Richard Haier: IQ Tests, Human Intelligence, and Group Differences

Lex Fridman (01:29.880)

and how do we measure it?

Richard Haier (01:31.560)

Everybody has an idea of what they mean by intelligence.

Lex Fridman (01:35.840)

In the vernacular, what I mean by intelligence

Richard Haier (01:40.040)

is just being smart, how well you reason,

Lex Fridman (01:42.560)

how well you figure things out,

Lex Fridman (01:45.160)

what you do when you don't know what to do.

Lex Fridman (01:48.680)

Those are just kind of everyday common sense definitions

Richard Haier (01:53.680)

of how people use the word intelligence.

Lex Fridman (01:56.660)

If you wanna do research on intelligence,

Richard Haier (01:59.500)

measuring something that you can study scientifically

Lex Fridman (02:03.360)

is a little trickier, and what almost all researchers

Richard Haier (02:09.120)

who study intelligence use is the concept

Lex Fridman (02:12.820)

called the G factor, general intelligence,

Lex Fridman (02:16.980)

and that is what is common, that is a mental ability

Lex Fridman (02:21.640)

that is common to virtually all tests of mental abilities.

Richard Haier (02:26.500)

What's the origin of the term G factor,

Lex Fridman (02:28.500)

by the way, such a funny word

Lex Fridman (02:29.820)

for such a fundamental human thing?

Lex Fridman (02:32.060)

The general factor, I really started with Charles Spearman,

Lex Fridman (02:36.540)

and he noticed, this is like, boy,

Lex Fridman (02:39.060)

more than 100 years ago, he noticed that

Richard Haier (02:44.140)

when you tested people with different tests,

Lex Fridman (02:47.780)

all the tests were correlated positively,

Lex Fridman (02:53.100)

and so he was looking at student exams and things,

Lex Fridman (02:57.380)

and he invented the correlation coefficient, essentially,

Lex Fridman (03:01.620)

and when he used it to look at student performance

Lex Fridman (03:06.380)

on various topics, he found all the scores

Richard Haier (03:09.920)

were correlated with each other,

Lex Fridman (03:11.920)

and they were all positive correlations,

Lex Fridman (03:14.480)

so he inferred from this that there must be

Lex Fridman (03:17.540)

some common factor that was irrespective

Richard Haier (03:21.220)

of the content of the test.

Lex Fridman (03:23.360)

And positive correlation means if you do well

Richard Haier (03:27.420)

on the first test, you're likely to do well

Lex Fridman (03:29.500)

on the second test, and presumably,

Richard Haier (03:31.760)

that holds for tests across even disciplines,

Lex Fridman (03:35.820)

so not within subject, but across subjects,

Lex Fridman (03:39.340)

so that's where the general comes in,

Lex Fridman (03:43.020)

something about general intelligence.

Lex Fridman (03:45.020)

So when you were talking about measuring intelligence

Lex Fridman (03:46.980)

and trying to figure out something difficult

Richard Haier (03:50.140)

about this world and how to solve the puzzles

Lex Fridman (03:52.300)

of this world, that means, generally speaking,

Richard Haier (03:54.740)

not some specific test, but across all tests.

Lex Fridman (03:58.140)

Absolutely right, and people get hung up on this

Richard Haier (04:02.340)

because they say, well, what about the ability

Lex Fridman (04:04.860)

to do X, isn't that independent?

Lex Fridman (04:08.860)

And they said, I know somebody who's very good at this

Lex Fridman (04:12.000)

but not so good at this, this other thing.

Lex Fridman (04:15.360)

And so there are a lot of examples like that,

Lex Fridman (04:17.460)

but it's a general tendency, so exceptions

Richard Haier (04:21.660)

really don't disprove, your everyday experience

Lex Fridman (04:26.500)

is not the same as what the data actually show.

Lex Fridman (04:30.740)

And your everyday experience, when you say,

Lex Fridman (04:32.700)

oh, I know someone who's good at X, but not so good at Y,

Richard Haier (04:36.580)

that doesn't contradict the statement of about,

Lex Fridman (04:39.100)

he's not so good, but he's not the opposite.

Richard Haier (04:43.380)

He's not, it's not a negative correlation.

Lex Fridman (04:46.520)

Okay, so we're not, our anecdotal data,

Richard Haier (04:49.880)

I know a guy who's really good at solving

Lex Fridman (04:53.740)

some kind of visual thing, that's not sufficient

Richard Haier (04:58.140)

for us to understand actually the depths

Lex Fridman (05:00.020)

of that person's intelligence.

Lex Fridman (05:01.380)

So how, this idea of G factor,

Lex Fridman (05:07.480)

how much evidence is there, how strong,

Richard Haier (05:11.180)

you know, given across the decades that this idea

Lex Fridman (05:13.700)

has been around, how much has it been held up

Richard Haier (05:16.420)

that there is a universal sort of horsepower

Lex Fridman (05:21.580)

of intelligence that's underneath all of it,

Richard Haier (05:24.020)

all the different tests we do to try to get to this thing

Lex Fridman (05:28.300)

in the depths of the human mind that's a universal,

Richard Haier (05:32.000)

stable measure of a person's intelligence.

Lex Fridman (05:34.820)

You used a couple of words in there, stable and.

Lex Fridman (05:38.820)

We have to be precise with words?

Lex Fridman (05:40.580)

I was hoping we can get away with being poetic.

Richard Haier (05:42.700)

We can, there's a lot about research in general,

Lex Fridman (05:46.260)

not just intelligence research that is poetic.

Richard Haier (05:49.580)

Science has a punetic aspect to it.

Lex Fridman (05:52.900)

Good scientists are very intuitive.

Richard Haier (05:55.840)

They're not just, hey, these are the numbers.

Lex Fridman (05:59.220)

You have to kind of step back and see the big picture.

Richard Haier (06:02.100)

When it comes to intelligence research,

Lex Fridman (06:05.960)

you asked how well has this general concept held up?

Lex Fridman (06:09.660)

And I think I can say without fear

Lex Fridman (06:13.140)

of being empirically contradicted,

Richard Haier (06:16.260)

that it is the most replicated finding in all of psychology.

Lex Fridman (06:21.100)

Now, some cynics may say, well, big deal,

Richard Haier (06:22.980)

psychology, we all know there's a replication crisis

Lex Fridman (06:25.460)

in psychology and a lot of this stuff doesn't replicate.

Richard Haier (06:28.420)

That's all true.

Lex Fridman (06:29.700)

There is no replication crisis when it comes to studying

Richard Haier (06:33.540)

the existence of this general factor.

Lex Fridman (06:36.780)

Let me tell you some things about it.

Richard Haier (06:38.940)

It looks like it's universal

Lex Fridman (06:42.420)

that you find it in all cultures.

Richard Haier (06:44.660)

The way you find it, step back one step,

Lex Fridman (06:47.760)

the way you find it is to give a battery of mental tests.

Lex Fridman (06:51.760)

What battery?

Lex Fridman (06:52.780)

You choose.

Richard Haier (06:53.800)

Take a battery of any mental tests you want,

Lex Fridman (06:57.100)

give it to a large number of diverse people,

Lex Fridman (07:01.520)

and you will be able to extract statistically

Lex Fridman (07:05.740)

the commonality among all those tests.

Richard Haier (07:09.020)

It's done by a technique called factor analysis.

Lex Fridman (07:12.460)

People think that this may be a statistical artifact

Richard Haier (07:17.660)

of some kind, it is not a statistical artifact.

Lex Fridman (07:21.340)

What is factor analysis?

Richard Haier (07:22.700)

Factor analysis is a way of looking at a big set of data

Lex Fridman (07:26.180)

and look at the correlation among the different test scores

Lex Fridman (07:29.900)

and then find empirically the clusters of scores

Lex Fridman (07:33.840)

that go together.

Lex Fridman (07:35.620)

And there are different factors.

Lex Fridman (07:37.340)

So if you have a bunch of mental tests,

Richard Haier (07:39.380)

there may be a verbal factor,

Lex Fridman (07:41.120)

there may be a numerical factor,

Richard Haier (07:43.320)

there may be a visual spatial factor,

Lex Fridman (07:45.900)

but those factors have variance in common with each other.

Lex Fridman (07:50.220)

And that is the common,

Lex Fridman (07:53.060)

that's what's common among all the tests

Lex Fridman (07:55.360)

and that's what gets labeled the G factor.

Lex Fridman (07:58.100)

So if you give a diverse battery of mental tests

Lex Fridman (08:01.460)

and you extract a G factor from it,

Lex Fridman (08:04.740)

that factor usually accounts for around half of the variance.

Richard Haier (08:08.240)

It's the single biggest factor, but it's not the only factor,

Lex Fridman (08:12.860)

but it is the most reliable, it is the most stable,

Lex Fridman (08:17.220)

and it seems to be very much influenced by genetics.

Lex Fridman (08:23.820)

It's very hard to change the G factor with training

Richard Haier (08:28.220)

or drugs or anything else.

Lex Fridman (08:32.860)

You don't know how to increase the G factor.

Richard Haier (08:34.980)

Okay, you said a lot of really interesting things there.

Lex Fridman (08:36.920)

So first, I mean, just to get people used to it

Richard Haier (08:40.940)

in case they're not familiar with this idea,

Lex Fridman (08:43.260)

G factor is what we mean.

Lex Fridman (08:45.820)

So often there's this term used IQ,

Lex Fridman (08:50.340)

which is the way IQ is used,

Richard Haier (08:53.980)

they really mean G factor in regular conversation.

Lex Fridman (08:58.900)

Because what we mean by IQ, we mean intelligence

Lex Fridman (09:02.640)

and what we mean by intelligence,

Lex Fridman (09:04.500)

we mean general intelligence and general intelligence

Richard Haier (09:07.300)

in the human mind from a psychology,

Lex Fridman (09:09.860)

from a serious rigorous scientific perspective

Richard Haier (09:12.320)

actually means G factor.

Lex Fridman (09:13.900)

So G factor equals intelligence,

Richard Haier (09:15.780)

just in this conversation to define terms.

Lex Fridman (09:18.460)

Okay, so there's this stable thing called G factor.

Richard Haier (09:22.220)

You said, now factor, you said factor many times,

Lex Fridman (09:27.220)

means a measure that potentially could be reduced

Richard Haier (09:33.500)

to a single number across the different factors

Lex Fridman (09:35.820)

you mentioned.

Lex Fridman (09:37.180)

And what you said, it accounts for half, halfish.

Lex Fridman (09:45.060)

Accounts for halfish of what?

Richard Haier (09:46.640)

Of variance across the different set of tests.

Lex Fridman (09:51.640)

Set of tests, so if you do for some reason

Lex Fridman (09:56.500)

well on some set of tests, what does that mean?

Lex Fridman (10:00.940)

So that means there's some unique capabilities

Richard Haier (10:03.340)

outside of the G factor that might account for that.

Lex Fridman (10:05.900)

And what are those?

Lex Fridman (10:07.500)

What else is there besides the raw horsepower,

Lex Fridman (10:10.420)

the engine inside your mind that generates intelligence?

Richard Haier (10:13.380)

There are test taking skills.

Lex Fridman (10:16.380)

There are specific abilities.

Richard Haier (10:20.880)

Someone might be particularly good at mathematical things,

Lex Fridman (10:28.300)

mathematical concepts, even simple arithmetic.

Richard Haier (10:32.140)

Some people are much better than others.

Lex Fridman (10:34.340)

You might know people who can memorize,

Lex Fridman (10:36.420)

and short term memory is another component of this.

Lex Fridman (10:42.540)

Short term memory is one of the cognitive processes

Richard Haier (10:46.900)

that's most highly correlated with the G factor.

Lex Fridman (10:51.260)

So all those things like memory,

Richard Haier (10:56.260)

test taking skills account for variability

Lex Fridman (10:59.940)

across the test performances.

Lex Fridman (11:02.240)

But so you can run, but you can't hide

Lex Fridman (11:06.460)

from the thing that God gave you.

Richard Haier (11:08.540)

The genetics, so that G factor,

Lex Fridman (11:12.940)

science says that G factor's there.

Richard Haier (11:15.120)

Each one of us have.

Lex Fridman (11:16.900)

Each one of us has a G factor.

Richard Haier (11:19.380)

Oh boy.

Lex Fridman (11:20.220)

Some have more than others.

