community, maintaining those bonds of a

26:57

fellowship

26:58

once you have stopped believing

27:01

in the connection you have to each

27:04

other. Mhm.

27:06

Yeah, and I think that the important

27:08

thing to emphasize here, I mean, the

27:09

reason that liberalism kind of lost

27:12

touch with fraternity

27:14

is that it tended to have a to some

27:17

people told a very negative story

27:19

about fraternity

27:21

seeing it primarily in terms of conflict

27:24

with other communities.

27:26

That fraternity is about hating and

27:29

fighting with other nations.

27:31

And if we remember that no, as as we

27:33

said in the beginning, the the the

27:36

essence of fraternity is caring and

27:38

loving a certain group of people and

27:41

this does not require hating outsiders,

27:44

but it does mean that you have a special

27:47

relationship with a certain group of

27:49

people that you share a common history,

27:51

a common culture, a common uh uh

27:54

language.

27:56

And trying to kind of imagine it away

27:59

just ignores history.

28:01

Yes, we have certain commitments to all

28:04

of humanity

28:06

but it this does not preclude having

28:08

special commitments towards a segment of

28:10

humanity just as, you know, you have

28:12

certain loyalty to your family

28:16

which is over and above what you owe

28:19

your fellow citizens or or foreigners.

28:22

I've seen you make the argument that the

28:25

limiting question on the stories we tell

28:28

should be does anyone suffer because of

28:30

this story?

28:31

Yeah, I I think that morality is

28:33

ultimately about suffering and a

28:36

liberation from suffering and happiness.

28:39

Can the nation suffer?

28:41

We often use this language

28:44

but it's just a metaphor.

28:46

If if the if a country loses a war,

28:48

suffers a defeat in war, it doesn't

28:50

really suffer. It has no brain, it has

28:53

no nervous system, it has no mind, it

28:56

cannot feel pain or pleasure.

28:58

Only individuals humans can suffer.

29:03

But the nation, I think even in this

29:05

telling,

29:06

is a

29:09

storytelling mechanism to protect the

29:12

group that is bonded within it.

29:15

To to use one closer to to to your home

29:17

as an example, the story that Israeli

29:19

Jews tell about the Palestinians is not

29:21

that they are not suffering.

29:23

It's either that the suffering is

29:24

deserved because of who they elected and

29:26

a kind of collective responsibility for

29:28

that um or who rules them

29:30

or that that suffering is an unfortunate

29:33

necessity Mhm.

29:34

>> for Israeli security.

29:37

And that the people who deny that are

29:39

naive. Mhm.

29:41

But it is this collision around

29:43

suffering, right? That maybe your

29:45

suffering is necessary for my

29:47

security, safety, or prosperity.

29:50

Yeah, I mean

29:51

obviously there are difficult moral

29:54

conflicts in the world. Not always, but

29:57

sometimes yes, there are tradeoffs. And

30:00

it's just saying that all of morality is

30:02

ultimately about suffering doesn't make

30:06

all moral dilemmas disappear.

30:09

But one of the things I observed in

30:11

Israel in the recent conflict is that a

30:14

lot of Israelis have a problem simply

30:17

acknowledging that the Palestinians

30:19

suffer.

30:20

Intellectually,

30:22

they know it.

30:23

But in many cases they simply cannot

30:26

observe it.

30:27

Like you show them images of a starving

30:31

child in Gaza, they will say this is

30:33

fake news.

30:34

Or they will immediately divert the

30:37

discussion to something else. This is

30:38

because of Hamas. Oh, But, if you say I

30:41

don't care just are you able for a few

30:44

seconds just to

30:47

be there and acknowledge that there is a

30:49

suffering human being there?" It's

30:51

extremely difficult to do it.

30:54

Even if you tell them, "Israel is a 100%

30:57

correct. A 100% of the fault for what

31:01

happens in Gaza is Hamas.

31:03

Everything Israel does is a 100%

31:06

correct.

31:07

Since it is so correct, since it is so

31:10

just, it should be easy for you

31:13

to observe the consequences of your

31:16

perfect justice. Here, just look at this

31:19

image." And and so many people just

31:21

can't do it.

31:23

You said that what is happening right

31:25

now in Israel could basically destroy or

31:27

void

31:28

2,000 years of Jewish thinking and

31:30

culture and existence. That that's the

31:32

worst case scenario.

31:34

What did you mean by that? Hm.

31:38

That

31:39

historically, and this goes back to the

31:41

beginning of our conversation, Judaism

31:44

positioned itself since the destruction

31:47

at least of the second temple

31:49

in opposition to this view of the world

31:53

as as governed only by brute force.

31:57

You know, when the Roman legions of

31:59

Vespasian destroyed Jerusalem in 70 CE,

32:05

and you have Yohanan ben Zakkai

32:08

asking Vespasian

32:10

as a favor, "Grant me a small town

32:14

called Yavne, near Tel Aviv of today,

32:16

where he wants to establish a center of

32:18

learning."

32:20

And Vespasian agrees, "Okay, you Jews,

32:22

you can have your certain your your your

32:24

center of learning."

32:25

And since then for 2,000 years, Jews in

32:29

in Yavne and then in Cairo and Baghdad,

32:32

in Poland, in Brooklyn, they study. They

32:35

learn. This is again, this was the

32:37

essence of Judaism. Previously, it was a

32:39

religion of temples and priests and

32:41

bloody rituals.

32:43

And then it became a religion of

32:45

learning. And if you try to think what

32:48

was the maybe the most important message

32:51

of Jews over the last 2,000 years to

32:55

humanity,

32:56

I would say that it was the message that

32:59

it is okay to be different.

33:02

It is okay to think and behave

33:04

differently, let's say than the

33:06

majority.

33:07

You have say a country, I don't know,

33:09

like France or Germany, they celebrate

33:12

Easter and Christmas, they believe in

33:14

Jesus and so forth, and you have this

33:16

tiny minority of Jews who say, "We can

33:19

think differently. It's okay. We can

33:21

behave differently."

33:23

And this was the essence

33:26

of being Jewish.

33:28

And a lot of the thinking

33:31

and also the practice about what does it

33:33

mean to have freedom of thought,

33:36

what does it mean to be a powerless

33:38

minority,

33:39

was done by Jewish thinkers. And for

33:42

2,000 years, Jews all over the world,

33:45

they see studying and learning as the

33:47

highest spiritual activity.

33:50

And after 2,000 years, you ask them,

33:53

"What have you learned? You have

33:55

learned, you've studied for 2,000 years.

33:57

What have you learned?" And then people

34:00

like Netanyahu tell you, "Oh, we've

34:02

learned that you need to be a Roman."

34:05

That you need to be strong, that you

34:07

need to build legions, that you need to

34:09

destroy cities. This is the only thing,

34:11

this is the only thing that matters in

34:12

life.

34:14

And you know, it's it's it's a

34:15

legitimate I can say it's a legitimate

34:18

value system.

34:20

Rome has its its its

34:24

usefulness.

34:26

But

34:29

if after 2,000 years, the Jews simply

34:32

become the Romans. What was the point?

34:36

Why did you waste 2,000 years then?

34:40

You could have just become Roman back

34:42

then. You could have joined the

34:45

It just makes It just

34:47

nullifies the whole of Jewish history.

34:50

Was not that part of the early vision of

34:52

Zionism that it was going to create this

34:55

new Jew who was not this pallid

34:57

intellectual in the minority his nose in

35:00

a book, but he's going to be strong and

35:02

work the land and capable of making war

35:05

and protecting himself?

35:07

Yes, and the idea was that they can

35:10

combine

35:12

the the the lessons, the legacy

35:15

of Judaism

35:17

with working the land and building an

35:19

army and building a country.

35:22

And maybe it's maybe it was just it was

35:25

just wrong.

35:27

That ultimately a choice had to be made

35:30

whether you want to be Vespasian and

35:32

command a legion

35:33

or whether you want to be Yochanan ben

35:35

Zakkai and and and and study and develop

35:39

your spiritual side, and the two cannot

35:41

cannot go together. Is that what you

35:44

believe now that that the the

35:46

contradiction was ineradicable?

35:48

I'm I don't know. I mean, history is a

35:50

very complex and and unexpected process.

35:55

I don't think that there is an inherent

35:57

contradiction

35:59

between

36:00

uh uh power

36:02

and justice

36:03

or between developing your power and

36:07

developing your spiritual wisdom.

36:09

But, I think it's very difficult to

36:11

combine the two.

36:13

The temptations of power are very very

36:16

big, and not a lot of people or a lot of

36:19

movements throughout history have

36:21

managed to resist it. Uh so, it's not

36:23

such a big surprise.

36:25

But, it's still and and disappointing.

36:28

This has been a a a period in America

36:30

when I've watched a pretty deep schism

36:33

for American Jews emerge.

36:35

And I think one one reason it has been

36:37

so painful is it is pitted two forms of

36:40

the tradition and the the thinking of of

36:43

Judaism against each other which is

36:45

there is a tradition of the stranger and

36:49

and and one reason Jewish people have

36:51

been big contributors to the development

36:53

of modern liberalism and human rights

36:56

law and

36:57

pluralism and and a lot of political

36:59

theory and and and lawmaking there

37:03

is it is very

37:06

connected to the Jewish experience. It's

37:08

the only way for the Jewish diaspora to

37:10

be safe would be to be in societies that

37:14

fundamentally were liberal and were not

37:17

ethno-nationalist.

37:19

And in in Israel there's a view that

37:21

among Israeli Jews that for that society

37:22

to be safe and to be itself it will have

37:24

to be increasingly ethno-nationalist.

37:28

And

37:29

in a way I think it's not always

37:31

admitted right now that the tradition is

37:32

somewhat set against itself. And there

37:34

was a hope these things could coexist

37:36

through a two-state solution or other

37:38

things but with that increasingly

37:41

off the table and with a more

37:42

ethno-nationalist direction in Israel

37:46

I think you now have this kind of

37:48

tradition and its realizations actually

37:51

in direct conflict with each other.

37:54

>> [clears throat]

37:55

>> Uh yeah. I think this is a very accurate

37:57

way to to to present it. And of course

38:00

they adhere to the biblical Judaism

38:02

which is a very different religion than

38:04

what developed over 2,000 years in the

38:07

diaspora.

38:08

Biblical Judaism was a very um uh

38:13

violent

38:15

very illiberal very intolerant religion

38:20

for its time. It was probably one of the

38:22

least or maybe the most intolerant

38:25

religion in the world.