Richard Haier (11:21.380)

I'm getting uncomfortable already.

Lex Fridman (11:22.820)

Well, IQ is a score, and IQ, an IQ score

Richard Haier (11:28.500)

is a very good estimate of the G factor.

Lex Fridman (11:32.420)

You can't measure G directly, there's no direct measure.

Richard Haier (11:36.100)

You estimate it from these statistical techniques.

Lex Fridman (11:39.880)

But an IQ score is a good estimate, why?

Richard Haier (11:43.080)

Because a standard IQ test is a battery

Lex Fridman (11:46.420)

of different mental abilities.

Richard Haier (11:48.640)

You combined it into one score,

Lex Fridman (11:51.380)

and that score is highly correlated with the G factor,

Richard Haier (11:55.700)

even if you get better scores on some subtests than others.

Lex Fridman (12:00.140)

Because again, it's what's common

Richard Haier (12:02.300)

to all these mental abilities.

Lex Fridman (12:04.300)

So a good IQ test, and I'll ask you about that,

Lex Fridman (12:08.200)

but a good IQ test tries to compress down that battery

Lex Fridman (12:13.340)

of tests, like tries to get a nice battery,

Richard Haier (12:16.200)

the nice selection of variable tests into one test.

Lex Fridman (12:21.420)

And so in that way, it sneaks up to this G factor.

Lex Fridman (12:24.180)

And that's another interesting thing about G factor.

Lex Fridman (12:28.380)

Now you give, first of all, you have a great book

Richard Haier (12:32.380)

on the neuroscience of intelligence.

Lex Fridman (12:34.180)

You have a great course, which is when I first learned,

Richard Haier (12:38.300)

you're a great teacher, let me just say.

Lex Fridman (12:39.820)

Thank you.

Richard Haier (12:42.140)

Your course at the teaching company,

Lex Fridman (12:44.180)

I hope I'm saying that correctly.

Richard Haier (12:45.820)

The Intelligent Brain.

Lex Fridman (12:47.180)

The Intelligent Brain is when I first heard

Richard Haier (12:50.520)

about this G factor, this mysterious thing

Lex Fridman (12:53.820)

that lurks in the darkness that we cannot quite shine

Richard Haier (12:56.300)

a light on, we're trying to sneak up on.

Lex Fridman (12:59.000)

So the fact that there's this measure,

Richard Haier (13:00.580)

a stable measure of intelligence, we can't measure directly.

Lex Fridman (13:04.820)

But we can come up with a battery test

Richard Haier (13:07.820)

or one test that includes a battery

Lex Fridman (13:10.300)

of variable type of questions that can reliably

Richard Haier (13:17.460)

or attempt to estimate in a stable way that G factor.

Lex Fridman (13:21.700)

That's a fascinating idea.

Lex Fridman (13:23.340)

So for me as an AI person, it's fascinating.

Lex Fridman (13:25.860)

It's fascinating there's something stable like that

Richard Haier (13:27.900)

about the human mind, especially if it's grounded in genetics.

Lex Fridman (13:32.020)

It's both fascinating that as a researcher

Richard Haier (13:37.100)

of the human mind and all the human psychological,

Lex Fridman (13:43.220)

sociological, ethical questions that start arising,

Richard Haier (13:46.560)

it makes me uncomfortable.

Lex Fridman (13:48.260)

But truth can be uncomfortable.

Richard Haier (13:51.340)

I get that a lot about being uncomfortable

Lex Fridman (13:54.180)

talking about this.

Richard Haier (13:56.540)

Let me go back and just say one more empirical thing.

Lex Fridman (14:02.060)

It doesn't matter which battery of tests you use.

Lex Fridman (14:08.700)

So there are countless tests.

Lex Fridman (14:10.700)

You can take any 12 of them at random,

Richard Haier (14:13.500)

extract a G factor and another 12 at random

Lex Fridman (14:17.100)

and extract a G factor and those G factors

Richard Haier (14:19.940)

will be highly correlated like over 0.9 with each other.

Lex Fridman (14:23.160)

That's very, so it is a ubiquitous.

Richard Haier (14:26.100)

It doesn't depend on the content of the test

Lex Fridman (14:28.500)

is what I'm trying to say.

Richard Haier (14:30.260)

It is general among all those tests of mental ability.

Lex Fridman (14:34.020)

And tests of mental, mental abilities include things like,

Richard Haier (14:37.940)

geez, playing poker.

Lex Fridman (14:41.260)

Your skill at poker is not unrelated to G.

Richard Haier (14:46.300)

Your skill at anything that requires reasoning

Lex Fridman (14:49.460)

and thinking, anything, spelling, arithmetic,

Richard Haier (14:54.300)

more complex things, this concept is ubiquitous.

Lex Fridman (15:00.500)

And when you do batteries of tests in different cultures,

Richard Haier (15:03.880)

you get the same thing.

Lex Fridman (15:05.800)

So this says something interesting about the human mind

Richard Haier (15:08.880)

that as a computer is designed to be general.

Lex Fridman (15:12.780)

So that means you can, so it's not easily made specialized.

Richard Haier (15:17.780)

Meaning if you're going to be good at one thing,

Lex Fridman (15:21.940)

Miyamoto Musashi has this quote, he's an ancient warrior,

Richard Haier (15:26.940)

famous for the Book of Five Rings in the martial arts world.

Lex Fridman (15:30.940)

And the quote goes, if you know the way broadly,

Richard Haier (15:34.180)

you will see it in everything.

Lex Fridman (15:36.420)

Meaning if you do one thing is going to generalize

Richard Haier (15:42.660)

to everything.

Lex Fridman (15:44.500)

And that's an interesting quote.

Lex Fridman (15:46.620)

And that's an interesting thing about the human mind.

Lex Fridman (15:50.660)

So that's what the G factor reveals.

Richard Haier (15:54.380)

Okay, so what's the difference,

Lex Fridman (15:57.060)

if you can elaborate a little bit further

Lex Fridman (15:58.900)

between IQ and G factor?

Lex Fridman (16:00.900)

Just because it's a source of confusion for people.

Lex Fridman (16:03.560)

And IQ is a score.

Lex Fridman (16:05.660)

People use the word IQ to mean intelligence.

Lex Fridman (16:08.260)

But IQ has a more technical meaning

Lex Fridman (16:11.020)

for people who work in the field.

Lex Fridman (16:12.740)

And it's an IQ score, a score on a test

Lex Fridman (16:16.540)

that estimates the G factor.

Lex Fridman (16:20.060)

And the G factor is what's common

Lex Fridman (16:22.020)

among all these tests of mental ability.

Lex Fridman (16:24.200)

So if you think about, it's not a Venn diagram,

Lex Fridman (16:27.040)

but I guess you could make a Venn diagram out of it,

Lex Fridman (16:30.580)

but the G factor would be really at the core,

Lex Fridman (16:33.900)

what's common to everything.

Lex Fridman (16:37.540)

And what IQ scores do is they allow a rank order

Lex Fridman (16:42.540)

of people on the score.

Lex Fridman (16:44.500)

And this is what makes people uncomfortable.

Lex Fridman (16:46.900)

This is where there's a lot of controversy

Richard Haier (16:48.920)

about whether IQ tests are biased

Lex Fridman (16:51.540)

toward any one group or another.

Lex Fridman (16:54.380)

And a lot of the answers to these questions are very clear,

Lex Fridman (16:59.060)

but they also have a technical aspect of it

Richard Haier (17:02.060)

that's not so easy to explain.

Lex Fridman (17:04.180)

Well, we'll talk about the fascinating

Lex Fridman (17:06.180)

and the difficult things about all of this.

Lex Fridman (17:10.700)

So by the way, when you say rank order,

Richard Haier (17:12.560)

that means you get a number and that means one person,

Lex Fridman (17:15.420)

you can now compare.

Richard Haier (17:17.540)

Like you could say that this other person

Lex Fridman (17:20.900)

is more intelligent than me.

Richard Haier (17:23.020)

Well, what you can say is IQ scores

Lex Fridman (17:25.820)

are interpreted really as percentiles.

Lex Fridman (17:29.220)

So that if you have an IQ of 140

Lex Fridman (17:33.300)

and somebody else has 70,

Richard Haier (17:35.580)

the metric is such that you cannot say

Lex Fridman (17:38.140)

the person with an IQ of 140 is twice as smart

Richard Haier (17:42.940)

as a person with an IQ of 70.

Lex Fridman (17:45.980)

That would require a ratio scale with an absolute zero.

Richard Haier (17:49.900)

Now you may think you know people with zero intelligence,

Lex Fridman (17:53.140)

but in fact, there is no absolute zero on an IQ scale.

Richard Haier (17:57.980)

It's relative to other people.

Lex Fridman (18:01.140)

So relative to other people,

Richard Haier (18:03.220)

somebody with an IQ score of 140

Lex Fridman (18:06.300)

is in the upper less than 1%,

Richard Haier (18:09.260)

whereas somebody with an IQ of 70

Lex Fridman (18:12.580)

is two standard deviations below the mean.

Richard Haier (18:15.500)

That's a different percentile.

Lex Fridman (18:18.660)

So it's similar to like in chess,

Richard Haier (18:20.920)

you have an ELO rating that's designed to rank order people.

Lex Fridman (18:27.740)

So you can't say it's twice one person.

Richard Haier (18:30.500)

If your ELO rating is twice another person,

Lex Fridman (18:33.500)

I don't think you're twice as good at chess.

Richard Haier (18:35.740)

It's not stable in that way,

Lex Fridman (18:37.580)

but because it's very difficult

Richard Haier (18:39.060)

to do these kinds of comparisons.

Lex Fridman (18:41.440)

But so what can we say about the number itself?

Lex Fridman (18:47.020)

Is that stable across tests and so on, or no?

Lex Fridman (18:50.500)

There are a number of statistical properties of any test.

Richard Haier (18:54.060)

They're called psychometric properties.

Lex Fridman (18:56.460)

You have validity, you have reliability,

Richard Haier (18:59.340)

reliability, there are many different kinds of reliability.

Lex Fridman (19:02.780)

They all essentially measure stability.

Lex Fridman (19:05.980)

And IQ tests are stable within an individual.

Lex Fridman (19:09.780)

There are some longitudinal studies

Richard Haier (19:11.940)

where children were measured at age 11.

Lex Fridman (19:15.840)

And again, when they were 70 years old

Lex Fridman (19:18.140)

and the two IQ scores are highly correlated with each other.

Lex Fridman (19:22.140)

This comes from a fascinating study from Scotland.

Richard Haier (19:26.260)

In the 1930s, some researchers decided to get an IQ test

Lex Fridman (19:31.260)

on every single child age 11 in the whole country.

Lex Fridman (19:35.900)

And they did.

Lex Fridman (19:37.340)

And those records were discovered in an old storeroom

Richard Haier (19:42.980)

at the University of Edinburgh by a friend of mine,

Lex Fridman (19:47.020)

Ian Deary, who found the records, digitized them,

Lex Fridman (19:52.180)

and has done a lot of research

Lex Fridman (19:53.940)

on the people who are still alive today

Richard Haier (19:57.220)

from that original study,

Lex Fridman (19:58.540)

including brain imaging research, by the way.

Richard Haier (1:00:03.000)

The film was called Charlie

Lex Fridman (1:00:04.800)

for the younger people who are listening to this.

Richard Haier (1:00:08.680)

You might be able to stream it on Netflix or something,

Lex Fridman (1:00:11.780)

but it was a story about a person

Richard Haier (1:00:16.780)

with very low IQ who underwent a surgical procedure

Lex Fridman (1:00:20.620)

in the brain and he slowly became a genius.

Lex Fridman (1:00:24.860)

And the tragedy of the story is the effect was temporary.

Lex Fridman (1:00:31.180)

It's a fascinating story really.

Richard Haier (1:00:33.120)

That goes in contrast to the basic human experience

Lex Fridman (1:00:36.580)

that each of us individually have,

Lex Fridman (1:00:38.420)

but it raises the question of the full range of people

Lex Fridman (1:00:43.420)

you might be able to be given different levels

Richard Haier (1:00:47.700)

of intelligence.

Lex Fridman (1:00:48.780)

You've mentioned the normal distribution.

Lex Fridman (1:00:52.920)

So let's talk about it.