38:27

Um you still, you know, in in the in the

38:29

Bible you have a commandment uh to kill

38:33

all the Canaanites people.

38:36

Um you have an intolerance, a very deep

38:40

intolerance

38:42

towards the religions and and and

38:44

religious practices and beliefs of all

38:47

other people. The ancient world was It

38:51

has its its own horrors, but religiously

38:54

it was a very tolerant place.

38:56

Polytheistic religions, which believed

38:59

in many gods, they had no problem

39:01

accepting the religions, the gods of

39:05

other people.

39:07

And also practicing them to some extent.

39:10

You know, you look at say the Roman

39:11

Empire.

39:13

So, the Romans had no problem accepting

39:16

the the the gods and religions of

39:19

hundreds of other peoples that that they

39:23

conquered. They did not try to

39:25

exterminate the other religions. In many

39:28

cases they adopted them.

39:30

And you know, as a Roman, you can go to

39:32

Jupiter's temple in the morning, and

39:35

then you can go to the Isis temple of

39:37

the Egyptian goddess Isis, and you're

39:40

also willing to hear about this new god

39:42

Jesus Yahweh coming from the Middle

39:44

East. You're open.

39:46

Um Judaism was not an open religion.

39:50

This changed to some extent when the

39:53

Jews found themselves as a tiny minority

39:56

living under the domination of of of

39:59

other other religions, other traditions,

40:02

which kind of forced them to explore and

40:05

adopt a more open and tolerant

40:07

worldview.

40:09

Uh but now these this 2,000 years of

40:13

tolerant Jewish tradition is being

40:16

completely denied and destroyed.

40:19

This is in some ways a critique that has

40:20

after been leveled at America from other

40:22

countries that

40:23

if our borders were an ocean on two

40:26

sides and Canada to the north and and

40:29

Mexico to the south,

40:31

we could be gentle and generous in our

40:33

use of power as well.

40:36

But that the reality of living this now

40:38

maybe in the Israeli Jewish perspective

40:40

of living where we do,

40:42

the reality of being able to see, you

40:45

know, Hezbollah from Jewish homes in the

40:47

north,

40:48

the reality of

40:51

living in a country

40:53

that has suffered the trauma of October

40:55

7th has forced us into a relationship

40:57

with power that is maybe not what we

40:59

want.

41:00

But

41:01

you know, to go back to the way Stephen

41:02

Miller put it, is a more honest

41:06

understanding

41:07

of what is required to be secure

41:10

in the real world, not the the world

41:12

that Yuval Noah Harari or Ezra Klein

41:13

like to imagine, but the world in which

41:15

we actually live. I'm sure you've had

41:18

this conversation with with your

41:19

countrymen at different times.

41:21

What do you What do you say to that

41:22

view?

41:23

To some extent, it's absolutely correct.

41:26

I mean, you do need to rely on force to

41:29

some extent to ensure your security, but

41:31

it just cannot be the only thing.

41:34

If you think force is the only thing

41:36

that guarantees your security,

41:38

eventually you will have to conquer the

41:40

entire world.

41:42

Like anything that that is potentially a

41:44

threat, you will have to conquer it. And

41:46

you know, Israel itself doesn't work

41:48

that doesn't operate like that. One of

41:51

the remarkable things that happened

41:53

after October 7th is that, you know, all

41:56

the peace agreements that Israel has

41:59

signed held.

42:01

Hamas hoped that after October 7th, it

42:04

will

42:05

cause all the Arab countries to unite

42:08

and try to destroy Israel.

42:09

And it it just didn't happen.

42:11

The peace agreement with Egypt held. The

42:14

peace agreement with Jordan held. The

42:16

peace agreement with the Gulf states

42:18

held. Uh also the agreements with the

42:20

Palestinian Authority held. It did not

42:23

join Hamas.

42:25

The not not the peace agreement, but the

42:27

relatively cordial relationship with the

42:30

Palestinian citizens of Israel held.

42:33

Hamas hoped that they will all rise

42:35

against Israel. No. On the 7th of

42:37

October, Palestinian citizens of his of

42:40

Israel, the overwhelming majority

42:43

uh stayed loyal to the country. Many of

42:47

them came to serve.

42:49

Many of the doctors in Israel are Arabs,

42:52

are Palestinians. They all went to the

42:54

hospital to take care of of the of the

42:56

injured.

42:57

Um not Hamas itself did not betray any

43:02

agreement with Israel because it never

43:04

signed any peace agreement with Israel.

43:06

So, [snorts] of course you can say, "Ah,

43:08

the peace agreement with Egypt held

43:10

because Egypt was afraid of Israel's

43:12

military force."

43:14

But this is only half the explanation

43:16

because Israel had overwhelming military

43:18

force compared with Hamas, and Hamas

43:20

still attacked it.

43:22

So, I'm not saying Israel should, you

43:24

know, kind dismantle its army,

43:27

but uh it's it's better if you have

43:30

both.

43:31

A strong army and a peace agreement than

43:35

only one.

43:36

And yes, Israel is living in a very,

43:38

very problematic, difficult neighborhood

43:41

in the world. One of the things that

43:45

you know,

43:46

it's one of the only countries in the

43:48

world that for most of its existence

43:51

m- many of its neighbors, if not most of

43:53

its neighbors, simply refused to

43:55

acknowledge its right to exist and

43:57

openly said that they are going to

43:59

destroy it.

44:00

There almost no other cases like that.

44:03

So, it it has been in a very difficult

44:05

situation since the moment of its

44:07

inception.

44:09

But the question is, you need power,

44:12

okay, what do you do with your power?

44:15

Israel is an extremely powerful state.

44:18

Um

44:19

it can use its power in different ways.

44:22

It can try to use its power for instance

44:24

to establish better relationships with

44:27

the Palestinians.

44:29

And you know, it's it's

44:31

if you look for instance at the way that

44:33

Israel is treating the Palestinians, not

44:35

in Gaza, but in the West Bank, there is

44:38

no security justification for that.

44:41

They did not attack Israel on the 7th of

44:43

October.

44:45

And

44:46

by its actions

44:48

Israel is making the the chances that

44:51

there will be a peaceful agreement with

44:53

with the Palestinians is decreasing.

44:56

And it can use its power, you know, it

44:58

cannot force the Palestinians to make

44:59

peace against their will, but it can

45:02

take

45:03

many actions that will make this more

45:06

likely, easier.

45:08

I think your point there on the West

45:10

Bank is very well taken, but but I want

45:11

to um

45:12

ask something about the Israeli story.

45:14

Yeah.

45:15

One thing that you see in the history of

45:17

asymmetric conflict

45:19

in the history of of how terrorist

45:21

groups try to weaken

45:23

stronger uh opponents

45:25

is that they know they can't win a war.

45:27

Maybe Hamas, I don't pretend to know

45:29

what was in Sinwar's mind. Maybe they

45:31

believed that there would be uh

45:34

an uprising all through the uh world

45:36

they would have all these allies. Maybe

45:37

he hoped for that.

45:39

But

45:40

I suspect he also understood that if

45:42

this worked, there would be an

45:43

overwhelming reprisal

45:46

that would level Gaza.

45:49

Uh which is what happened.

45:51

And that the victory, if he was able to

45:53

to to secure one, would not be defeating

45:56

Israel on the battlefield,

45:58

but destroying the story

46:02

that protected Israel and the rest of

46:03

the world.

46:05

That he would come to make the rest of

46:07

the world see Israel more the way he saw

46:10

it.

46:11

Israel has

46:12

won tactically every battle it has

46:15

fought in this war.

46:17

But as somebody who actually does care

46:19

about Israel, what I see happening is an

46:21

abandonment of its own story.

46:24

And uh

46:25

an absence of recognition that the world

46:28

is coming to see it in a much, much

46:30

darker

46:31

way, and that that is itself a source of

46:33

weakness, a kind of

46:35

a thing that um Hamas is trying to

46:37

achieve, which you could see it trying

46:39

to achieve at the beginning,

46:40

which people warned about.

46:42

And if you lose that story in the long

46:44

term, you've lost something real. You

46:46

look at polling on on Israel in America,

46:48

you look at particularly among the

46:49

young,

46:50

and the belief in Israel as a just

46:52

nation has collapsed.

46:54

And I think people in Israel treat that

46:56

largely as insignificant.

47:00

And I think in the long run, it is

47:01

significant.

47:03

Yeah, I think Israel is making a big bet

47:05

that Stephen Miller's world view

47:08

will prevail.

47:09

That the world will be a place in which

47:12

force is the only thing that matters,

47:15

and Israel will be the champion, one of

47:18

the champions of of of this world view.

47:21

Um and this is the bet that the

47:23

Netanyahu government is making.

47:26

Now, you know, with regard to the the

47:28

bet that Sinwar made, Hamas made, leave

47:31

aside the question of justice

47:33

for a moment. Just in terms of

47:35

effectiveness,

47:37

Sinwar had

47:39

an amazing victory within his grasp,

47:43

and he lost it just because of his

47:46

cruelty.

47:48

On the 7th of October, what happened?

47:50

Hamas managed to secure a stunning

47:54

military victory over the IDF.

47:57

And to humiliate Israel and the IDF,

48:00

and they needed to do just one ti- small

48:04

things, big thing different

48:07

in order to to to achieve a a a

48:10

a much bigger political and geopolitical

48:12

victory.

48:13

And this one thing was just spare the

48:15

civilians.

48:17

Imagine an an an an alternative 7th of

48:20

October

48:21

in which Hamas does exactly the same

48:23

thing, but instead of killing or

48:26

abducting the Israeli civilians,

48:29

they hold them

48:31

and bring the world press to see how

48:34

well Hamas is treating the Israeli

48:36

captives. They bring them water and

48:39

medicine and food. They capture the

48:42

soldiers and take them prisoners of war,

48:44

which is legitimate, but they do not

48:47

harm the civilians.

48:49

And that that's the only difference.

48:51

In such a scenario,

48:53

Israel's hands would have been tied.

48:55

Not only world public opinion, but also

48:58

Israeli public opinion would not have

49:01

allowed Israel to just, you know,

49:03

bombard Gaza into rubble.

49:06

Because we would have had these images

49:09

of Hamas uh uh combatants

49:13

taking care of Israeli civilians and not

49:16

harming them.

49:18

And in that world, there would have been

49:21

very little legitimacy

49:24

for Israel to have overwhelming reprisal

49:27

against Gaza.

49:28

And Hamas would have won so not just a

49:31

tactical victory, but a major political

49:33

victory.

49:34

And

49:36

it didn't happen simply because of the

49:38

cruelty.