Lex Fridman (1:00:54.520)

There's a book called The Bell Curve written in 1994,

Richard Haier (1:00:58.360)

written by psychologist Richard Herrnstein

Lex Fridman (1:01:01.340)

and political scientist Charles Murray.

Lex Fridman (1:01:04.500)

Why was this book so controversial?

Lex Fridman (1:01:08.540)

This is a fascinating book.

Richard Haier (1:01:10.720)

I know Charles Murray.

Lex Fridman (1:01:12.620)

I've had many conversations with him.

Lex Fridman (1:01:15.340)

Yeah, what is the book about?

Lex Fridman (1:01:16.660)

The book is about the importance of intelligence

Richard Haier (1:01:22.720)

in everyday life.

Lex Fridman (1:01:25.420)

That's what the book is about.

Richard Haier (1:01:27.520)

It's an empirical book.

Lex Fridman (1:01:29.220)

It has statistical analyses of very large databases

Richard Haier (1:01:34.180)

that show that essentially IQ scores or their equivalent

Lex Fridman (1:01:39.180)

are correlated to all kinds of social problems

Lex Fridman (1:01:44.640)

and social benefits.

Lex Fridman (1:01:46.800)

And that in itself is not where the controversy

Richard Haier (1:01:51.740)

about that book came.

Lex Fridman (1:01:53.640)

The controversy was about one chapter in that book.

Lex Fridman (1:01:57.600)

And that is a chapter about the average difference

Lex Fridman (1:02:02.240)

in mean scores between black Americans and white Americans.

Lex Fridman (1:02:06.720)

And these are the terms that were used in the book

Lex Fridman (1:02:08.960)

at the time and are still used to some extent.

Lex Fridman (1:02:14.080)

And historically, or really for decades,

Lex Fridman (1:02:21.240)

it has been observed that disadvantaged groups

Richard Haier (1:02:28.020)

score on average lower than Caucasians

Lex Fridman (1:02:35.480)

on academic tests, tests of mental ability,

Lex Fridman (1:02:38.280)

and especially on IQ tests.

Lex Fridman (1:02:40.440)

And the difference is about a standard deviation,

Richard Haier (1:02:43.120)

which is about 15 points, which is a substantial difference.

Lex Fridman (1:02:48.920)

In the book, Herrnstein and Murray in this one chapter

Richard Haier (1:02:54.160)

assert clearly and unambiguously

Lex Fridman (1:02:59.200)

that whether this average difference

Richard Haier (1:03:01.960)

is due to genetics or not, they are agnostic.

Lex Fridman (1:03:06.960)

They don't know.

Richard Haier (1:03:08.280)

Moreover, they assert they don't care

Lex Fridman (1:03:11.160)

because you wouldn't treat anybody differently

Richard Haier (1:03:13.880)

knowing if there was a genetic component or not

Lex Fridman (1:03:17.640)

because that's a group average finding.

Richard Haier (1:03:20.560)

Every individual has to be treated as an individual.

Lex Fridman (1:03:23.960)

You can't make any assumption

Richard Haier (1:03:26.040)

about what that person's intellectual ability might be

Lex Fridman (1:03:30.640)

from the fact of a average group difference.

Richard Haier (1:03:33.160)

They're very clear about this.

Lex Fridman (1:03:34.880)

Nonetheless, people took away,

Richard Haier (1:03:41.680)

I'm gonna choose my words carefully

Lex Fridman (1:03:43.240)

because I have a feeling that many critics

Richard Haier (1:03:44.800)

didn't actually read the book.

Lex Fridman (1:03:49.160)

They took away that Herrnstein and Murray were saying

Richard Haier (1:03:51.880)

that blacks are genetically inferior.

Lex Fridman (1:03:54.960)

That was the take home message.

Lex Fridman (1:03:56.560)

And if they weren't saying it, they were implying it

Lex Fridman (1:03:59.680)

because they had a chapter that discussed

Richard Haier (1:04:02.720)

this empirical observation of a difference.

Lex Fridman (1:04:07.000)

And isn't this horrible?

Lex Fridman (1:04:10.560)

And so the reaction to that book was incendiary.

Lex Fridman (1:04:18.400)

What do we know about from that book

Lex Fridman (1:04:22.560)

and the research beyond about race differences

Lex Fridman (1:04:28.960)

and intelligence?

Richard Haier (1:04:30.620)

It's still the most incendiary topic in psychology.

Lex Fridman (1:04:33.800)

Nothing has changed that.

Richard Haier (1:04:35.880)

Anybody who even discusses it is easily called a racist

Lex Fridman (1:04:41.200)

just for discussing it.

Richard Haier (1:04:43.040)

It's become fashionable to find racism

Lex Fridman (1:04:45.560)

in any discussion like this.

Richard Haier (1:04:49.960)

It's unfortunate.

Lex Fridman (1:04:53.560)

The short answer to your question is

Richard Haier (1:04:57.080)

there's been very little actual research

Lex Fridman (1:05:00.160)

on this topic since 19...

Richard Haier (1:05:03.840)

Since the bell curve.

Lex Fridman (1:05:05.200)

Since the bell curve, even before.

Richard Haier (1:05:07.920)

This really became incendiary in 1969

Lex Fridman (1:05:12.460)

with an article published by an educational psychologist

Richard Haier (1:05:15.760)

named Arthur Jensen.

Lex Fridman (1:05:17.760)

Let's just take a minute and go back to that

Richard Haier (1:05:20.520)

to see the bell curve in a little bit more

Lex Fridman (1:05:22.920)

historical perspective.

Richard Haier (1:05:25.040)

Arthur Jensen was a educational psychologist

Lex Fridman (1:05:28.160)

at UC Berkeley.

Richard Haier (1:05:29.320)

I knew him as well.

Lex Fridman (1:05:31.440)

And in 1969 or 68, the Harvard Educational Review

Richard Haier (1:05:37.800)

asked him to do a review article

Lex Fridman (1:05:42.640)

on the early childhood education programs

Richard Haier (1:05:47.560)

that were designed to raise the IQs of minority students.

Lex Fridman (1:05:54.800)

This was before the federally funded Head Start program.

Richard Haier (1:05:58.560)

Head Start had not really gotten underway

Lex Fridman (1:06:01.280)

at the time Jensen undertook his review

Richard Haier (1:06:04.440)

of what were a number of demonstration programs.

Lex Fridman (1:06:08.640)

And these demonstration programs were for young children

Richard Haier (1:06:13.920)

who were around kindergarten age.

Lex Fridman (1:06:15.720)

And they were specially designed to be

Richard Haier (1:06:18.200)

cognitively stimulating, to provide lunches,

Lex Fridman (1:06:23.020)

do all the things that people thought would

Richard Haier (1:06:27.200)

minimize this average gap of intelligence tests.

Lex Fridman (1:06:31.800)

There was a strong belief among virtually all psychologists

Richard Haier (1:06:37.200)

that the cause of the gap was unequal opportunity

Lex Fridman (1:06:40.760)

due to racism, due to all negative things in the society.

Lex Fridman (1:06:45.600)

And if you could compensate for this, the gap would go away.

Lex Fridman (1:06:51.200)

So early childhood education back then was called

Richard Haier (1:06:54.740)

literally compensatory education.

Lex Fridman (1:06:58.640)

Jensen looked at these programs.

Richard Haier (1:07:00.920)

He was an empirical guy.

Lex Fridman (1:07:02.200)

He understood psychometrics.

Lex Fridman (1:07:04.600)

And he wrote a, it was over a hundred page article

Lex Fridman (1:07:08.800)

detailing these programs

Lex Fridman (1:07:12.120)

and the flaws in their research design.

Lex Fridman (1:07:15.360)

Some of the programs reported IQ gains

Richard Haier (1:07:17.640)

of on average five points,

Lex Fridman (1:07:20.400)

but a few reported 10, 20 and even 30 point gains.

Richard Haier (1:07:24.720)

One was called the miracle in Milwaukee.

Lex Fridman (1:07:28.000)

That investigator went to jail ultimately

Richard Haier (1:07:30.440)

for fabricating data.

Lex Fridman (1:07:33.280)

But the point is that Jensen wrote an article that said,

Richard Haier (1:07:36.000)

look, the opening sentence of his article is classic.

Lex Fridman (1:07:40.000)

The opening sentence is, I may not quote it exactly right,

Lex Fridman (1:07:43.520)

but it's, we have tried compensatory education

Lex Fridman (1:07:47.120)

and it has failed.

Lex Fridman (1:07:48.420)

And he showed that these gains were essentially nothing.

Lex Fridman (1:07:54.140)

You couldn't really document empirically any gains at all

Richard Haier (1:07:58.660)

from these really earnest efforts to increase IQ.

Lex Fridman (1:08:03.340)

But he went a step further, a fateful step further.

Richard Haier (1:08:08.060)

He said, not only have these efforts failed,

Lex Fridman (1:08:11.740)

but because they have had essentially no impact,

Richard Haier (1:08:15.220)

we have to reexamine our assumption

Lex Fridman (1:08:17.660)

that these differences are caused by environmental things

Richard Haier (1:08:22.100)

that we can address with education.

Lex Fridman (1:08:24.420)

We need to consider a genetic influence,

Richard Haier (1:08:28.500)

whether there's a genetic influence

Lex Fridman (1:08:30.580)

on this group difference.

Lex Fridman (1:08:32.420)

So you said that this is one of the more controversial works

Lex Fridman (1:08:36.060)

ever in science. I think it's the most infamous paper

Richard Haier (1:08:37.900)

in all of psychology, I would go on to say.

Lex Fridman (1:08:41.680)

Because in 1969, the genetic data was very skimpy

Richard Haier (1:08:46.680)

on this question, skimpy and controversial.

Lex Fridman (1:08:49.640)

It's always been controversial,

Lex Fridman (1:08:50.820)

but it was even skimpy and controversial.

Lex Fridman (1:08:53.820)

It's kind of a long story that I go into a little bit

Richard Haier (1:08:56.480)

in more detail in the book, Neuroscience of Intelligence.

Lex Fridman (1:09:02.200)

But to say he was vilified is an understatement.

Richard Haier (1:09:06.320)

I mean, he couldn't talk at the American

Lex Fridman (1:09:08.880)

Psychological Association without bomb threats

Richard Haier (1:09:13.140)

clearing the lecture hall.

Lex Fridman (1:09:15.240)

Campus security watched him all the time.

Richard Haier (1:09:18.120)

They opened his mail.

Lex Fridman (1:09:20.320)

He had to retreat to a different address.

Richard Haier (1:09:24.680)

This was one of the earliest kinds,

Lex Fridman (1:09:28.760)

this is before the internet

Lex Fridman (1:09:30.500)

and kind of internet social media mobs.

Lex Fridman (1:09:35.080)

But it was that intense.

Lex Fridman (1:09:38.120)

And I have written that overnight,

Lex Fridman (1:09:42.740)

after the publication of this article,

Richard Haier (1:09:45.960)

all intelligence research became radioactive.

Lex Fridman (1:09:49.760)

Nobody wanted to talk about it.

Richard Haier (1:09:56.200)

Nobody was doing more research.

Lex Fridman (1:09:58.920)

And then the bell curve came along.

Lex Fridman (1:10:02.080)

And the Jensen controversy was dying down.

Lex Fridman (1:10:05.720)

I have stories that Jensen told me about his interaction

Richard Haier (1:10:08.960)

with the Nixon White House on this issue.

Lex Fridman (1:10:10.920)

I mean, this was like a really big deal.

Richard Haier (1:10:14.360)

It was some unbelievable stories,

Lex Fridman (1:10:16.160)

but he told me this, so I kind of believe these stories.

Richard Haier (1:10:20.100)

Nonetheless.

Lex Fridman (1:10:21.640)

25 years later.

Richard Haier (1:10:22.920)

25 years later.

Lex Fridman (1:10:24.240)

All this silence, basically, saying,

Richard Haier (1:10:30.040)

nobody wants to do this kind of research.

Lex Fridman (1:10:32.280)

There's so much pressure, so much attack

Richard Haier (1:10:34.620)

against this kind of research.

Lex Fridman (1:10:36.360)

And here's sort of a bold, stupid, crazy people

Richard Haier (1:10:41.960)

that decide to dive right back in.

Lex Fridman (1:10:44.880)

I wonder how much discussion that was.

Lex Fridman (1:10:46.520)

Do we include this chapter or not?

Lex Fridman (1:10:48.720)

Murray has said they discussed it,

Lex Fridman (1:10:51.100)

and they felt they should include it.

Lex Fridman (1:10:55.640)

And they were very careful in the way they wrote it,

Richard Haier (1:10:59.580)

which did them no good.

Lex Fridman (1:11:01.040)

So, as a matter of fact, when the bell curve came out,

Richard Haier (1:11:06.560)

it was so controversial.

Lex Fridman (1:11:08.480)

I got a call from a television show called Nightline.

Richard Haier (1:11:13.320)

It was with a broadcaster called Ted Koppel.

Lex Fridman (1:11:16.840)

We had this evening show, I think it was on late at night.

Richard Haier (1:11:20.600)

Talked about news.

Lex Fridman (1:11:21.600)

It was a straight up news thing.

Lex Fridman (1:11:24.240)

And a producer called and asked if I would be on it

Lex Fridman (1:11:28.200)

to talk about the bell curve.

Lex Fridman (1:11:31.680)

And I said, she asked me what I thought

Lex Fridman (1:11:35.000)

about the bell curve as a book.

Lex Fridman (1:11:36.640)

And I said, look, it's a very good book.

Lex Fridman (1:11:38.560)

It talks about the role of intelligence in society.

Lex Fridman (1:11:43.080)

And she said, no, no, what do you think

Lex Fridman (1:11:44.980)

about the chapter on race?

Richard Haier (1:11:47.160)

That's what we want you to talk about.

Lex Fridman (1:11:49.940)

I remember this conversation.

Richard Haier (1:11:52.400)

I said, well, she said, what would you say

Lex Fridman (1:11:56.600)

if you were on TV?

Lex Fridman (1:11:58.800)

And I said, well, what I would say is that

Lex Fridman (1:12:02.240)

it's not at all clear if there's any genetic component

Richard Haier (1:12:07.760)

to intelligence, any differences.

Lex Fridman (1:12:13.600)

But if there were a strong genetic component,

Richard Haier (1:12:17.440)

that would be a good thing.

Lex Fridman (1:12:21.140)

And complete silence on the other end of the phone.

Lex Fridman (1:12:25.280)

And she said, well, what do you mean?

Lex Fridman (1:12:28.180)

And I said, well, if it's the more genetic

Richard Haier (1:12:31.160)

any difference is, the more it's biological.

Lex Fridman (1:12:35.080)

And if it's biological, we can figure out how to fix it.

Richard Haier (1:12:39.960)

I see, that's interesting.

Lex Fridman (1:12:41.480)

She said, would you say that on television?

Richard Haier (1:12:43.680)

Yes.

Lex Fridman (1:12:44.520)

And I said, no.

Lex Fridman (1:12:45.340)

And so that was the end of that.

Lex Fridman (1:12:47.400)

So that's for more like biology is within the reach

Richard Haier (1:12:56.400)

of science and the environment is a public policy,

Lex Fridman (1:13:02.000)

is social and all those kinds of things.

Richard Haier (1:13:05.500)

From your perspective, whichever one you think

Lex Fridman (1:13:09.120)

is more amenable to solutions in the short term

Richard Haier (1:13:11.760)

is the one that excites you.

Lex Fridman (1:13:13.560)

But you saying that is good, the truth of genetic differences,

Richard Haier (1:13:22.900)

no matter what, between groups is a painful, harmful,

Lex Fridman (1:13:32.680)

potentially dangerous thing.

Lex Fridman (1:13:35.320)

So let me ask you to this question,

Lex Fridman (1:13:38.080)

whether it's bell curve or any research

Richard Haier (1:13:40.160)

on race differences, can that be used to increase

Lex Fridman (1:13:47.080)

the amount of racism in the world?

Richard Haier (1:13:49.400)

Can that be used to increase the amount of hate

Lex Fridman (1:13:51.680)

in the world?

Lex Fridman (1:13:52.820)

Do you think about this kind of stuff?

Lex Fridman (1:13:54.400)

I've thought about this a lot, not as a scientist,

Lex Fridman (1:13:57.400)

but as a person.

Lex Fridman (1:14:00.680)

And my sense is there is such enormous reservoirs

Richard Haier (1:14:05.680)

of hate and racism that have nothing to do

Lex Fridman (1:14:13.040)

with scientific knowledge of the data,

Richard Haier (1:14:16.600)

that speak against that.

Lex Fridman (1:14:19.200)

That no, I don't wanna give racist groups a veto power

Richard Haier (1:14:25.360)

over what scientists study.

Lex Fridman (1:14:27.680)

If you think that the differences, and by the way,

Richard Haier (1:14:31.440)

virtually no one disagrees that there are differences

Lex Fridman (1:14:34.360)

in scores, it's all about what causes them

Lex Fridman (1:14:37.200)

and how to fix it.

Lex Fridman (1:14:39.400)

So if you think this is a cultural problem,

Richard Haier (1:14:42.920)

then you must ask the problem,

Lex Fridman (1:14:44.840)

what do you want to change anything about the culture?

Lex Fridman (1:14:49.360)

Or are you okay with the culture?

Lex Fridman (1:14:51.800)

Cause you don't feel it's appropriate

Richard Haier (1:14:53.380)

to change a person's culture.

Lex Fridman (1:14:55.600)

So are you okay with that?

Lex Fridman (1:14:57.160)

And the fact that that may lead to disadvantages

Lex Fridman (1:14:59.720)

in school achievement.

Richard Haier (1:15:01.600)

It's a question.

Lex Fridman (1:15:02.500)

If you think it's environmental,

Lex Fridman (1:15:05.940)

what are the environmental parameters that can be fixed?

Lex Fridman (1:15:10.380)

I'll tell you one, lead from gasoline in the atmosphere.

Richard Haier (1:15:15.860)

Lead in paint, lead in water.

Lex Fridman (1:15:18.780)

That's an environmental toxin that society

Richard Haier (1:15:22.340)

has the means to eliminate, and they should.

Lex Fridman (1:15:25.980)

Yeah, just to sort of try and define some insights

Lex Fridman (1:15:30.260)

and conclusion to this very difficult topic.

Lex Fridman (1:15:33.860)

Is there been research on environment versus genetics,

Richard Haier (1:15:38.740)

nature versus nurture, on this question

Lex Fridman (1:15:40.420)

of race differences?

Richard Haier (1:15:43.100)

There is not, no one wants to do this research.

Lex Fridman (1:15:46.380)

First of all, it's hard research to do.

Richard Haier (1:15:48.080)

Second of all, it's a minefield.

Lex Fridman (1:15:50.620)

No one wants to spend their career on it.

Richard Haier (1:15:52.940)

Tenured people don't want to do it, let alone students.

Lex Fridman (1:15:56.320)

The way I talk about it,

Richard Haier (1:16:00.600)

well, before I tell you the way I talk about it,

Lex Fridman (1:16:02.440)

I want to say one more thing about Jensen.

Richard Haier (1:16:05.640)

He was once asked by a journalist straight out,

Lex Fridman (1:16:08.120)

are you a racist?

Richard Haier (1:16:10.480)

His answer was very interesting.

Lex Fridman (1:16:12.520)

His answer was, I've thought about that a lot,

Lex Fridman (1:16:16.960)

and I've concluded it doesn't matter.

Lex Fridman (1:16:22.040)

Now, I know what he meant by this.

Richard Haier (1:16:23.640)

The guts to say that, wow.

Lex Fridman (1:16:25.840)

He was a very unusual person.

Richard Haier (1:16:27.560)

I think he had a touch of Asperger's syndrome,

Lex Fridman (1:16:29.920)

to tell you the truth,

Richard Haier (1:16:30.940)

because I saw him in many circumstances.

Lex Fridman (1:16:34.160)

He would be canceled on Twitter immediately

Richard Haier (1:16:36.140)

with that sentence.

Lex Fridman (1:16:37.520)

But what he meant was he had a hypothesis,

Lex Fridman (1:16:42.440)

and with respect to group differences,

Lex Fridman (1:16:44.920)

he called it the default hypothesis.

Richard Haier (1:16:47.400)

He said, whatever factors affect individual intelligence

Lex Fridman (1:16:51.240)

are likely the same factors that affect group differences.

Richard Haier (1:16:54.320)

It was the default.

Lex Fridman (1:16:55.600)

But it was a hypothesis.

Richard Haier (1:16:58.280)

It should be tested, and if it turned out

Lex Fridman (1:17:01.040)

empirical tests didn't support the hypothesis,

Richard Haier (1:17:03.680)

he was happy to move on to something else.

Lex Fridman (1:17:06.480)

He was absolutely committed to that scientific ideal,

Richard Haier (1:17:12.480)

that it's an empirical question,

Lex Fridman (1:17:16.040)

we should look at it, and let's see what happens.

Richard Haier (1:17:18.800)

The scientific method cannot be racist,

Lex Fridman (1:17:22.080)

from his perspective.

Richard Haier (1:17:23.480)

It doesn't matter what the scientists,

Lex Fridman (1:17:26.240)

if they follow the scientific method,

Richard Haier (1:17:30.440)

it doesn't matter what they believe.

Lex Fridman (1:17:32.080)

And if they are biased, and they consciously

Richard Haier (1:17:35.520)

or unconsciously bias the data,

Lex Fridman (1:17:39.520)

other people will come along to replicate it,

Richard Haier (1:17:42.680)

they will fail, and the process over time will work.

Lex Fridman (1:17:48.200)

So let me push back on this idea.

Richard Haier (1:17:50.980)

Because psychology to me is full of gray areas.

Lex Fridman (1:17:57.380)

And what I've observed about psychology,

Richard Haier (1:18:01.460)

even replication crisis aside,

Lex Fridman (1:18:04.060)

is that something about the media,

Richard Haier (1:18:06.220)

something about journalism,

Lex Fridman (1:18:08.020)

something about the virality of ideas in the public sphere,

Richard Haier (1:18:13.700)

they misinterpret, they take up things from studies,

Lex Fridman (1:18:18.360)

willfully or from ignorance, misinterpret findings,

Lex Fridman (1:18:23.640)

and tell narratives around that.

Lex Fridman (1:18:27.120)

I personally believe, for me,

Richard Haier (1:18:29.820)

I'm not saying that broadly about science,

Lex Fridman (1:18:31.520)

but for me, it's my responsibility to anticipate

Richard Haier (1:18:35.760)

the ways in which findings will be misinterpreted.

Lex Fridman (1:18:40.420)

So I thought about this a lot,

Richard Haier (1:18:42.840)

because I published papers on semi autonomous vehicles,

Lex Fridman (1:18:47.020)

and those cars, people die in cars.

Richard Haier (1:18:52.960)

There's people that have written me letters saying emails,

Lex Fridman (1:18:57.200)

nobody writes letters, I wish they did,

Richard Haier (1:19:00.000)

that have blood on my hands,

Lex Fridman (1:19:01.800)

because of things that I would say positive or negative,

Richard Haier (1:19:04.800)

there's consequences.

Lex Fridman (1:19:06.200)

In the same way, when you're a researcher of intelligence,

Richard Haier (1:19:09.260)

I'm sure you might get emails,

Lex Fridman (1:19:12.040)

or at least people might believe

Richard Haier (1:19:14.500)

that a finding of your study is going to be used

Lex Fridman (1:19:17.920)

by a large number of people

Richard Haier (1:19:19.080)

to increase the amount of hate in the world.

Lex Fridman (1:19:22.640)

I think there's some responsibility on scientists,

Lex Fridman (1:19:26.060)

but for me, I think there's a great responsibility

Lex Fridman (1:19:30.480)

to anticipate the ways things will be misinterpreted,

Lex Fridman (1:19:35.400)

and there, you have to, first of all,

Lex Fridman (1:19:37.880)

decide whether you want to say a thing at all,

Richard Haier (1:19:40.980)

do the study at all, publish the study at all,

Lex Fridman (1:19:43.440)

and two, the words with which you explain it.