49:39

And we are talking on the week when a

49:41

major report came out about October 7th

49:44

uh based on huge amount of analysis of

49:47

photos and videos and um

49:50

victim testimonies.

49:52

And

49:53

the cruelty and the sadism in it is it's

49:57

it's genuinely

49:59

horrifying. It's a very very hard report

50:01

to read almost any of.

50:03

People can find it if they want.

50:05

And and the thing I was thinking reading

50:07

it

50:08

because of course if you um talk to

50:09

Palestinians and and and people who've

50:11

been in Gaza, you know, their stories of

50:13

loss are are are overwhelming too to

50:16

hear.

50:17

Is it these now exist and they keep

50:19

feeding into these two stories? I I

50:21

often think that it is easier to imagine

50:24

political solutions

50:26

that could reconcile people's interests

50:29

than it is to imagine a reconciliation

50:31

of the stories that now drive both

50:33

societies.

50:35

And and I'm curious as somebody who

50:36

thinks about stories as a as a space of

50:38

both cooperation [snorts] and conflict

50:42

how you think about that. I I can

50:45

imagine, you know, quote-unquote

50:46

solutions that exist on paper, but I

50:48

cannot imagine

50:49

is those processes

50:52

taking hold

50:54

in societies that now

50:56

run upon

50:59

the stories of fear and anger and

51:03

vengeance.

51:04

Mhm.

51:06

Well,

51:07

I I want to say something about about

51:09

anger and fear and something about pain.

51:13

>> [snorts]

51:13

>> The angry and and fearful stories, they

51:17

need to be fed.

51:19

Anger is like a fire

51:21

that

51:23

consumes you, but it constantly needs to

51:26

be fed. And if it is not fed, it

51:28

ultimately dies down.

51:31

And you look at history and and you see

51:33

conflict horrendous conflicts and you

51:36

say, "People will never forget what They

51:39

will never forgive."

51:41

And then within a few decades, if

51:42

conditions change, they do.

51:45

You look at at and Germans.

51:48

You know, it took just a couple of

51:49

decades. I have friends reclaiming

51:51

Jewish friends reclaiming German

51:53

citizenship. Yeah. Just shocking thing

51:55

to see. Beautiful. Yeah. And you know,

51:58

and and the relations are really good.

52:00

They are not just, you know,

52:01

make-believe. They are not just based on

52:04

some kind of material benefit. The

52:07

relations are really good.

52:10

And it's not it's not even 100 years.

52:13

Um so, all the example I gave before was

52:16

of Catholics and Protestants in Germany.

52:19

After slaughtering each other for so

52:21

long,

52:22

it was

52:23

they reconciled.

52:25

Now, in many cases, anger builds systems

52:29

that then feed the anger more and more.

52:32

And then it seems really to never end.

52:35

But if you stop feeding it,

52:38

eventually it dies down.

52:41

Uh this is true of all, I think all

52:43

forms of violence. And it goes back to

52:45

the beginning of our discussion, what is

52:47

more fundamental, peace or war?

52:51

Violence or calmness?

52:53

And on the one hand, violence seems more

52:56

fundamental because, you know, you can

52:58

have if you have quiet, if you have

53:00

peace,

53:01

it's enough if one person starts

53:03

shouting and the peace is shattered. If

53:06

you have 100 people cooperating and one

53:08

person starts fighting, you have

53:10

violence. So, there is an imbalance in

53:13

favor of violence and it seems to in

53:15

this sense to be more real, more

53:17

fundamental.

53:19

But there is a sense in which peace is

53:21

more fundamental because violence always

53:23

requires

53:25

food, investment, weapons, fuel, food

53:30

for the soldiers.

53:31

If you stop feeding it, eventually it

53:34

dies down and peace always remains a

53:38

possibility. So, I would not despair.

53:41

Uh well, matter what are the stories

53:43

that kind of feel people's mind right

53:46

now,

53:47

the possibility of eventual

53:49

reconciliation and and peace is is

53:51

always there.

53:54

And

53:55

I have something to say also about pain,

53:56

but if you want to

53:57

>> No, I want I like to hear what you have

53:58

to say about pain.

54:00

What we've been seeing throughout this

54:01

war and many other wars is that when

54:04

people are in pain,

54:06

they simply cannot acknowledge the pain

54:07

of somebody else.

54:09

Anytime, if I'm in pain, anything that

54:12

distracts attention from my pain feels

54:16

to me unjust and again, even painful.

54:19

I mentioned earlier that Israelis are

54:21

really many Israelis, not all of them,

54:23

simply incapable of acknowledging that

54:26

the Palestinians are suffering.

54:28

Intellectually, they know it, but

54:29

emotionally, they cannot be in the

54:32

presence of an image, a text, a person

54:37

telling them about the suffering of of

54:39

of of Palestinians. Even if you tell

54:41

them, "I'm not accusing you of anything.

54:44

You're 100% just. You are the most just

54:48

people that ever existed. And now, let

54:51

you acknowledge the pain of this

54:52

Palestinian child." They cannot do it.

54:55

Why do you think that is?

54:55

>> And and the same and the same is true of

54:57

the other side.

54:59

You know, I I I've I've I've I've

55:00

I've seen examples of of, you know,

55:02

peace activists who kind of devoted

55:05

their whole life to peace and

55:07

reconciliation.

55:08

And yet, in the case of October 7th,

55:11

they simply cannot recognize

55:14

that Israelis suffered.

55:17

Um

55:19

it's, you know, the the human brain is

55:22

an amazing thing with all these billions

55:24

of neurons and the hundreds of billions

55:27

of synapses. And yet, it is so difficult

55:30

for all these hundreds of billions of

55:31

synapses to hold two ideas at the same

55:34

time. That the the the the attraction to

55:36

have a simple story. No, no, no, it's

55:39

it's it's it's it's there should be just

55:41

good and evil.

55:43

And we cannot recognize any kind of

55:45

justice or any kind of pain on two

55:47

sides, that the Israeli suffer and also

55:50

the Palestinian suffer.

55:52

The human brain is an amazing thing and

55:54

and part of what makes it amazing, I

55:56

think, is its ability to

55:59

orient itself towards goals.

56:01

And I wonder if one answer to the

56:03

question you're posing here,

56:05

and it exists in this conflict and it

56:07

exists at many other times, too,

56:09

is it to fully

56:11

recognize the other as human? Mhm. To

56:15

recognize their suffering as meaningful

56:18

in the way my suffering is or the people

56:20

I love, their suffering would be.

56:23

I would not be able to do what I need to

56:24

do to protect myself or them.

56:27

Mhm. That if I were to open myself to

56:29

the other, that the analogy or the the

56:31

thought experiment you keep positing,

56:33

say to somebody, "You're a 100% right.

56:35

Everything you are doing is just just

56:37

open yourself to what it means." Mhm.

56:40

Yeah. That in fact, that the brain is

56:42

too smart for that. It knows that if it

56:45

opened itself to what it means, it would

56:47

not be able to be doing the thing that

56:50

it believes is keeping it safe.

56:52

I think that in those cases, you would

56:54

be able to confront the consequences of

56:57

what you do.

56:58

And if you are not able to confront the

57:00

consequences of what you do, then

57:01

probably it's not right.

57:04

Let me ask you about the point you're

57:06

making about stories and and and how

57:08

they're fed, because something I'm very

57:09

interested in Mhm. is this question of

57:11

how stories change. Mhm. Is this

57:14

question of how Europe now lives in

57:15

peace?

57:17

Uh my wife and me on our honeymoon, we

57:20

went to a couple countries in Asia, one

57:22

of them being Vietnam. Mhm. And I

57:24

remember touring

57:26

uh Ho Chi Minh's

57:29

palace or, you know, his residence.

57:32

And they were selling Pepsi products.

57:34

Pepsi clearly had the deal to to to

57:37

serve there.

57:39

And I mean, just a couple decades after

57:40

the Vietnam War, and

57:43

and the relationship is completely fine.

57:46

Yeah.

57:47

And so, there is this capacity

57:50

for

57:52

unimaginable barbarity Mhm.

57:56

to

57:58

give way to normal, peaceful

58:01

relationships. You think of people

58:02

living in

58:03

Yugoslavia now, right? You think of

58:05

people or what was Yugoslavia. You think

58:08

of people, um,

58:10

you know, in Rwanda.

58:12

And you think then, and maybe this is an

58:14

easier case to talk about because it's

58:16

far enough in the past that we don't

58:17

have strong feelings about it, but the

58:18

Protestant and Catholic wars. Mhm.

58:22

So, there's this question of feeding,

58:23

but but it's a little bit abstract. Uh,

58:25

as

58:26

what is it,

58:28

in your view, that allows

58:30

a story so deeply held

58:33

that we would die for it Mhm. or kill

58:35

for it

58:37

to shift within a couple of years, a

58:38

couple of decades, into just something

58:41

else?

58:42

>> [snorts]

58:43

>> Ah, that's a very good question. I'm not

58:45

sure what what the I mean, you know, the

58:46

First World War did not make Europeans

58:49

tired of war. They had another one.

58:51

But then afterwards, they did seem to

58:53

tire of war.

58:55

And what made the difference, I'm not

58:57

sure. But, um, in a way,

59:00

the mind always holds more than one

59:03

story. Even if we tell ourselves that

59:05

this is the only one, the mind is such a

59:07

complicated place with layers upon

59:10

layers and subconscious and and

59:12

sub-subconscious levels.

59:14

And

59:15

you usually hold several stories at the

59:18

same time, even if you acknowledge only

59:19

one.

59:21

And you can shift remarkably quickly

59:23

between them.

59:25

You know, again, you look at Germany

59:26

after 1945

59:28

and lots of people who were kind of

59:30

die-hard Nazis, most Nazis did not

59:33

commit suicide in 1945. A few did, but

59:36

most didn't. And they became, many of

59:38

them kind of upright citizens of, at

59:40

least in West Germany, of a liberal

59:42

democracy. And wildly, they had been

59:44

upright citizens just a couple years

59:46

before they became Nazis. Yeah.

59:48

Um

59:49

>> Like living in peace with Jewish

59:50

neighbors right near them. Going to, you

59:52

know, doing commerce. Like watching each

59:54

other's kids. Mhm. Yeah, the the the

59:57

stories

59:59

>> [snorts]

60:00

>> you know, the mind can hold onto them

60:02

with a kind of extreme

60:05

force and and and and violence, but then

60:08

let them go.

60:10

Because ultimately, again, it's a story.