Richard Haier (1:19:49.000)

I find this on Twitter a lot, actually,

Lex Fridman (1:19:50.640)

which is, when I write a tweet,

Lex Fridman (1:19:53.200)

and I'm usually just doing it so innocently,

Lex Fridman (1:19:58.560)

I'll write it, it takes me five seconds to write it,

Richard Haier (1:20:02.040)

or whatever, 30 seconds to write it,

Lex Fridman (1:20:04.320)

and then I'll think, all right, I close my eyes open,

Lex Fridman (1:20:08.240)

and try to see how will the world interpret this,

Lex Fridman (1:20:11.680)

what are the ways in which this will be misinterpreted,

Lex Fridman (1:20:14.400)

and I'll sometimes adjust that tweet to see,

Lex Fridman (1:20:18.480)

yeah, so in my mind, it's clear,

Lex Fridman (1:20:20.940)

but that's because it's my mind from which this tweet came,

Lex Fridman (1:20:24.040)

but you have to think, in a fresh mind that sees this,

Lex Fridman (1:20:28.160)

and it's spread across a large number of other minds,

Lex Fridman (1:20:32.840)

how will the interpretation morph?

Richard Haier (1:20:36.040)

I mean, for a tweet, it's a silly thing, it doesn't matter,

Lex Fridman (1:20:38.260)

but for a scientific paper and study and finding,

Richard Haier (1:20:45.640)

I think it matters.

Lex Fridman (1:20:47.240)

So I don't know what your thoughts are on that,

Richard Haier (1:20:49.720)

because maybe for Jensen, the data's there,

Lex Fridman (1:20:54.600)

what do you want me to do?

Richard Haier (1:20:55.600)

This is a scientific process that's been carried out,

Lex Fridman (1:20:59.160)

if you think the data was polluted by bias,

Richard Haier (1:21:02.200)

do other studies that reveal the bias,

Lex Fridman (1:21:05.800)

but the data's there.

Lex Fridman (1:21:07.440)

And I'm not a poet, I'm not a literary writer,

Lex Fridman (1:21:14.720)

what do you want me to do?

Richard Haier (1:21:15.640)

I'm just presenting you the data.

Lex Fridman (1:21:17.560)

What do you think on that spectrum?

Lex Fridman (1:21:19.360)

What's the role of a scientist?

Lex Fridman (1:21:21.120)

The reason I do podcasts,

Richard Haier (1:21:23.540)

the reason I write books for the public

Lex Fridman (1:21:27.280)

is to explain what I think the data mean

Lex Fridman (1:21:30.420)

and what I think the data don't mean.

Lex Fridman (1:21:32.680)

I don't do very much on Twitter other than to retweet

Richard Haier (1:21:38.240)

references to papers.

Lex Fridman (1:21:39.920)

I don't think it's my role to explain these,

Richard Haier (1:21:42.440)

because they're complicated, they're nuanced.

Lex Fridman (1:21:46.480)

But when you decide not to do a scientific study

Richard Haier (1:21:51.560)

because you're, or not to publish a result

Lex Fridman (1:21:54.360)

because you're afraid the result could be harmful

Richard Haier (1:21:59.360)

or insensitive, that's not an unreasonable thought.

Lex Fridman (1:22:05.080)

And people will make different conclusions

Lex Fridman (1:22:09.200)

and decisions about that.

Lex Fridman (1:22:11.560)

I wrote about this, I'm the editor

Richard Haier (1:22:14.680)

of a journal called Intelligence,

Lex Fridman (1:22:17.000)

which publishes scientific papers.

Richard Haier (1:22:20.560)

Sometimes we publish papers on group differences.

Lex Fridman (1:22:24.200)

Those papers sometimes are controversial.

Richard Haier (1:22:27.040)

These papers are written for a scientific audience.

Lex Fridman (1:22:29.600)

They're not written for the Twitter audience.

Richard Haier (1:22:32.160)

I don't promote them very much on Twitter.

Lex Fridman (1:22:37.480)

But in a scientific paper,

Richard Haier (1:22:41.240)

you have to now choose your words carefully also,

Lex Fridman (1:22:44.600)

because those papers are picked up by non scientists,

Richard Haier (1:22:49.480)

by writers of various kinds,

Lex Fridman (1:22:52.000)

and you have to be available to discuss what you're saying

Lex Fridman (1:22:56.120)

and what you're not saying.

Lex Fridman (1:22:58.720)

Sometimes you are successful at having a good conversation

Richard Haier (1:23:04.640)

like we are today, that doesn't start out pejorative.

Lex Fridman (1:23:09.200)

Other times I've been asked to participate in debates

Richard Haier (1:23:12.520)

where my role would be to justify race science.

Lex Fridman (1:23:16.500)

Well, you can see you start out.

Richard Haier (1:23:21.280)

That was a BBC request that I received.

Lex Fridman (1:23:25.680)

I have so much, it's a love hate relationship,

Richard Haier (1:23:28.280)

mostly hate with these shallow journalism organizations.

Lex Fridman (1:23:33.000)

So they would want to use you

Richard Haier (1:23:36.240)

as a kind of in a debate setting to communicate

Lex Fridman (1:23:39.180)

as to like there is raised differences between groups

Lex Fridman (1:23:42.300)

and make that into debate and put you in a role of...

Lex Fridman (1:23:47.880)

Justifying racism.

Richard Haier (1:23:49.280)

Justifying racism.

Lex Fridman (1:23:50.120)

That's what they're asking me to do.

Richard Haier (1:23:51.120)

Courses like educating about this field

Lex Fridman (1:23:54.360)

of the science of intelligence, yeah.

Richard Haier (1:23:56.040)

I wanna say one more thing

Lex Fridman (1:23:57.980)

before we get off the normal distribution.

Lex Fridman (1:24:01.320)

You also asked me what is the science after the bell curve?

Lex Fridman (1:24:06.160)

And the short answer is there's not much new work,

Lex Fridman (1:24:09.640)

but whatever work there is supports the idea

Lex Fridman (1:24:13.420)

that there still are group differences.

Richard Haier (1:24:16.280)

It's arguable whether those differences

Lex Fridman (1:24:18.560)

have diminished at all or not.

Lex Fridman (1:24:20.700)

And there is still a major problem

Lex Fridman (1:24:24.940)

in underperformance for school achievement

Richard Haier (1:24:29.100)

for many disadvantaged and minority students.

Lex Fridman (1:24:33.400)

And there so far is no way to fix it.

Lex Fridman (1:24:37.020)

What do we do with this information?

Lex Fridman (1:24:39.780)

Is this now a task?

Richard Haier (1:24:42.060)

Now we'll talk about the future

Lex Fridman (1:24:45.620)

on the neuroscience and the biology side,

Lex Fridman (1:24:47.700)

but in terms of this information as a society

Lex Fridman (1:24:51.100)

in the public policy, in the political space,

Lex Fridman (1:24:53.480)

in the social space, what do we do with this information?

Lex Fridman (1:24:56.260)

I've thought a lot about this.

Richard Haier (1:24:57.880)

The first step is to have people interested in policy

Lex Fridman (1:25:03.220)

understand what the data actually show

Richard Haier (1:25:06.940)

to pay attention to intelligence data.

Lex Fridman (1:25:09.800)

You can read policy papers about education

Lex Fridman (1:25:13.460)

and using your word processor,

Lex Fridman (1:25:15.840)

you can search for the word intelligence.

Richard Haier (1:25:17.860)

You can search a 20,000 word document in a second

Lex Fridman (1:25:22.100)

and find out the word intelligence does not appear anywhere.

Richard Haier (1:25:26.780)

In most discussions about what to do about achievement gaps,

Lex Fridman (1:25:32.500)

I'm not talking about test gaps,

Richard Haier (1:25:33.640)

I'm talking about actual achievement gaps in schools,

Lex Fridman (1:25:37.340)

which everyone agrees is a problem,

Richard Haier (1:25:40.140)

the word intelligence doesn't appear among educators.

Lex Fridman (1:25:43.940)

That's fascinating.

Richard Haier (1:25:44.780)

As a matter of fact, in California,

Lex Fridman (1:25:47.660)

there has been tremendous controversy

Richard Haier (1:25:50.020)

about recent attempts to revise the curriculum

Lex Fridman (1:25:53.100)

for math in high schools.

Lex Fridman (1:25:56.380)

And we had a Stanford professor of education

Lex Fridman (1:25:59.140)

who was running this review assert

Richard Haier (1:26:02.820)

there's no such thing as talent, mathematical talent.

Lex Fridman (1:26:07.860)

And she wanted to get rid of the advanced classes in math

Richard Haier (1:26:12.300)

because not everyone could do that.

Lex Fridman (1:26:15.340)

Now, of course, this has been very controversial,

Richard Haier (1:26:17.260)

they've retreated somewhat,

Lex Fridman (1:26:19.260)

but the idea that a university professor

Richard Haier (1:26:21.620)

was in charge of this who believes

Lex Fridman (1:26:26.140)

that there's no talent, that it doesn't exist,

Richard Haier (1:26:31.840)

this is rather shocking,

Lex Fridman (1:26:33.540)

let alone the complete absence of intelligence data.

Richard Haier (1:26:37.460)

By the way, let me tell you something

Lex Fridman (1:26:38.980)

about what the intelligence data show.

Richard Haier (1:26:41.980)

Let's take race out of it.

Lex Fridman (1:26:45.540)

Even though the origins of these studies

Richard Haier (1:26:48.260)

were a long time ago,

Lex Fridman (1:26:52.020)

I'm blocking on the name of the report,

Richard Haier (1:26:53.900)

the Coleman report was a famous report about education.

Lex Fridman (1:26:57.700)

And they measured all kinds of variables about schools,

Richard Haier (1:27:01.900)

about teachers,

Lex Fridman (1:27:03.460)

and they looked at academic achievement as an outcome.

Lex Fridman (1:27:08.100)

And they found the most predictive variables

Lex Fridman (1:27:12.660)

of education outcome were the variables

Richard Haier (1:27:15.980)

the student brought with him or her into the school,

Lex Fridman (1:27:20.340)

essentially their ability.

Lex Fridman (1:27:23.380)

And that when you combine the school

Lex Fridman (1:27:26.380)

and the teacher variables together,

Richard Haier (1:27:29.460)

the quality of the school, the funding of the school,

Lex Fridman (1:27:31.860)

the quality of the teachers, their education,

Richard Haier (1:27:34.620)

you put all the teacher and school variables together,

Lex Fridman (1:27:37.540)

it barely accounted for 10% of the variance.

Lex Fridman (1:27:41.660)

And this has been replicated now.

Lex Fridman (1:27:45.340)

So the best research we have shows that school variables

Lex Fridman (1:27:51.660)

and teacher variables together account

Lex Fridman (1:27:54.740)

for about 10% of student academic achievement.

Richard Haier (1:27:59.500)

Now, you wanna have some policy

Lex Fridman (1:28:02.060)

on improving academic achievement,

Lex Fridman (1:28:04.780)

how much money do you wanna put into teacher education?

Lex Fridman (1:28:08.340)

How much money do you wanna put into the quality

Lex Fridman (1:28:11.620)

of the school administration?

Lex Fridman (1:28:14.220)

You know who you can ask?

Richard Haier (1:28:15.340)

You can ask the Gates Foundation,

Lex Fridman (1:28:18.100)

because they spent a tremendous amount of money doing that.

Lex Fridman (1:28:21.500)

And at the end of it, because they're measurement people,

Lex Fridman (1:28:25.100)

they wanna know the data,

Richard Haier (1:28:27.540)

they found it had no impact at all.

Lex Fridman (1:28:29.660)

And they've kind of pulled out of that kind of program.

Richard Haier (1:28:33.460)

So, oh boy.

Lex Fridman (1:28:36.220)

Let me ask you, this is me talking, but there's...

Richard Haier (1:28:41.380)

Just the two of us.

Lex Fridman (1:28:42.660)

Just the two of us, but I'm gonna say

Richard Haier (1:28:44.900)

some funny and ridiculous things,

Lex Fridman (1:28:46.660)

so you're surely not approving of it.

Lex Fridman (1:28:51.260)

But there's a movie called Clerks.

Lex Fridman (1:28:53.940)

You probably...

Richard Haier (1:28:54.780)

I've seen it, I've seen it, yeah.

Lex Fridman (1:28:56.380)

There's a funny scene in there where a lovely couple

Richard Haier (1:28:59.860)

are talking about the number

Lex Fridman (1:29:01.460)

of previous sexual partners they had.

Lex Fridman (1:29:03.940)

And the woman says that,

Lex Fridman (1:29:07.820)

I believe she just had a handful,

Richard Haier (1:29:09.580)

like two or three or something like that sexual partners,

Lex Fridman (1:29:12.620)

but then she also mentioned that she...