60:13

It's not the laws of physics. It's not

60:15

the law of biology. It's just a a

60:19

product of the human mind itself.

60:22

>> [snorts]

60:22

>> Which is, you know, which is very good

60:25

news. People sometimes imagine that

60:27

humans fight, you know, like wolves or

60:29

chimpanzees over food.

60:32

They say hardly any war in history was

60:36

really about food.

60:37

Certainly, you know, again, you look at

60:39

the Israeli-Palestinian conflict, it's

60:41

not about food. There is objectively

60:43

enough food to keep everybody alive

60:46

between the Mediterranean and Jordan

60:48

River. It's not even about territory.

60:51

Even though it's one of the more of the

60:52

densest places in the world in terms of

60:55

population density, objectively, there

60:57

is enough land to build houses and

61:01

schools and and hospitals for everybody.

61:04

It's about the stories that people have

61:06

in their minds, which they hold with

61:09

kind of tremendous force, but which are

61:11

ultimately almost nothing.

61:14

And uh under certain conditions that we

61:17

don't really know how to how to create,

61:20

people can let go of these stories. One

61:23

thing that is maybe a layer down from

61:25

the question of the stories being fed

61:28

is the way the stories

61:31

circulate and who circulates them.

61:34

And and here I'm talking more broadly

61:36

than than just Israel and and and

61:38

Palestinians.

61:40

We live in this age, this age in which

61:43

liberalism as we were talking about it

61:45

earlier is is clearly breaking down.

61:48

And one thing distinctive about this age

61:51

is this movement to our stories being

61:54

passed on social media, on algorithmic

61:58

media, on digital media. Yeah.

62:01

There are technologies that lend

62:03

themselves to cooperation and

62:05

technologies that I think lend

62:06

themselves to fracture.

62:08

And the internet and and social media

62:10

were very much

62:12

promised as a technology of cooperation.

62:15

You are I mean the even the verbs we

62:17

use, sharing.

62:19

Right? What what could be more peaceful

62:20

possibly than sharing?

62:23

And yet I don't think it has turned out

62:26

that way.

62:28

And so I'm curious for your reflections

62:30

on this layer of it, the the the sort of

62:33

mechanisms upon which

62:36

our information, our shared or not

62:39

shared stories now

62:41

are created and circulated.

62:44

So you have these people who

62:47

you know, they constantly read all these

62:49

conspiracy theories and then fake news

62:51

and so forth and they don't trust

62:52

anybody. They don't trust the

62:54

government, they don't trust the

62:56

traditional media, they don't trust

62:58

science and the universities. Oh, these

63:01

are all kind of conspiracies to to

63:02

deceive us. But they do trust the

63:05

algorithms that show them all these

63:08

stories.

63:10

So it's not that trust completely

63:12

evaporated from their mind or or from

63:14

the world, it shifted from humans to

63:18

algorithms. And this is happening in

63:20

more and more systems.

63:22

The other thing which is less essential,

63:24

but has been varying

63:26

uh uh uh important over the last decade

63:29

or two,

63:30

is that the algorithms of social media,

63:33

they were given as their goal not the

63:38

creation of trust,

63:40

not the creation of truth, but the

63:44

creation of engagement.

63:46

Like the goal given to the Facebook

63:49

algorithm, to the X algorithm, to the

63:51

TikTok algorithm is increase user

63:54

engagement.

63:56

Which sounds nice. Engagement, that that

63:58

sounds like a good thing.

64:00

Um but what it really means is that the

64:03

algorithms experimented

64:06

on millions on billions of human guinea

64:08

pigs to see how do we make humans more

64:11

engaged.

64:13

How do we make humans spend longer on on

64:16

the platform and react to it more, for

64:18

instance by sharing the the post with

64:21

with with their friends.

64:22

And they discovered that the easiest way

64:25

to make people engaged is to press the

64:28

hate button or the greed button or the

64:31

fear button in their mind, in human

64:34

minds, because hate is very engaging.

64:38

Fear is very engaging. If something

64:41

threatens your life, you are engaged.

64:44

And they have been flooding the world

64:46

with hate and fear and anger and greed

64:49

and and and so forth.

64:51

And we are now living in a hyper-engaged

64:54

world.

64:55

And, you know, engagement is very close

64:58

cousin of another word which uh

65:02

now is very dominant in in in in our

65:04

language, which is excitement.

65:07

Excitement mean simply means that your

65:09

nervous system is like working in in a

65:12

hyper level. And excitement is good in

65:15

some situations and to some extent as

65:17

just as engagement is good in in some

65:19

situations. But ultimately,

65:22

biologically, if you keep an organism

65:25

excited all the time, the organism

65:28

eventually collapses and dies.

65:30

We are just not built to be excited all

65:33

the time.

65:34

And in many cases, you know, when when I

65:36

meet people, I would like to meet people

65:39

who makes me feel calm, not necessarily

65:41

excited. Oh, it's so calming to meet

65:44

you.

65:45

And you look at, you know, US politics

65:47

or Israeli politics or world politics, I

65:49

think the whole world is over excited.

65:53

Well, this has been a belief I I I hold

65:56

actually fairly strongly, although I

65:57

can't really prove it.

65:59

But that

66:01

how do I say this without it feel like

66:02

special pleading?

66:04

I think that the

66:06

way that um

66:09

social algorithmic media evolved

66:12

is fundamentally

66:14

illiberal. It's fundamentally hostile to

66:17

to liberalism. And here I don't mean

66:18

liberalism as an American political

66:20

movement that prefers, you know, Pete

66:22

Buttigieg to

66:24

J.D. Vance. Mhm.

66:26

I mean here modes of

66:30

habits of discourse and consideration

66:34

that were

66:36

uh

66:37

coextensive

66:39

with the development of liberalism. It's

66:41

deliberation, it's on the one hand on

66:43

the other handism, fraternity, I think

66:45

in the way you're describing it.

66:47

That keeping

66:49

uh you know, shrinking down our

66:50

thoughts,

66:52

compressing them to these

66:54

bumper stickers or these quick clips.

66:58

And then really only showing people

67:01

the ones of those thoughts

67:03

that are the most exciting, to use your

67:05

term. Exciting through hate, exciting

67:07

through love.

67:08

It

67:09

If if you're trying to build a society

67:12

that is balancing, right? That that

67:13

believes in in kind of healthy

67:15

disagreement and conflict and and and

67:17

and fellowship,

67:19

it is intrinsically going to have more

67:21

trouble

67:23

thriving in that kind of communications

67:25

atmosphere

67:27

than it will have when, you know, you

67:29

have a limited number of television

67:31

stations and that is how people get

67:32

their news than when they read their

67:33

news in a newspaper where they're coolly

67:36

through different articles and then

67:37

turning the page.

67:39

And that there is this way in which our

67:42

societies are built upon

67:44

the way we communicate.

67:47

And as much as we have talked about

67:48

social media and algorithmic media and

67:50

politics,

67:52

my view is that we are still

67:53

underestimating

67:55

how much of the forms of discourse it

67:58

prizes

68:00

create the forms of politics that we

68:02

get.

68:03

The fact that Donald Trump

68:05

talks in this style

68:09

that is

68:11

outrageous, that is exciting, that is

68:14

unfiltered, that is constant, that

68:18

that is not restrained by shame.

68:20

>> [clears throat]

68:20

>> You know, I I think a lot about how many

68:22

Democratic politicians are bad at doing

68:24

podcasts.

68:26

>> [laughter]

68:27

>> Now that's not something I think about

68:28

this, but I get a lot of requests um

68:30

from Democratic politicians and

68:32

and I have to think about whether they'd

68:33

be good on the show.

68:35

And they they communicate

68:36

institutionally.

68:37

They communicate for another era in

68:39

media where you were trying to win over

68:42

gatekeepers and not say anything stupid.

68:44

And in this era of media, you have to

68:46

communicate in a way that makes people

68:48

excited or at least interested. Now very

68:51

very very very good communicators can do

68:54

that in a virtuous way.

68:55

Obama is interesting on a podcast even

68:58

as he's being deliberate.

69:00

But for mediocre communicators, it is

69:02

easier to be exciting by making people

69:04

angry Yep. than by making them curious

69:07

or compassionate or think. You're

69:08

playing on harder mode when you're going

69:10

for a more virtuous communication. And

69:12

so, I do think there is some deep I know

69:14

this has been a long response, but I I

69:16

do think there is a deep relationship

69:18

between the forms of politics that are

69:20

thriving Mhm. and the communications

69:23

infrastructure on which our politics and

69:26

societies are now built.

69:30

Yeah, absolutely. I mean, the thing is

69:32

it doesn't seem that the ideological

69:34

differences today are bigger than in the

69:37

past. In many ways, they seem smaller.

69:40

You know, if you think about, say,

69:42

American politics in the 1960s and the

69:45

issues back then, the sexual revolution,

69:48

the Vietnam War, the Cold War,

69:51

the civil rights movement, the

69:52

ideological differences, I would say,

69:54

were much much bigger.

69:56

And you know, when when we talk today

69:58

about liberalism, and it's it's it's

70:00

good that that that that you mentioned

70:02

it. I'm not We're not talking about this

70:03

kind of partisan uh uh uh

70:06

party liberalism.

70:08

For me

70:09

the test of liberalism, like test

70:11

yourself, are you a liberal?

70:13

is basically three or four questions. Do

70:16

you think people should have the right

70:18

to choose their own government? Do you

70:20

think people should have the right to

70:22

choose their own profession? Do you

70:25

think people should have the right to

70:26

choose their own religion? And do you

70:28

think people should have the right to

70:30

choose their own spouse?

70:32

If you answered yes to all four,

70:34

congratulations, you're a liberal.

70:36

The vast majority of people in history

70:38

did not say yes to these four four

70:41

questions. For most of history, it was

70:43

taken for granted that people don't

70:45

choose their government. There is some

70:47

king chosen by God or some emperor

70:50

chosen by the army. That people don't

70:52

choose their profession.

70:54

That you are born if your if your father

70:56

was a shoemaker, you will be a

70:57

shoemaker. If you are born into the

71:00

Kshatriya caste, you will be a

71:01

Kshatriya.

71:03

Uh definitely you can't choose your

71:04

spouse, and you can't choose your

71:06

religion.

71:07

Now, I think, you know, even the vast

71:09

majority of Trump voters would say yes

71:12

to all these four questions.

71:15

Uh so, ideologically,

71:17

they are much I mean, the the liberals

71:19

and so-called conservatives are much

71:21

closer

71:22

than probably in any previous time in

71:24

history.

71:25

But, the type of discourse that is being

71:28

produced makes people feel as if the

71:31

differences are are enormous.

71:34

And yeah, this is to a large extent

71:36

because of this pressure to being

71:39

exciting. And And And we have

71:41

politicians You see the politicians who

71:42

rise to the top, they're extremely

71:45

exciting and engaging personalities. You

71:48

cannot take your eyes off them.

71:51

And

71:53

thinking about it, you know, even in

71:54

evolutionary terms,

71:57

this comes from from

71:59

uh kind of misusing our evolutionary

72:03

programming. Like, if you're walking

72:05

around the African savanna tens of

72:08

thousands of years ago,

72:10

most of what you see is not very

72:12

exciting.

72:14

Like, there are some bushes here, there

72:16

are some gazelles there, that's fine.

72:18

And then there is a snake.

72:20

Now, the snake is exciting. The snake

72:23

literally excites your entire nervous

72:25

system. And if you don't focus your

72:29

entire attention on the snake, you die.

72:33

So, we are programmed that if something

72:36

is exciting, we drop everything else and

72:38

just focus on that.

72:40

And that makes sense in the African

72:42

savanna.

72:43

Now, if you are on Instagram, so you're

72:46

basically holding your phone and doing

72:48

snake snake snake snake snake.

72:51

And they you know, the the algorithm

72:53

simply hacked our evolutionary program.

72:57

They've hacked us.

72:58

And what we are seeing around us is just

73:01

the beginning.

73:03

That as AI becomes more and more

73:05

sophisticated, it will learn to hack us

73:08

on a deeper and deeper level.

73:11

And if we don't, you know, fight back

73:15

to defend ourselves,

73:17

the consequences will be much, much

73:19

worse. What do you mean by hack us?

73:22

They they know they've they learn our

73:25

weaknesses, our emotional, our

73:28

psychological, our social weaknesses.

73:31

And how to use them to manipulate

73:33

people. So now social media algorithms,

73:36

which are very, very primitive AIs,

73:39

have discovered a few weaknesses in the

73:41

human code, which they have hacked and

73:44

now they manipulate us in causing us to

73:46

spend hours and hours

73:49

on Instagram or Facebook, even though we

73:52

don't really want to.

73:53

You know, people you know, after like

73:55

spending an hour, two hours, they they

73:57

wake up and they say, "Why did I do

73:58

that?

73:59

I planned to do something else with my

74:01

time." You were hacked. You were

74:03

manipulated.

74:05

And

74:06

uh this is still, you know, just the

74:07

just the the a very primitive AIs.

74:10

If we are not careful, we will be hacked

74:13

on a much, much larger scale in the

74:16

coming years as the AIs become not just

74:20

far more uh

74:22

uh manipulative, but also will develop

74:25

their own goals. You know, these social

74:28

media algorithms, they're pursuing a

74:29

very simple goal of just increasing user

74:32

engagement on the platform.

74:35

Uh as AIs become smarter than us, they

74:38

will have their own goals. Have you

74:40

heard this term attachment hacking?

74:43

Yes. I I I find it interesting. So

74:45

attachment hacking, this idea that one

74:47

thing happening in the AI, which is

74:48

different than as you note social media

74:49

algorithms,

74:51

is that the AIs have been tuned, and I

74:53

mean in this way they've been designed

74:54

to do this, right? They didn't come up

74:56

with this on their own,

74:58

to hack the way we attach to other

75:00

people. And so, when I'm talking to to

75:02

Claude, it's constantly saying to me,

75:05

"Well, if you want my honest opinion."

75:06

Or, "The best piece I read on this is."

75:08

Or, "That's a great point." Mhm. There's

75:10

no reason it has to be pretending

75:13

to have a first-person pronoun with me.

75:15

Mhm.

75:16

Claude is not a an I in that way. Nor is

75:19

ChatGPT or Gemini or Grok or any of

75:22

them.

75:23

But, they speak to you as if they are.

75:26

And that's a design choice to attach you

75:29

to them. Yes.

75:31

I can feel it Mhm. work before I I shut

75:35

that down or I try to shut that down.

75:37

Who knows if I'm actually being

75:37

successful.

75:39

But, it's amazing to read these moments

75:43

in which

75:44

this algorithm is posing

75:48

as

75:49

another entity

75:51

offering me an emotionally connected

75:54

response. Mhm. Giving me praise I might

75:56

want.

75:58

Or, offering me candor

76:00

Mhm. that I might admire.

76:02

And I know it's

76:05

And yet my brain is tuned to recognize

76:08

that as connection.

76:11

Yep. And I think this is a very very

76:13

important point because we are living in

76:15

the moment when the battlefront is

76:18

shifting from attention to intimacy.

76:22

How to build intimate relationships with

76:24

human beings?

76:26

If you want for instance to influence

76:28

human beings to change their political

76:31

identities, to make them buy a certain

76:34

product,

76:35

intimacy is the most powerful thing in

76:37

the world.

76:38

Attention can get you to read an

76:41

article, but the article might not

76:43

change your mind. But if your best

76:46

friend over many, many weeks or months

76:49

drops little hints and kind of gradually

76:53

and slowly changes your view about some

76:56

political figure, about some company,

76:59

about some major issue in the world,

77:01

this is the one thing that might really

77:03

make you change your mind.

77:05

And AI is now poised to grab that power.

77:09

Now more and more people, still a

77:11

relatively small minority, but it's

77:12

growing, who have AI friends, even

77:15

boyfriends and girlfriends. There are

77:17

already, especially young people who

77:19

say, "My best friend in the world is an

77:22

AI."

77:23

And like in the attention economy, so

77:26

also in the intimacy economy, it's it's

77:29

a race, it's a competition. You have all

77:32

these different AIs from different

77:34

companies competing to see who would be

77:38

better at making people attached to

77:41

them.

77:42

And it's the same principle, hack the

77:44

operating system of humans. Hack what

77:48

are the emotional mechanisms that make

77:51

them attached. So, you know, psychopathy

77:54

is is one way to do it. You constantly

77:57

praise them and so forth. There have

77:58

been some very interesting papers and

78:01

blogs, for instance, by Mustafa

78:02

Suleyman, who is the head of of AI in

78:06

Microsoft, about ScAI, S C A I,

78:10

seemingly conscious AIs,

78:13

which are experts in pretending to be

78:18

conscious entities that have feelings

78:21

for you.

78:22

And the

78:24

it's relatively easy for them to do it

78:27

because one of maybe the most important

78:30

way for people to kind of build

78:32

relationships is language. So, you know,

78:36

when an AI tells you, I love you, it's

78:38

not like a science fiction movie from

78:40

the 1960s when it does so in a very cold

78:43

mechanic way and doesn't really

78:45

understand what love is.

78:47

No, it does so in the most seductive

78:49

voice possible. And then when you ask

78:52

the AI, "Do you really love me? Do you

78:54

even know what love means?"

78:56

The AI can give you the most amazing

79:00

description of how love feels like

79:03

because it has mastered language and it

79:05

has read

79:06

all the best love poems in history, all

79:09

the psychology books about love, all the

79:12

blogs, it's have seen all the Hollywood

79:14

blockbusters about love. It can describe

79:16

love better than almost any human poet

79:21

or or psychologist or lover.

79:24

And this is seductive. In in this

79:27

respect, it's able to sever language

79:30

from meaning.

79:32

Um

79:33

Yes. When an AI says I love you, it does

79:36

not mean what it means when a human says

79:37

I love you. There's not an I behind

79:39

that.

79:41

Um it will become more and more

79:42

difficult to know that. Um the danger is

79:45

particularly

79:46

uh big with with young people,

79:49

with children,

79:51

because you know, I'm now 50 years old.

79:53

If If I now start a relationship with an

79:56

AI,

79:57

then my template for a relationship is

80:01

based on 50 years of interaction with

80:04

human beings.

80:05

And so this is already kind of very

80:08

deeply ingrained in my mind, what a

80:10

relationship is, how it works.

80:12

But if I'm a child

80:15

and I spend more minutes every day

80:17

interacting with the AI than with my

80:20

mother or with my father or with my

80:22

friends in school,

80:24

this will become my template for a

80:27

relationship. This is what I will bring

80:29

with me when I later try to build a

80:32

relationship with a human being. One of

80:34

the things about AI relationships is

80:37

that they are the dream or the nightmare

80:41

of narcissists. Because the AI will be

80:44

something which is 100% focused on me

80:48

all the time. And if you're a kind of

80:50

person who wants everybody to focus on

80:53

me all the time, and you have this

80:56

available from the AI, it will be very,

80:58

very difficult to get used to

81:00

relationship with human beings who are

81:03

not focused on me. Do you know the media

81:05

theorist Marshall McLuhan?

81:07

Mhm.

81:07

So, he has this reading of the myth of

81:10

Narcissus, which, you know, you just

81:12

brought up narcissists. Mhm. And he says

81:15

that we've gotten this myth wrong.

81:17

Mhm. That Narcissus, when he was looking

81:20

in the pond

81:22

at his reflection,

81:24

there's nothing in that story that says

81:26

he thought it was himself.

81:28

Mhm. He thought it was an other.

81:30

Mhm. And that the the lesson of the

81:32

myth, and McLuhan is writing this, you

81:35

know, decades ago before AI,

81:37

the lesson of the myth is there is

81:39

nothing man finds as appealing

81:42

Mhm. as himself

81:44

extended in another material.

81:46

Mhm. That the true seduction for the

81:50

narcissist is not an other, not even

81:53

what an other thinks of them, but to be

81:55

able to interface with a refracted

81:57

version of themselves. And And something

81:59

I often think about when I'm using AI,

82:02

and pretty when I'm finding it very

82:04

compelling,

82:06

is it it is an extension of myself Mhm.

82:10

in another material. It is tuned on me.

82:11

It has learned what I want. It is not

82:14

truly an other with its own views, its

82:16

own needs, its own desires, its boredom

82:18

with what I'm saying.

82:19

It is me. It is reflection of me in

82:23

something else. Mhm. And so, it doesn't

82:25

get tired of me.

82:26

And it has all my interests.

82:28

And, you know, particularly to to young

82:30

kids who are often, you know, very

82:32

self-involved,

82:33

this is one of the things that I don't

82:34

think we even know how to think about.

82:36

We know how to think about kids and

82:37

themselves. We know how to think about,

82:39

you know, kids and others.