Lex Fridman (1:29:17.860)

What's that called?

Lex Fridman (1:29:19.020)

Fallacia, what's the scientific?

Lex Fridman (1:29:20.380)

But she went, you know, gave a blow job

Lex Fridman (1:29:23.060)

to 37 guys, I believe it is.

Lex Fridman (1:29:26.180)

And so that has to do with the truth.

Lex Fridman (1:29:29.980)

So sometimes, knowing the truth

Richard Haier (1:29:35.940)

can get in the way of a successful relationship

Lex Fridman (1:29:39.540)

of love of some of the human flourishing.

Lex Fridman (1:29:43.020)

And that seems to me that's at the core here,

Lex Fridman (1:29:46.380)

that facing some kind of truth

Richard Haier (1:29:49.860)

that's not able to be changed

Lex Fridman (1:29:53.740)

makes it difficult to sort of...

Richard Haier (1:29:56.260)

Is limiting as opposed to empowering.

Lex Fridman (1:29:59.940)

That's the concern.

Richard Haier (1:30:01.300)

If you sort of test for intelligence

Lex Fridman (1:30:03.940)

and lay the data out,

Richard Haier (1:30:06.180)

it feels like you will give up on certain people.

Lex Fridman (1:30:09.140)

You will sort of start bidding people,

Richard Haier (1:30:12.580)

it's like, well, this person is like,

Lex Fridman (1:30:15.620)

let's focus on the average people

Richard Haier (1:30:18.060)

or let's focus on the very intelligent people.

Lex Fridman (1:30:20.060)

That's the concern.

Lex Fridman (1:30:21.500)

And there's a kind of intuition

Lex Fridman (1:30:26.700)

that if we just don't measure

Lex Fridman (1:30:29.580)

and we don't use that data,

Lex Fridman (1:30:31.660)

that we will treat everybody equal

Lex Fridman (1:30:33.620)

and give everybody equal opportunity.

Lex Fridman (1:30:37.860)

If we have the data in front of us,

Richard Haier (1:30:39.660)

we're likely to misdistribute

Lex Fridman (1:30:43.820)

the amount of sort of attention we allocate,

Richard Haier (1:30:46.340)

resources we allocate to people.

Lex Fridman (1:30:49.420)

That's probably the concern.

Richard Haier (1:30:52.180)

It's a realistic concern,

Lex Fridman (1:30:55.100)

but I think it's a misplaced concern

Richard Haier (1:30:57.660)

if you wanna fix the problem.

Lex Fridman (1:31:00.540)

If you wanna fix the problem,

Richard Haier (1:31:02.060)

you have to know what the problem is.

Lex Fridman (1:31:03.900)

Yep.

Richard Haier (1:31:05.140)

Now, let me tell you this,

Lex Fridman (1:31:06.820)

let's go back to the bell curve,

Richard Haier (1:31:08.860)

not the bell curve, but the normal distribution.

Lex Fridman (1:31:11.340)

Yes, 16% of the population on average has an IQ under 85,

Richard Haier (1:31:20.540)

which means they're very hard.

Lex Fridman (1:31:22.100)

If you have an IQ under 85,

Richard Haier (1:31:24.220)

it's very hard to find gainful employment

Lex Fridman (1:31:26.780)

at a salary that sustains you

Lex Fridman (1:31:31.260)

at least minimally in modern life, okay?

Lex Fridman (1:31:35.420)

Not impossible, but it's very difficult.

Richard Haier (1:31:37.540)

16% of the population of the United States

Lex Fridman (1:31:41.380)

is about 51 or 52 million people with IQs under 85.

Richard Haier (1:31:47.420)

This is not a small issue.

Lex Fridman (1:31:51.620)

14 million children have IQs under 85.

Lex Fridman (1:31:57.100)

Is this something we wanna ignore?

Lex Fridman (1:31:59.820)

Does this have any, what is the Venn diagram between,

Richard Haier (1:32:03.540)

you know, when you have people with IQs under 85,

Lex Fridman (1:32:07.140)

and you have achievement in school or achievement in life?

Richard Haier (1:32:13.540)

There's a lot of overlap there.

Lex Fridman (1:32:16.020)

This is why, to go back to the IQ pill,

Richard Haier (1:32:18.940)

if there were a way to shift that curve toward the higher end,

Lex Fridman (1:32:27.380)

that would have a big impact.

Richard Haier (1:32:30.980)

If I could maybe, before we talk about the impact on life

Lex Fridman (1:32:34.860)

and so on, some of the criticisms of the bell curve.

Lex Fridman (1:32:38.500)

So Steven Jay Gould wrote that the bell curve

Lex Fridman (1:32:41.940)

rests on four incorrect assumptions.

Richard Haier (1:32:44.860)

It would be just interesting to get your thoughts

Lex Fridman (1:32:46.900)

on the four assumptions, which are,

Richard Haier (1:32:48.940)

intelligence must be reducible to a single number,

Lex Fridman (1:32:51.980)

intelligence must be capable of rank ordering people

Richard Haier (1:32:54.740)

in a linear order,

Lex Fridman (1:32:56.100)

intelligence must be primarily genetically based,

Lex Fridman (1:32:59.860)

and intelligence must be essentially immutable.

Lex Fridman (1:33:03.860)

Maybe not as criticisms, but as thoughts about intelligence.

Richard Haier (1:33:09.740)

Oh yeah, we could spend a lot of time on him.

Lex Fridman (1:33:13.620)

On Steven Jay Gould?

Richard Haier (1:33:14.700)

Yes.

Lex Fridman (1:33:15.700)

He wrote that in what, about 1985, 1984?

Richard Haier (1:33:20.700)

His views were overtly political, not scientific.

Lex Fridman (1:33:25.980)

He was a scientist,

Lex Fridman (1:33:27.380)

but his views on this were overtly political,

Lex Fridman (1:33:30.700)

and I would encourage people listening to this,

Richard Haier (1:33:33.820)

if they really want to understand his criticisms,

Lex Fridman (1:33:38.940)

they should just Google what he had to say,

Lex Fridman (1:33:44.420)

and Google the scientific reviews of his book,

Lex Fridman (1:33:49.740)

The Mismeasure of Man,

Lex Fridman (1:33:51.500)

and they will take these statements apart.

Lex Fridman (1:33:54.740)

They were wrong, not only were they wrong,

Lex Fridman (1:33:57.940)

but when he asserted in his first book

Lex Fridman (1:34:00.380)

that there was no biological basis essentially to IQ,

Richard Haier (1:34:05.620)

by the time the second edition came around,

Lex Fridman (1:34:08.260)

there were studies of MRIs showing that brain size,

Richard Haier (1:34:13.980)

brain volume were correlated to IQ scores,

Lex Fridman (1:34:16.860)

which he declined to put in his book.

Lex Fridman (1:34:19.940)

So I'm learning a lot today.

Lex Fridman (1:34:21.900)

I didn't know actually the extent of his work.

Richard Haier (1:34:25.820)

I was just using a few little snippets of criticism.

Lex Fridman (1:34:28.940)

That's interesting.

Richard Haier (1:34:29.940)

There was a battle here.

Lex Fridman (1:34:30.940)

He wrote a book, Mismeasure of Man,

Richard Haier (1:34:32.820)

that's missing a lot of the scientific grounding.

Lex Fridman (1:34:36.220)

His book is highly popular in colleges today.

Richard Haier (1:34:39.700)

You can find it in any college bookstore

Lex Fridman (1:34:41.660)

under assigned reading.

Richard Haier (1:34:42.980)

It's highly popular.

Lex Fridman (1:34:44.220)

The Mismeasure of Man?

Richard Haier (1:34:45.220)

Yes, highly influential.

Lex Fridman (1:34:46.220)

Can you speak to the Mismeasure of Man?

Richard Haier (1:34:48.580)

I'm undereducated about this.

Lex Fridman (1:34:50.140)

So is this the book basically criticizing the ideas in the book?

Richard Haier (1:34:54.580)

Yeah, where those four things came from.

Lex Fridman (1:34:57.580)

And it is really a book that was really taken apart point by point

Richard Haier (1:35:04.580)

by a number of people who actually understood the data.

Lex Fridman (1:35:08.100)

And he didn't care.

Richard Haier (1:35:09.100)

Yeah.

Lex Fridman (1:35:10.100)

He didn't care.

Richard Haier (1:35:11.100)

He didn't modify anything.

Lex Fridman (1:35:12.100)

Listen, because this is such a sensitive topic,

Richard Haier (1:35:16.700)

like I said, I believe the impact of the work,

Lex Fridman (1:35:24.800)

because it is misinterpreted, has to be considered.

Richard Haier (1:35:28.620)

Because it's not just going to be scientific discourse,

Lex Fridman (1:35:31.380)

it's going to be political discourse,

Richard Haier (1:35:32.940)

there's going to be debates,

Lex Fridman (1:35:34.700)

there's going to be politically motivated people

Richard Haier (1:35:39.380)

that will use messages in each direction,

Lex Fridman (1:35:42.780)

make something like the bulk of the enemy

Richard Haier (1:35:45.060)

or the support for one's racist beliefs.

Lex Fridman (1:35:52.180)

And so I think you have to consider that.

Lex Fridman (1:35:55.260)

But it's difficult because Nietzsche was used by Hitler

Lex Fridman (1:35:59.300)

to justify a lot of his beliefs.

Lex Fridman (1:36:02.380)

And it's not exactly on Nietzsche to anticipate Hitler

Lex Fridman (1:36:09.060)

or how his ideas will be misinterpreted and used for evil.

Lex Fridman (1:36:12.820)

But there is a balance there.

Lex Fridman (1:36:14.680)

So I understand.

Richard Haier (1:36:15.680)

This is really interesting.

Lex Fridman (1:36:16.680)

I didn't know.

Richard Haier (1:36:17.680)

Is there any criticism of the book you find compelling

Lex Fridman (1:36:20.740)

or interesting or challenging to you from a scientific perspective?

Richard Haier (1:36:23.740)

There were factual criticisms about the nature of the statistics

Lex Fridman (1:36:29.900)

that were used, the statistical analyses.

Richard Haier (1:36:32.620)

These are more technical criticisms.

Lex Fridman (1:36:34.940)

And they were addressed by Murray in a couple of articles

Richard Haier (1:36:38.020)

where he took all the criticisms and spoke to them.

Lex Fridman (1:36:41.940)

And people listening to this podcast

Richard Haier (1:36:44.940)

can certainly find all those online.

Lex Fridman (1:36:47.940)

And it's very interesting.

Richard Haier (1:36:48.980)

Murray went on to write some additional books,

Lex Fridman (1:36:52.740)

two in the last couple of years, one about human diversity

Richard Haier (1:36:57.860)

where he goes through the data refuting the idea that race

Lex Fridman (1:37:02.980)

is only a social construct with no biological meaning.

Richard Haier (1:37:07.980)

He discusses the data.

Lex Fridman (1:37:09.820)

It's a very good discussion.

Richard Haier (1:37:11.020)

You don't have to agree with it.

Lex Fridman (1:37:12.500)

But he presents data in a cogent way.

Lex Fridman (1:37:16.420)

And he talks about the critics of that.

Lex Fridman (1:37:19.060)

And he talks about their data in a cogent, nonpersonal way.

Richard Haier (1:37:23.460)

It's a very informative discussion.

Lex Fridman (1:37:26.860)

The book is called Human Diversity.

Richard Haier (1:37:28.940)

He talks about race.

Lex Fridman (1:37:29.900)

And he talks about gender, same thing, about sex differences.

Lex Fridman (1:37:34.620)

And more recently, he's written what

Lex Fridman (1:37:37.180)

might be his final say on this, a book called Facing Reality

Richard Haier (1:37:43.060)

where he talks about this again.

Lex Fridman (1:37:46.220)

So he can certainly defend himself.

Richard Haier (1:37:49.580)

He doesn't need me to do that.

Lex Fridman (1:37:52.100)

But I would urge people who have heard

Richard Haier (1:37:55.060)

about him and the bell curve and who

Lex Fridman (1:37:58.500)

think they know what's in it, you are likely incorrect.

Lex Fridman (1:38:03.500)

And you need to read it for yourself.

Lex Fridman (1:38:06.300)

But it is, scientifically, it's a serious subject.