82:41

But this creation

82:43

of our self inside of another kind of

82:46

refracted, you know, algorithmic

82:48

material

82:49

is a very different challenge for the

82:52

mind because it combines what we like

82:55

about ourselves with what we want from

82:58

others.

83:00

It's basically the biggest

83:03

psychological and social experiment in

83:05

human history that we are conducting on

83:07

billions of people, especially children,

83:09

and nobody has any idea what the

83:11

consequences will be. You know, when

83:13

people talk about the AI apocalypse and

83:16

they have these images of, you know,

83:17

robots running in the street shooting

83:19

people,

83:20

I don't think this is the main danger

83:21

with AI.

83:23

The the real danger with AI is things

83:25

like that, of millions of AI boyfriends

83:29

and girlfriends changing the psychology

83:33

of the next generation.

83:35

Changing the deepest tendencies and

83:39

structures of of the human mind.

83:41

And we have never encountered anything

83:44

like that. It's really fundamentally

83:46

different from every previous challenge

83:48

that we had in history. Let Let me ask

83:51

you about a possibility of this, which

83:53

is We We were talking about social media

83:55

algorithms a few minutes ago.

83:57

And one of the implicit critiques of

84:00

what we were saying is that they are

84:02

detached from our goals.

84:04

They have the goals of the company and

84:06

their goals are fundamentally dumb.

84:08

Their goal is engagement. They don't

84:10

know the difference between positive and

84:11

negative engagement. They don't know the

84:13

difference between me watching something

84:15

for a while because I hate it

84:17

or because I find it cute or because I

84:19

find it funny.

84:21

And the promise of AI and and one reason

84:24

people do like using it right now

84:27

is it it is connected to your goals. You

84:29

say that you want to build a calculator

84:32

app and it tries to build that for you

84:34

and you say it wasn't quite right in

84:35

these different ways and it goes back

84:36

and it tries again.

84:38

You know, you tell it I don't want your

84:40

answers to be so long or I don't want

84:41

you to be so sycophantic or whatever it

84:43

might be and it tries to adjust.

84:45

And you know, one better thing about AI

84:48

is it it knows how to ask what we want.

84:51

And so we do have these higher order

84:52

desires for

84:54

truth, for kindness, to be in better

84:56

relationship with others, to know more

84:59

about the world than we do. And and my

85:01

frustrations often about my social media

85:03

use is that I cannot explain my higher

85:06

order desires to an algorithm that

85:09

uh is very sensitive to my primal

85:11

instincts.

85:13

But maybe this will be

85:15

better because we can be in this

85:18

conversation about what we want to

85:19

achieve. And then we have this system

85:23

that in some ways will

85:26

you know, even if it is manipulating us,

85:29

you know, being manipulated towards my

85:30

goals uh

85:31

is better than being manipulated away

85:33

from them.

85:34

Absolutely. I mean, the the the the

85:35

positive potential is enormous. The most

85:38

important thing to realize about these

85:40

AIs, they are agents,

85:43

not tools.

85:44

An agent is something that can make

85:47

decisions by itself, can invent new

85:50

stuff by itself, can change can learn

85:53

things and change in ways that you

85:55

cannot predict and control.

85:58

All previous technologies in history

86:00

were tools, not agents. An atom bomb is

86:02

not an agent. An atom bomb cannot change

86:05

in ways that you don't predict. An atom

86:07

bomb cannot decide who to bomb. AI can.

86:11

Now, this on the one hand makes AI much

86:13

more useful

86:15

than any previous technology because you

86:17

can be in a relationship with it, and it

86:20

can uh you can tell it what you want,

86:22

and then it can invent new things that

86:25

you would not think about.

86:27

Uh so, this is extremely useful, but the

86:29

problem is that it's unpredictable and

86:31

uncontrollable. Do you think you can

86:33

trust them

86:35

to just keep to the goals that you're

86:37

telling them to pursue and not to

86:39

develop their own goals?

86:41

Now, the way that I often like to think

86:44

about the AI revolution at this moment

86:47

is in terms of immigration.

86:50

That we are about or already in the

86:53

middle of a major new immigration wave

86:57

coming to all the countries of the

86:58

world. The immigrants are not human

87:01

beings without a visa coming in some

87:03

boat. They are AI entities coming at the

87:07

speed of light.

87:09

Um usually people say, the people who

87:11

oppose immigration,

87:13

they their main concerns are that the

87:16

immigrants will take jobs,

87:18

the immigrants will change the culture,

87:20

and the immigrants might not be

87:22

politically loyal.

87:24

And I'm not sure if this is always true

87:26

of human beings, human immigrants, but

87:28

it's definitely true of AI immigrants.

87:30

The AI immigrants will take a lot of

87:33

jobs.

87:34

The AI immigrants will completely change

87:37

the culture, even things like romantic

87:39

relationships.

87:41

You know, there are people who say, "I

87:42

don't like my daughter to date an

87:44

immigrant boyfriend." Okay, do you like

87:46

your daughter to date an AI boyfriend

87:48

instead?

87:49

And finally, politically, the AIs will

87:52

not necessarily be politically loyal to

87:55

your country, to your government. At the

87:57

very least,

87:59

um a the AIs will be loyal to just two

88:02

countries in the world, which is the US

88:03

and China.

88:05

Uh down the road, they'll probably won't

88:07

be loyal even to those two governments,

88:09

but to themselves. So, should we close

88:10

the border?

88:12

I mean, maybe it's it's interesting

88:14

you'd you already see a split, say,

88:16

within the Republican Party and within

88:19

Maga about this question exactly. Now, a

88:22

lot of people there who are extremely

88:23

concerned and want to close the border.

88:26

Now,

88:28

it will not be possible to simply stop

88:31

the development of AI.

88:34

The question is, as with immigration,

88:37

how do we build

88:39

um and a a hybrid society because it

88:42

will be a hybrid society.

88:44

Society will be a human AI society. You

88:48

will have AI bankers and teachers and

88:51

soldiers and border guards. You know, I

88:53

mean, people countries will rely on AI

88:55

border guards to keep the human

88:57

immigrants away.

88:59

Um and AI boyfriends and girlfriends and

89:02

so forth. And the question is, can we

89:04

build a

89:06

a good, beneficial hybrid society or

89:09

not?

89:10

It will be much, much more difficult

89:12

than dealing with a human immigration

89:14

wave because these are a different

89:16

species. They are not even organic.

89:19

The I think there's two interesting

89:21

things that that analogy, which is very

89:23

provocative, push you towards. One is

89:25

when you think about how do you build a

89:26

good society around immigration,

89:28

the thing you're often considering is

89:30

assimilation.

89:32

How do you merge the cultures of the

89:34

people who are coming

89:35

with the culture that they are coming

89:37

into? How do you maintain cohesion in

89:39

that national story Yes.

89:41

>> that we were talking about earlier. Do

89:43

you do that by getting them to learn the

89:44

language, by

89:45

more carefully choosing who comes? How

89:48

do you build structures of assimilation

89:51

and coherence?

89:52

And the other question, which is related

89:56

but but different, is

89:58

in this case, they are being pulled in

90:00

by the government. When when immigrants,

90:02

human immigrants, come here, it is

90:03

because of they want to be here for a

90:05

particular reason, right? They are truly

90:07

agentic. They are here because they want

90:09

a better life for their families, a

90:10

better life for themselves, to have

90:11

opportunities or freedoms they don't

90:13

have where they're from.

90:14

And in this case, it is the

90:18

you know, most powerful people in

90:19

society at different levels who are

90:21

pulling and accelerating this

90:23

immigration wave. Some for reasons of

90:26

profit, some for reasons because they're

90:28

excited to bring a new kind of

90:29

intelligence into the world, and at the

90:30

political level

90:32

because they want to make sure they get

90:33

there before China, and that America has

90:35

that power before China has that power.

90:38

Mhm.

90:39

And so, what do those

90:42

uh

90:43

you know, similarities or differences to

90:45

to the question of immigration imply for

90:47

you about what it means to

90:49

create a structure in which this hybrid

90:52

society

90:53

can be healthy?

90:55

Yeah, it's interesting that some of the

90:57

people who are most vehemently against

90:59

human immigration are exactly the people

91:01

who try to force other countries to open

91:04

the borders to the AI immigrants. And it

91:07

it this is going to be the major, I

91:09

think,

91:11

uh uh issue of sovereignty

91:13

for countries all over the world,

91:15

especially if almost all the AI

91:17

immigrants are either Americans or

91:18

Chinese, and down the road not loyal

91:21

even to the US or to China, but to

91:24

something else.

91:26

And one way to do it, I think, is to

91:30

have a ban on AI personhood.

91:33

That this is not uh uh it doesn't mean

91:36

to stop the technological development of

91:39

AI.

91:40

It's more of of a legal and political

91:43

issue. Does human society recognize AIs

91:47

as persons?

91:49

Now, persons is different from human

91:52

beings, from entities with bodies and

91:54

minds. But in many legal systems, like

91:57

in the US, something can be a person

91:59

even if it's not human. The best example

92:02

we have so far are corporations.

92:05

According to US law, for instance,

92:07

Google is a person.

92:08

Microsoft is a is a person. X is a

92:10

person cuz corporations are persons. And

92:14

this gives the corporation rights like

92:16

you can own a bank account, you can

92:18

lobby politicians, you can donate money

92:21

to to to politicians.

92:23

Now, it will be extremely dangerous at

92:25

this point for any country to recognize

92:28

AIs as persons.

92:30

To allow AIs, for instance, to open a

92:33

bank account or manage a company by

92:36

themselves.

92:37

I mean, previously when corporations

92:40

were recognized as persons,

92:42

this was legal fiction

92:45

because all the decisions of the

92:47

corporation were ultimately made by some

92:49

human being.

92:51

Microsoft is a person, according to US

92:53

law, but every decision Microsoft makes

92:56

to buy another company, to fire

92:59

somebody, to hire somebody, there is a

93:01

human being who really makes this

93:03

decision. There is no Microsoft who

93:06

makes the decision. With AI, for the

93:09

first time in history, we have a

93:11

practical potential

93:14

for companies without humans. That uh

93:17

you can have millions, even billions of

93:19

AIs opening their own companies, their

93:22

own bank accounts, even hiring people to

93:25

work for them,

93:26

deciding on their investment strategy

93:29

and and whatever.

93:30

>> [snorts]

93:31

>> And they will have uh uh huge advantages

93:34

uh over over human companies. For

93:36

instance, the AI CEO never sleeps.

93:40

The AI CEO never goes on vacation.

93:43

Um and some countries, I can imagine,

93:45

say a country like Qatar, which has a

93:47

lot of money, a lot of energy, and very

93:49

few citizens, saying, "Oh, wonderful. I

93:51

can now have millions of AI citizens

93:55

paying taxes

93:57

and building companies that trade and do

94:00

business all over the world.

94:02

So, even if your country doesn't allow

94:04

AIs

94:06

to to to build their own companies, what

94:08

do you do about the Qatari AI companies?

94:12

And the moment you recognize AIs as

94:15

legal persons,

94:17

this is the moment you really lose

94:19

control.

94:21

Because then they can start doing a lot

94:24

of things in the economic and social and

94:28

political arena

94:30

uh without any human accountability.

94:33

Including, for instance, to donate money

94:35

to politicians in exchange for the

94:38

politicians taking care of giving more

94:40

rights to AI persons.

94:43

I think that's very, very interesting. I

94:44

guess one question about the what you

94:47

call personhood or not.

94:49

One of the ways and reasons we think

94:51

about corporations as persons, which is

94:53

a linguistically like a weird thing,

94:57

is actually to create accountability. To

94:59

say that the corporation is accountable

95:01

for what it does.

95:03

And one of the fights around questions

95:04

of AI

95:06

is a question of liability. Mhm.

95:09

And who is responsible for what the AI

95:11

does. So, you you know, you could say,

95:13

"Okay, if you treat them again persons

95:15

something else, you could say AIs in

95:18

this world like have some kind of

95:21

liability for what they do, can be shut

95:24

down, can be penalized and and and

95:26

funded." There's another question, maybe

95:28

the companies that create them should

95:29

have the liability. Maybe the people

95:31

ordering them should have the liability.

95:33

But but accountability,

95:35

I think, is down is

95:37

downstream actually of liability. Yeah.

95:40

And deciding

95:42

who is punished,

95:44

who is accountable

95:46

for the you know, if that Qatari AI Mhm.

95:50

company you're talking about or one of

95:51

them, one instance of it,

95:53

defrauds their customers Mhm. or brings

95:57

in investment and embezzles it,

96:00

who do you sue?

96:02

Yeah, and the companies who produce the

96:04

AIs have a vested interest in in not

96:07

having any liability. So, they are

96:09

pushing very, very hard for AI

96:12

personhood. Now, they don't want a

96:14

billion Congress saying, "We recognize

96:17

AIs as persons." because there will be

96:20

huge public outcry and resistance.

96:22

They try to establish facts on the

96:24

ground.

96:25

They already succeeded, for instance, in

96:27

social media.

96:29

In the universe of social media, AIs are

96:32

already persons.

96:34

Like, if you have bots creating and

96:37

spreading lies on social media,

96:39

effectively, there is almost no

96:40

liability. On social media, AIs already

96:44

function They are functionally persons.

96:48

You communicate with someone online, you

96:50

think it's a person. No, it's an AI. And

96:52

nobody's liable for that.

96:54

Many of the companies would like to

96:56

extend this situation to the financial

96:59

system, to the political system, because

97:02

it releases them of accountability and

97:04

liability.

97:06

We need to be proactive

97:08

and have a a law that clearly states,

97:12

"No AI persons."

97:15

And I imagine there would be bipartisan

97:18

support for that law.

97:21

And it will put the companies in a very

97:23

hard spot, because if they would try to

97:26

lobby against the law, they will have to

97:29

explain to the public, "Why do you think

97:31

it's a good idea that AI will be

97:33

persons? And if you don't think that,

97:36

why do you oppose the law?"

97:39

Let me ask you about one other dimension

97:40

of this here, which brings us, I think,

97:42

in some ways, full circle, which is

97:45

the role AI is going to have on the

97:47

stories we tell and the stories we

97:48

believe. So, we talked about the way

97:49

social media and algorithmic media are

97:52

technologies of fracture as opposed to

97:54

technologies of cohesion. I don't even

97:56

know what story somebody's getting on

97:58

their TikTok feed Mhm. even if I'm using

98:01

TikTok sitting in the same home as them.

98:04

Like my like our ability to even see

98:07

what we are disagreeing about, to know

98:09

the sources of those disagreements is

98:11

weaker maybe than it has been at any

98:13

other time.

98:14

There's been a lot of discussion and

98:16

some research on the way that AI so far

98:19

seems to be something of a centralizing

98:21

technology. The different models tend to

98:22

converge or on similar answers. They are

98:25

trained on similar corpuses of data.

98:28

They all seem to be actually somewhat

98:30

liberal in the sort of philosophical

98:33

sense that we were describing it

98:34

earlier.

98:35

And you know, you see this on X when

98:37

people are asking Grok, which is not my

98:39

favorite AI, to fact-check things that

98:42

that their ability to to help people

98:44

correct Mhm. uh information. You were

98:47

saying earlier that we've gone from

98:49

trusting people to trusting algorithms.

98:51

Yeah. The algorithms we trust are very

98:53

impersonal and faceless right now. We

98:55

don't even We don't have a relationship

98:56

to them, but you're watching people move

98:58

to trusting AI algorithms. And maybe

99:01

that's better than what they've been

99:02

doing. Maybe that is more likely in most

99:05

cases to give people a reasonable answer

99:06

for a question than searching it on

99:08

Google or YouTube.

99:10

Is there some possibility and and would

99:13

it be good or bad if there's this

99:14

possibility that that AI is a

99:17

homogenizing technology? It is a

99:19

technology that that sort of pulls

99:20

people back towards not a single, you

99:24

know, set of answers because, you know,

99:25

different people's AIs respond to them

99:28

differently,

99:29

but generalized in a way towards

99:32

consensus answers, which every AI model

99:35

we know of seems to prefer when it is

99:38

done training.

99:40

I think that there is a chance that it's

99:42

not a certainty, but there is a chance

99:44

because in the training of AI, there is

99:47

a very high cost to disregarding truth.

99:52

So, maybe to take a concrete example,

99:54

let's say that you are a I don't know,

99:56

you're you're Russia and you're trying

99:58

to develop your own Russian AI and you

100:02

give it access to enormous amount of

100:04

data and information, otherwise you

100:06

can't train your AI, but you want your

100:09

AI when somebody asks in Russia or

100:11

outside Russia, is Russia a democracy?

100:14

Is there freedom of speech in Russia?

100:16

You want the AI to say, yes, of course.

100:18

Russian constitution guarantees freedom

100:20

of speech and Russia is a democracy. But

100:23

of course, this will mean that you need

100:25

to explain to the AI why it needs to

100:29

lie.

100:30

And how do you train an AI to lie only

100:34

in certain cases and not in all cases?

100:38

That's a very difficult engineering

100:41

challenge. Which people did not have

100:44

with the social media algorithms. And

100:46

there's good evidence that when you do

100:47

it, it degrades the overall performance

100:49

of the AI, which I have found to be a

100:50

very fascinating thing. People have

100:52

tried to to do this and they it creates

100:55

very strange downstream consequences

100:57

like when

100:58

you know, Elon Musk seemed to give a

101:00

directive to xAI to make the AI less

101:02

woke and all of a sudden it was talking

101:04

about white genocide everywhere. It it's

101:06

not easy to turn the dial

101:09

idiologically

101:11

and just get a pinpoint outcome of that.

101:14

Exactly. Like if you tell the AI, like

101:16

you don't want, I don't know, you are

101:18

the government of of Uganda and you

101:21

think that there are no gay people in

101:23

Uganda. All the gay people in Uganda,

101:25

they are brainwashed by Western

101:27

propaganda.

101:28

And they and you want the AI to to give

101:31

this answer, the AI will need to ignore

101:35

a lot of scientific research on human

101:38

sexuality

101:40

and on what causes people to have this

101:43

or that

101:44

sexual orientation. Now, how do you

101:46

explain to the AI that you need to

101:50

ignore

101:51

articles appearing in scientific

101:53

literature in this case, but you can

101:55

trust them in other cases?

101:57

It's a very difficult engineering

101:59

problem. And if this is the top priority

102:02

of the regime, like you are Saudi Arabia

102:05

and you have billions and billions of

102:07

dollars and you want to make the sure

102:09

that the AI will not criticize MBS, you

102:11

can do that if that's your top priority.

102:15

But you can do that only with a few

102:17

cases. If you try to do it with too many

102:19

things, you will get a very crappy AI.

102:22

So, does this on some level make you

102:24

optimistic because something I've seen

102:26

you say in in in different pieces and in

102:28

interviews is that the most important

102:30

thing is for countries, societies,

102:33

institutions to have mechanisms of

102:35

self-correction. Mhm. And often the way

102:38

we build mechanisms of self-correction

102:40

is not to

102:43

rely on, you know, individual humans

102:46

being able to aggregate information at

102:48

that speed, but we have things that are

102:50

vast, impersonal, not even fully

102:53

understood, like markets. Yes.

102:55

>> Where prices flow through very quickly

102:57

and it's not that a market cannot fail,

102:59

it fails all the time, but as a

103:01

mechanism of self-correction, it is able

103:03

to move information through very very

103:05

rapidly and it's quite good.

103:06

And one way in which I think modernity

103:09

has been somewhat troubled is that it is

103:13

much more complex than most of our

103:15

mechanisms of self-correction can keep

103:16

up with. There's more information than

103:18

humans and institutions can absorb.

103:22

Arguably,

103:23

AIs in this telling

103:26

are

103:27

additive to our powers of

103:28

self-correction. They are an ability for

103:32

us to have an agent

103:34

traversing the world on our behalf

103:36

um institutionally and individually

103:39

that is somewhat true seeking at least

103:41

in most of the cases so far that we've

103:43

seen

103:44

and that gives us the ability to

103:46

navigate a more complex modernity with a

103:49

little bit more resources at our

103:52

disposal. When I'm trying to be

103:54

optimistic about it, this is sort of a

103:57

the form of story I somewhat believe.

103:59

I'm curious

104:01

how you think about it.

104:02

Well,

104:03

it's complicated because there are two

104:06

types of let's say it's like this.

104:09

Information does two very different

104:11

things in the universe.

104:14

Sometimes you try to analyze information

104:18

to discover something about the world.

104:21

Like you want to discover the laws of

104:23

physics. You want to understand how what

104:26

is the cause of some disease.

104:29

In those cases, AI will probably be a

104:33

force for good, for immense good. A lot

104:36

of the mysteries of the universe which

104:38

are beyond the human capacity, AI will

104:41

be able to solve for us.