Richard Haier (1:38:12.220)

It's a difficult subject.

Lex Fridman (1:38:13.580)

Ethically, it's a difficult subject.

Richard Haier (1:38:15.980)

Everything you said here, calmly and thoughtfully,

Lex Fridman (1:38:19.740)

is difficult. It's difficult for me

Richard Haier (1:38:21.620)

to even consider that G factor exists.

Lex Fridman (1:38:26.420)

I don't mean from like that somehow G factor is inherently

Richard Haier (1:38:29.860)

racist or sexist or whatever.

Lex Fridman (1:38:32.260)

It's just it's difficult in the way

Richard Haier (1:38:35.260)

that considering the fact that we die one day is difficult.

Lex Fridman (1:38:38.220)

That we are limited by our biology is difficult.

Lex Fridman (1:38:42.820)

And at least from an American perspective,

Lex Fridman (1:38:47.740)

you would like to believe that everything

Richard Haier (1:38:49.300)

is possible in this world.

Lex Fridman (1:38:51.380)

Well, that leads us to what I think

Richard Haier (1:38:56.220)

we should do with this information.

Lex Fridman (1:38:59.980)

And what I think we should do with this information

Richard Haier (1:39:03.300)

is unusual.

Lex Fridman (1:39:07.660)

Because I think what we need to do

Richard Haier (1:39:09.540)

is fund more neuroscience research on the molecular

Lex Fridman (1:39:13.420)

biology of learning and memory.

Richard Haier (1:39:16.780)

Because one definition of intelligence

Lex Fridman (1:39:22.060)

is based on how much you can learn

Lex Fridman (1:39:24.580)

and how much you can remember.

Lex Fridman (1:39:27.140)

And if you accept that definition of intelligence,

Richard Haier (1:39:30.860)

then there are molecular studies going on now,

Lex Fridman (1:39:35.580)

and Nobel Prizes being won on molecular biology

Richard Haier (1:39:40.820)

or molecular neurobiology of learning and memory.

Lex Fridman (1:39:45.420)

Now, the step those researchers, those scientists

Richard Haier (1:39:49.500)

need to take when it comes to intelligence

Lex Fridman (1:39:53.180)

is to focus on the concept of individual differences.

Richard Haier (1:39:58.260)

Intelligence research has individual differences

Lex Fridman (1:40:03.740)

as its heart because it assumes that people

Richard Haier (1:40:08.900)

differ on this variable.

Lex Fridman (1:40:10.700)

And those differences are meaningful

Lex Fridman (1:40:13.020)

and need understanding.

Lex Fridman (1:40:15.900)

Cognitive psychologists who have morphed

Richard Haier (1:40:19.340)

into molecular biologists studying learning and memory

Lex Fridman (1:40:23.420)

hate the concept of individual differences historically.

Richard Haier (1:40:27.460)

Some now are coming around to it.

Lex Fridman (1:40:30.180)

I once sat next to a Nobel Prize winner

Richard Haier (1:40:34.740)

for his work on memory.

Lex Fridman (1:40:37.860)

And I asked him about individual differences.

Lex Fridman (1:40:41.340)

And he said, don't go there.

Lex Fridman (1:40:42.820)

It'll set us back 50 years.

Lex Fridman (1:40:46.580)

But I said, don't you think they're

Lex Fridman (1:40:48.660)

the key, though, to understand?

Lex Fridman (1:40:50.420)

Why can some people remember more than others?

Lex Fridman (1:40:53.980)

He said, you don't want to go there.

Richard Haier (1:40:55.980)

I think the 21st century will be remembered

Lex Fridman (1:40:58.700)

by the technology and the science that

Richard Haier (1:41:02.140)

goes to individual differences.

Lex Fridman (1:41:04.100)

Because we have now data.

Richard Haier (1:41:05.540)

We have now the tools to much, much better

Lex Fridman (1:41:07.380)

to start to measure, start to estimate,

Richard Haier (1:41:10.100)

not just on the sort of through tests and IQ test type

Lex Fridman (1:41:13.940)

of things, sort of outside the body kind of things,

Lex Fridman (1:41:18.220)

but measuring all kinds of stuff about the body.

Lex Fridman (1:41:20.380)

So yeah, truly go into the molecular biology,

Richard Haier (1:41:23.140)

to the neurobiology, to the neuroscience.

Lex Fridman (1:41:27.260)

Let me ask you about life.

Lex Fridman (1:41:31.940)

How does intelligence correlate with or lead to

Lex Fridman (1:41:36.820)

or has anything to do with career success?

Richard Haier (1:41:39.060)

You've mentioned these kinds of things.

Lex Fridman (1:41:40.660)

And is there any data?

Richard Haier (1:41:43.380)

You had an excellent conversation

Lex Fridman (1:41:44.780)

with Jordan Peterson, for example.

Richard Haier (1:41:46.740)

Is there any data on what intelligent

Lex Fridman (1:41:49.820)

means for success in life?

Richard Haier (1:41:53.380)

Success in life.

Lex Fridman (1:41:54.540)

There is a tremendous amount of validity data

Richard Haier (1:42:00.620)

that looked at intelligence test scores and various measures

Lex Fridman (1:42:08.260)

of life success.

Richard Haier (1:42:11.380)

Now, of course, life success is a pretty broad topic.

Lex Fridman (1:42:17.420)

And not everybody agrees on what success means.

Lex Fridman (1:42:22.100)

But there's general agreement on certain aspects of success

Lex Fridman (1:42:28.460)

that can be measured.

Richard Haier (1:42:33.180)

Including life expectancy, like you said.

Lex Fridman (1:42:35.060)

Life expectancy.

Richard Haier (1:42:36.500)

Now, there's life success.

Lex Fridman (1:42:42.020)

Life expectancy, I mean, that is such an interesting finding.

Lex Fridman (1:42:47.460)

But IQ scores are also correlated to things like income.

Lex Fridman (1:42:54.020)

Now, OK, so who thinks income means you're successful?

Richard Haier (1:42:59.900)

That's not the point.

Lex Fridman (1:43:01.340)

The point is that income is one empirical measure

Richard Haier (1:43:06.540)

in this culture that says something

Lex Fridman (1:43:09.300)

about your level of success.

Richard Haier (1:43:11.500)

You can define success in ways that

Lex Fridman (1:43:13.620)

have nothing to do with income.

Richard Haier (1:43:15.980)

You can define success based on your evolutionary natural

Lex Fridman (1:43:21.100)

selection success.

Lex Fridman (1:43:25.220)

But for variables, and even that, by the way,

Lex Fridman (1:43:29.780)

is correlated to IQ in some studies.

Lex Fridman (1:43:33.940)

So however you want to define success, IQ is important.

Lex Fridman (1:43:42.300)

It's not the only determinant.

Lex Fridman (1:43:44.260)

People get hung up on, well, what about personality?

Lex Fridman (1:43:46.780)

What about so called emotional intelligence?

Richard Haier (1:43:49.300)

Yes, all those things matter.

Lex Fridman (1:43:51.980)

The thing that matters empirically,

Richard Haier (1:43:54.460)

the single thing that matters the most

Lex Fridman (1:43:56.380)

is your general ability, your general mental intellectual

Richard Haier (1:44:01.180)

ability, your reasoning ability.

Lex Fridman (1:44:03.580)

And the more complex your vocation,

Richard Haier (1:44:07.100)

the more complex your job, the more G matters.

Lex Fridman (1:44:11.500)

G doesn't matter in a lot of occupations

Richard Haier (1:44:14.580)

don't require complex thinking.

Lex Fridman (1:44:17.020)

And there are occupations like that, and G doesn't matter.

Richard Haier (1:44:20.940)

Within an occupation, the G might not matter so much.

Lex Fridman (1:44:28.300)

So that if you look at all the professors at MIT

Lex Fridman (1:44:36.260)

and had a way to rank order them,

Lex Fridman (1:44:40.220)

there's a ceiling effect is what I'm saying.

Richard Haier (1:44:43.060)

That, you know.

Lex Fridman (1:44:45.140)

Also, when you get past a certain threshold,

Richard Haier (1:44:47.140)

then there's impact on wealth, for example,

Lex Fridman (1:44:49.900)

or career success, however that's

Richard Haier (1:44:52.460)

defined in each individual discipline.

Lex Fridman (1:44:54.220)

But after a certain point, it doesn't matter.

Richard Haier (1:44:56.780)

Actually, it does matter in certain things.

Lex Fridman (1:44:59.380)

So for example, there is a very classic study

Richard Haier (1:45:04.500)

that was started at Johns Hopkins when

Lex Fridman (1:45:07.220)

I was a graduate student there.

Lex Fridman (1:45:08.780)

And I actually worked on this study at the very beginning.

Lex Fridman (1:45:11.260)

It's the study of mathematically and scientifically

Richard Haier (1:45:13.700)

precocious youth.

Lex Fridman (1:45:15.660)

And they gave junior high school students

Richard Haier (1:45:20.140)

age 11 and 12 the standard SAT math exam.

Lex Fridman (1:45:27.140)

And they found a very large number of students

Richard Haier (1:45:31.420)

scored very high on this exam.

Lex Fridman (1:45:33.900)

Not a large number.

Richard Haier (1:45:35.100)

I mean, they found many students when

Lex Fridman (1:45:37.700)

they cast the net to all of Baltimore.

Richard Haier (1:45:40.700)

They found a number of students who

Lex Fridman (1:45:42.900)

scored as high on the SAT math when

Richard Haier (1:45:45.460)

they were 12 years old as incoming Hopkins freshmen.

Lex Fridman (1:45:50.220)

And they said, gee, now this is interesting.

Lex Fridman (1:45:53.580)

What shall we do now?

Lex Fridman (1:45:56.860)

And on a case by case basis, they

Richard Haier (1:46:00.140)

got some of those kids into their local community college

Lex Fridman (1:46:03.860)

math programs.

Richard Haier (1:46:06.660)

Many of those kids went on to be very successful.

Lex Fridman (1:46:10.260)

And now there's a 50 year follow up of those kids.

Lex Fridman (1:46:14.940)

And it turns out these kids were in the top 1%.

Lex Fridman (1:46:21.380)

So everybody in this study is in the top 1%.

Richard Haier (1:46:24.820)

If you take that group, that rarefied group,

Lex Fridman (1:46:28.100)

and divide them into quartiles so that you have the top 25%

Richard Haier (1:46:33.100)

of the top 1% and the bottom 25% of the top 1%,

Lex Fridman (1:46:39.780)

you can find unmeasurable variables of success.

Richard Haier (1:46:48.020)

The top quartile does better than the bottom quartile

Lex Fridman (1:46:51.740)

in the top 1%.

Richard Haier (1:46:53.780)

They have more patents.

Lex Fridman (1:46:54.920)

They have more publications.

Richard Haier (1:46:56.360)

They have more tenure at universities.

Lex Fridman (1:46:59.520)

And this is based on, you're dividing them

Richard Haier (1:47:03.140)

based on their score at age 12.

Lex Fridman (1:47:05.980)

I wonder how much interesting data

Richard Haier (1:47:10.220)

is in the variability in the differences.

Lex Fridman (1:47:12.700)

So but that's really, oh, boy.

Richard Haier (1:47:16.460)

That's very interesting.

Lex Fridman (1:47:17.620)

But it's also, I don't know, somehow painful.

Richard Haier (1:47:21.020)

I don't know why it's so painful that that G

Lex Fridman (1:47:25.020)

factor is so determinant of even in the nuanced top percent.

Richard Haier (1:47:31.140)

This is interesting that you find that painful.

Lex Fridman (1:47:32.780)

Do you find it painful that people with charisma

Richard Haier (1:47:37.180)

are very successful, can be very successful in life,

Lex Fridman (1:47:40.740)

even though having no other attributes other than they're

Lex Fridman (1:47:43.900)

famous and people like them?

Lex Fridman (1:47:45.780)

Do you find that painful?

Richard Haier (1:47:47.380)

Yes, if that charisma is untrainable.

Lex Fridman (1:47:51.100)

So one of the things, again, this

Richard Haier (1:47:53.500)

is like I learned psychology from the Johnny Depp trial.