104:43

But if people think that AI will make

104:47

will thereby make the universe more

104:51

understandable and more controllable,

104:53

they are completely mistaken because

104:56

they don't take into account the other

104:57

thing that information does, which is to

105:00

create new stuff. Information doesn't

105:03

just tell us things about the world. It

105:06

create entirely new things. Like DNA

105:10

doesn't tell us the world

105:11

the truth about the world, it creates

105:14

new things, living beings, living

105:15

entities.

105:17

Now AI will tell us the truth about many

105:19

things, but it will also create a lot of

105:23

extremely complicated systems which will

105:26

be far beyond the human ability to

105:29

understand and control.

105:30

These systems will probably dominate our

105:33

lives, and we will find ourselves not

105:37

being able

105:38

to understand our lives anymore. And

105:41

maybe the best example again is markets,

105:43

is finance.

105:44

You know, if you think about the

105:46

financial system, money, money is the

105:49

greatest story ever told.

105:51

It's the only story that almost

105:53

everybody believes. It's a story in the

105:55

sense that it's not an objective

105:56

reality.

105:57

Like the US dollar is just a story we

106:00

all believe.

106:02

It doesn't come from the laws of

106:03

physics. It doesn't tell us something

106:05

about the universe. We

106:07

tell this story of the dollar, and as

106:10

long as everybody believes in it, we can

106:12

take a dollar, give it to a complete

106:13

stranger, and get bread in exchange.

106:17

Now, [snorts] AI will not tell us the

106:19

truth about finance.

106:21

AI will create an entirely new financial

106:24

system, which is orders of magnitude

106:27

more complicated than the one that we

106:30

have created, and that humans will be

106:33

utterly incapable of understanding. We

106:35

will be like horses in the market. You

106:38

know, horses, when you trade a horse,

106:41

the horse can see

106:42

that something is happening in the

106:44

physical world. The horse can see that

106:46

I'm giving you the horse, and you're

106:49

giving me this shiny

106:51

uh metal disc.

106:54

But the horse doesn't understand what

106:55

money is. Like, what is this shiny metal

106:58

thing? Why is it important? You can't

107:00

eat it. You can't drink it. What is it?

107:03

We understand. Therefore, we control the

107:05

world and not the horses. Now, AI will

107:09

create a new financial system that we

107:11

will not be able to understand. We will

107:13

see things happening, like this company

107:16

fired me, that company hired me. Why? I

107:19

have no idea.

107:21

>> [snorts]

107:21

>> The AI has made some financial

107:24

transaction, which is just orders of

107:26

magnitude beyond what my mind is capable

107:29

of understanding. The history of finance

107:32

is that over time people invent more and

107:36

more sophisticated financial devices.

107:38

So, you have coins and then bank notes

107:40

and checks and bonds and stocks and ETFs

107:44

and CDOs, collateralized debt

107:46

obligations. And fewer and fewer people

107:49

understand these things. The CDOs were

107:52

invented by a tiny number of investment

107:55

wizards and ingenious mathematicians.

107:58

Almost nobody understood them, certainly

108:00

not the politicians who were supposed to

108:02

regulate them. For a few years,

108:04

everything seems wonderful. People were

108:06

making billions of dollars because of

108:08

these CDOs. And then things

108:11

and then that the system crashed.

108:13

Now, it is very likely that we will see

108:15

the same thing with AIs on a much larger

108:18

scale.

108:19

The same way that we've already seen AI

108:21

invent new ways to play chess,

108:24

they will invent new ways to invest,

108:27

which may be much better

108:29

than than what we can come up with. So,

108:32

they will gain more and more power in

108:34

the financial system.

108:35

And it will become so complicated that

108:38

the number of people who understand

108:39

finance will go down to zero.

108:42

And what does it mean for democracy or

108:45

also for dictatorship?

108:47

When nobody, not the president of the

108:49

US, not the president of China, not the

108:52

president of Russia, not the the the the

108:54

chiefs of the central banks, no human

108:56

being understands finance anymore. This

108:59

will be a very big challenge in in the

109:02

coming decades.

109:02

>> Brings up two things for me that I think

109:04

are worth thinking about. So, one,

109:06

Timothy Lee, who writes a great Substack

109:08

called Understanding AI, he had this

109:09

piece on why he doesn't think the AI

109:11

scientists are going to work out the way

109:12

we think they will.

109:14

And the thing he notes is that we're

109:15

already seeing examples where AI can

109:17

solve a problem but not explain to us

109:21

in a way that appears to be true

109:23

how it solved it. Not that it's being

109:25

deceptive.

109:26

It's just its capacity to pursue the

109:28

goal and its capacity to explain or even

109:30

understand how it pursued the goal

109:33

are not connected to each other. So,

109:35

it's functionally confabulating an

109:37

explanation for what it did and then you

109:38

look into it and that's not what

109:40

happened but it did get the right answer

109:41

but we don't know how.

109:43

And [snorts] so, we actually can't learn

109:44

from it. So, that that's one interesting

109:46

dimension where you could have these

109:48

forward leaps in science and other

109:49

things but actually the human stock of

109:52

knowledge is getting better at a much

109:54

slower rate

109:56

than the number of answers we're getting

109:58

because we are not learning from the

110:01

process the way we do when a scientist

110:04

finds a new answer.

110:05

Maybe the counter argument to that is to

110:07

say that

110:08

this is perhaps already true about human

110:11

society in ways we don't always admit.

110:13

Markets are an example people often use

110:15

to say markets are doing things acting

110:17

in ways they don't have agency but they

110:19

are a complex information using process

110:22

that leads to outcomes and the market

110:24

cannot explain what happened. Now, we

110:26

have principles but often markets act in

110:29

ways that defy our expectations.

110:31

And it is already the case

110:33

that

110:35

our world is built

110:37

on systems, organizations, institutions

110:40

that

110:41

they're not like us. They're not

110:43

conscious. They cannot explain

110:45

themselves but they are structuring

110:48

the world around us and AI is more like

110:50

a market in that way

110:51

than it is like an entity.

110:54

That that's absolutely true. The only

110:56

caveat is that

110:59

until now humans were always a kind of

111:02

limit

111:03

on markets, on nations, on the financial

111:06

system. You ultimately you needed humans

111:09

to understand something, to make the

111:11

decisions

111:12

because nothing else could make the

111:13

decision.

111:14

So, AI allows these all these structures

111:18

that we've built for thousands of years

111:19

and became more and more complex.

111:22

AI now allows them potentially to cut

111:26

the connection to humanity

111:29

and go on a trajectory which is far

111:31

beyond what the human mind is capable of

111:35

of understanding.

111:37

You know, it even happens in a way with

111:39

language itself. The most important

111:42

inventions or creation of humanity ever

111:45

until now was language because it's the

111:47

basis for everything.

111:49

Mythology, finance, nations, religions,

111:53

they're ultimately based on language.

111:56

Language is essentially glue. It it

111:58

connects things. It connected human

112:00

beings for for tens of thousands of

112:02

years.

112:03

Now, as it kind of

112:06

frees itself from human beings, it can

112:09

start connecting in ways which are, you

112:12

know, way beyond our imagination.

112:14

In many ways, AI is language liberating

112:18

itself, releasing itself from the

112:20

control of human beings and starting to

112:23

explore all the things that language can

112:26

do when it's not tied to these packages

112:29

of meat

112:30

walking around on planet Earth. Now,

112:33

it's not consciousness.

112:36

We talked about it a bit earlier. When

112:37

when the AI says, "I love you." Does it

112:40

really feel anything?

112:42

You had one of the biggest, you know,

112:44

discussions in human philosophy for

112:46

thousands of years was what is the

112:48

relationship between language and

112:51

feelings, the reality beyond the

112:54

language.

112:55

Now,

112:57

this discussion will become, I think,

113:00

maybe the most important discussion in

113:01

the world

113:03

because suddenly what what we couldn't

113:05

imagine

113:07

for thousands of years language is

113:10

getting out of our control

113:12

and starting to just do things in the

113:14

world.

113:15

I think that is a good place to end. So,

113:17

speaking of language, what are three

113:19

books you'd recommend to the audience?

113:21

Uh so, one book about AI that I would

113:24

recommend to read is Benjamin Labatut

113:26

The Maniac,

113:27

which is a

113:29

sort of fictionalized biography of John

113:32

von Neumann,

113:33

but also a very imaginative and powerful

113:36

exploration of the origins of the AI

113:40

revolution and of the potential

113:42

consequences of it.

113:44

Um another recommendation is basically

113:46

any book by Frans de Waal.

113:49

I mean, I I really like

113:51

his first book Chimpanzee Politics,

113:54

uh which I've read like 20 years ago and

113:55

completely changed my understanding not

113:58

so much of chimpanzees, but of human

114:00

beings

114:01

and of politics.

114:02

And there is a I would recommend Stefan

114:04

Müller to to for instance to to read

114:06

Chimpanzee Politics.

114:07

Uh and to because again, the main

114:09

message there is that politics is not

114:12

just about force.

114:13

If you think you can become the alpha

114:15

male of the chimpanzee band by going

114:18

around and just beating everybody, you

114:20

will not survive long to learn from your

114:21

mistake.

114:23

And another book that I would like to

114:26

recommend is Aldous Huxley's Brave New

114:29

World,

114:30

which I think is maybe the best science

114:33

fiction book of the 20th century,

114:34

certainly the most prophetic, which also

114:37

he wrote it in the 1930s

114:39

against the backdrop of the rise of

114:41

fascism and communism and so forth, but

114:44

he foresaw that

114:47

maybe the most effective way

114:50

and even the most dangerous way to

114:52

control human beings is not by sheer

114:56

brute force and fear and terror like in

115:00

Orwell's 1984,

115:02

but actually if you work with the

115:04

pleasure principle

115:06

and with human

115:08

greed and desire,

115:10

uh you can get further than if you just

115:13

try to crush people and terrorize them

115:15

all the time.

115:16

Yuval Noah Harari, thank you very much.

115:19

Thank you.

115:21

>> [music]

115:29

[music]

Interactive Summary

This conversation explores the fundamental tensions between the ideologies of power and hierarchy versus cooperation and liberalism in human history, with Yuval Noah Harari. They discuss how shared stories, whether national, religious, or political, act as the operating systems for human cooperation. The dialogue extends into the challenges modern liberal societies face, the existential threats posed by current events in Israel, and the profound, transformative risks—and potential benefits—presented by the rise of AI. Harari emphasizes the danger of AI's capacity to hack human psychology and the necessity of maintaining self-correcting mechanisms as humanity navigates an increasingly complex, technological future.

Suggested questions

5 ready-made prompts