Lex Fridman (1:47:56.740)

But one of the things the psychologist, the personality

Richard Haier (1:48:01.700)

psychologist, he can maybe speak to this

Lex Fridman (1:48:03.380)

because he had an interest in this for a time,

Richard Haier (1:48:07.220)

is she was saying that personality, technically

Lex Fridman (1:48:12.580)

speaking, is the thing that doesn't change over a lifetime.

Richard Haier (1:48:16.980)

It's the thing you're, I don't know if she was actually

Lex Fridman (1:48:20.140)

implying that you're born with it.

Richard Haier (1:48:21.660)

Well, it's a trait.

Lex Fridman (1:48:22.780)

It's a trait that's state.

Richard Haier (1:48:24.580)

It's a trait that's relatively stable over time.

Lex Fridman (1:48:27.220)

I think that's generally correct.

Lex Fridman (1:48:28.940)

So to the degree your personality

Lex Fridman (1:48:31.060)

is stable over time, yes, that too is painful.

Richard Haier (1:48:36.820)

Because what's not painful is the thing,

Lex Fridman (1:48:40.700)

if I'm fat and out of shape, I can

Richard Haier (1:48:42.660)

exercise and become healthier in that way.

Lex Fridman (1:48:47.900)

If my diet is a giant mess and that's

Richard Haier (1:48:51.020)

resulting in some kind of conditions

Lex Fridman (1:48:53.860)

that my body is experiencing, I can fix that

Richard Haier (1:48:55.980)

by having a better diet.

Lex Fridman (1:48:58.100)

That sort of my actions, my willed actions

Richard Haier (1:49:02.100)

can make a change.

Lex Fridman (1:49:03.860)

If charisma is part of the personality that's,

Richard Haier (1:49:07.940)

the part of the charisma that is part of the personality that

Lex Fridman (1:49:10.900)

is stable, yeah, yeah, that's painful too.

Richard Haier (1:49:15.380)

Because it's like, oh shit, I'm stuck with this.

Lex Fridman (1:49:18.420)

I'm stuck with this.

Richard Haier (1:49:19.860)

Well, and this pretty much generalizes

Lex Fridman (1:49:22.860)

to every aspect of your being.

Richard Haier (1:49:24.980)

This is who you are.

Lex Fridman (1:49:26.460)

You've got to deal with it.

Lex Fridman (1:49:27.540)

And what it undermines, of course,

Lex Fridman (1:49:29.340)

is a realistic appreciation for this,

Richard Haier (1:49:32.580)

undermines the fairly recent idea prevalent in this country

Lex Fridman (1:49:40.740)

that if you work hard, you can be anything you want to be,

Richard Haier (1:49:44.020)

which has morphed from the original idea

Lex Fridman (1:49:47.420)

that if you work hard, you can be successful.

Richard Haier (1:49:50.860)

Those are two different things.

Lex Fridman (1:49:53.900)

And now we have if you work hard,

Richard Haier (1:49:57.860)

you can be anything you want to be.

Lex Fridman (1:50:00.860)

This is completely unrealistic.

Richard Haier (1:50:03.460)

Sorry.

Lex Fridman (1:50:04.500)

It just is.

Richard Haier (1:50:05.180)

Now, you can work hard and be successful.

Lex Fridman (1:50:06.900)

There's no question.

Lex Fridman (1:50:08.500)

But you know what?

Lex Fridman (1:50:09.980)

I could work very hard, and I am not

Richard Haier (1:50:12.420)

going to be a successful theoretical physicist.

Lex Fridman (1:50:16.820)

I'm just not.

Richard Haier (1:50:18.740)

That said, I mean, we should, because we

Lex Fridman (1:50:21.020)

had this conversation already, but it's good to repeat.

Richard Haier (1:50:26.340)

The fact that you're not going to be

Lex Fridman (1:50:27.700)

a theoretical physicist is not judgment

Richard Haier (1:50:31.100)

on your basic humanity.

Lex Fridman (1:50:32.820)

We're turning again to the all men, which means

Richard Haier (1:50:36.180)

men and women are created equal.

Lex Fridman (1:50:39.100)

So again, some of the differences

Richard Haier (1:50:40.620)

we're talking about in quote unquote success, wealth,

Lex Fridman (1:50:47.020)

number of whether you win a Nobel Prize or not,

Richard Haier (1:50:50.620)

that doesn't put a measure on your basic humanity

Lex Fridman (1:50:55.820)

and basic value and even goodness of you

Richard Haier (1:51:00.940)

as a human being.

Lex Fridman (1:51:02.500)

Because your basic role and value in society

Richard Haier (1:51:06.900)

is largely within your control.

Lex Fridman (1:51:11.940)

It's some of these measures that we're talking about.

Richard Haier (1:51:16.740)

It's good to remember this.

Lex Fridman (1:51:19.380)

One question about the Flynn effect.

Lex Fridman (1:51:22.260)

What is it?

Lex Fridman (1:51:23.260)

Are humans getting smarter over the years, over the decades,

Lex Fridman (1:51:27.140)

over the centuries?

Lex Fridman (1:51:28.820)

The Flynn effect is James Flynn, who passed away about a year

Richard Haier (1:51:33.500)

ago, published a set of analyses going back

Lex Fridman (1:51:42.420)

a couple of decades when he first noticed this,

Richard Haier (1:51:46.020)

that IQ scores, when you looked over the years,

Lex Fridman (1:51:51.460)

seemed to be drifting up.

Richard Haier (1:51:54.420)

Now, this was not unknown to the people who make the test

Lex Fridman (1:51:58.820)

because they renorm the test periodically

Lex Fridman (1:52:02.220)

and they have to renorm the test periodically

Lex Fridman (1:52:05.260)

because what 10 items correct meant

Richard Haier (1:52:09.700)

relative to other people 50 years ago

Lex Fridman (1:52:13.700)

is not the same as what 10 items mean relative today.

Richard Haier (1:52:18.580)

People are getting more things correct.

Lex Fridman (1:52:21.220)

Now, the scores have been drifting up about three points.

Richard Haier (1:52:24.980)

IQ scores have been drifting up about three points per decade.

Lex Fridman (1:52:30.180)

This is not a personal effect.

Richard Haier (1:52:31.700)

This is a cohort effect.

Lex Fridman (1:52:34.020)

Well, it's not for an individual, but.

Lex Fridman (1:52:37.100)

The world, how do you explain?

Lex Fridman (1:52:38.700)

So what's that?

Lex Fridman (1:52:39.500)

And this has presented intelligence researchers

Lex Fridman (1:52:42.420)

with a great mystery.

Richard Haier (1:52:44.580)

Two questions.

Lex Fridman (1:52:46.220)

First, is it effect on the 50% of the variance that's

Lex Fridman (1:52:51.140)

the G factor or on the other 50%?

Lex Fridman (1:52:55.060)

And there's evidence that it is a G factor effect.

Lex Fridman (1:52:59.620)

And second, what on earth causes this?

Lex Fridman (1:53:02.660)

And doesn't this mean intelligence and G factor

Richard Haier (1:53:05.780)

cannot be genetic because the scale of natural selection

Lex Fridman (1:53:10.220)

is much, much longer than a couple of decades ago?

Lex Fridman (1:53:15.660)

And so it's been used to try to undermine the idea

Lex Fridman (1:53:19.700)

that there can be a genetic influence on intelligence.

Lex Fridman (1:53:24.420)

But certainly, it can be the Flynn effect

Lex Fridman (1:53:28.540)

can affect the nongenetic aspects of intelligence

Richard Haier (1:53:32.060)

because genes account for maybe 50% of the variance.

Lex Fridman (1:53:37.700)

Maybe higher, could be as high as 80% for adults,

Lex Fridman (1:53:40.900)

but let's just say 50% for discussion.

Lex Fridman (1:53:46.420)

So the Flynn effect, it's still a mystery.

Richard Haier (1:53:50.260)

It's still a mystery.

Lex Fridman (1:53:51.180)

That's interesting.

Richard Haier (1:53:51.780)

It's still a mystery, although the evidence is coming out.

Lex Fridman (1:53:54.060)

I told you before I edited a journal on intelligence,

Lex Fridman (1:53:56.700)

and we're doing a special issue in honor of James Flynn.

Lex Fridman (1:54:00.460)

So I'm starting to see papers now on really

Richard Haier (1:54:03.060)

the latest research on this.

Lex Fridman (1:54:06.300)

I think most people who specialize

Richard Haier (1:54:08.820)

in this area of trying to understand the Flynn effect

Lex Fridman (1:54:12.460)

are coming to the view based on data

Richard Haier (1:54:16.460)

that it has to do with advances in nutrition and health care.

Lex Fridman (1:54:20.180)

And there's also evidence that the effect is slowing down

Lex Fridman (1:54:25.100)

and possibly reversing.

Lex Fridman (1:54:27.500)

Oh, boy.

Lex Fridman (1:54:28.340)

So how would nutrition and health,

Lex Fridman (1:54:30.580)

so nutrition would still be connected to the G factor.

Lex Fridman (1:54:36.300)

So nutrition as it relates to the G factor,

Lex Fridman (1:54:38.660)

so the biology that leads to the intelligence.

Richard Haier (1:54:42.060)

Yes.

Lex Fridman (1:54:42.580)

That would be the claim.

Richard Haier (1:54:43.820)

Like the hypothesis being tested by the research.

Lex Fridman (1:54:50.300)

Yes.

Lex Fridman (1:54:50.820)

And there's some evidence from infants

Lex Fridman (1:54:54.460)

that nutrition has made a difference.

Lex Fridman (1:54:58.540)

So it's not an unreasonable connection.

Lex Fridman (1:55:02.140)

But does it negate the idea that there's a genetic influence?

Richard Haier (1:55:05.740)

Not logically at all.

Lex Fridman (1:55:07.980)

But it is very interesting.

Lex Fridman (1:55:09.860)

So that if you take an IQ test today but you take the score

Lex Fridman (1:55:17.460)

and use the tables that were available in 1940,

Richard Haier (1:55:22.900)

you're going to wind up with a much higher IQ number.

Lex Fridman (1:55:27.020)

So are we really smarter than a couple of generations ago?

Richard Haier (1:55:32.700)

No, but we might be able to solve problems a little better.

Lex Fridman (1:55:38.300)

And make use of our G because of things like Sesame Street

Lex Fridman (1:55:43.140)

and other curricula in school.

Lex Fridman (1:55:45.380)

More people are going to school.

Lex Fridman (1:55:49.540)

So there are a lot of factors here to disentangle.

Lex Fridman (1:55:53.820)

It's fascinating though.

Richard Haier (1:55:55.060)

It's fascinating that there's not clear answers yet.

Lex Fridman (1:55:58.860)

That as a population, we're getting smarter.

Richard Haier (1:56:02.620)

When you just zoom out, that's what it looks like.

Lex Fridman (1:56:04.500)

As a population, we're getting smarter.

Lex Fridman (1:56:06.020)

And it's interesting to see what the effects of that are.

Lex Fridman (1:56:08.740)

I mean, this raises the question.

Richard Haier (1:56:10.060)

We've mentioned it many times but haven't clearly addressed it,

Lex Fridman (1:56:14.420)

which is nature versus nurture question.

Lex Fridman (1:56:17.260)

So how much of intelligence is nature?

Lex Fridman (1:56:20.060)

How much of it is nurture?

Lex Fridman (1:56:21.860)

How much of it is determined by genetics versus environment?

Lex Fridman (1:56:25.700)

All of it.

Richard Haier (1:56:26.700)

All of it is genetics.

Lex Fridman (1:56:28.100)

No, all of it is nature and nurture.

Lex Fridman (1:56:32.620)

So yes.

Lex Fridman (1:56:33.620)

Yes.

Richard Haier (1:56:34.620)

Okay.

Lex Fridman (1:56:36.620)

But how much of the variance can you apportion to either?

Richard Haier (1:56:44.940)

Most of the people who work in this field say that the framing of that, if the question

Richard Haier (1:56:51.580)

is framed that way, it can't be answered because nature and nurture are not two independent

Richard Haier (1:56:57.500)

influences.

Lex Fridman (1:56:59.460)

They interact with each other.

Lex Fridman (1:57:01.420)

And understanding those interactions is so complex that many behavioral geneticists say

Richard Haier (1:57:10.220)

it is today impossible and always will be impossible to disentangle that, no matter

Lex Fridman (1:57:17.380)

what kind of advances there are in DNA technology and genomic informatics.

Lex Fridman (1:57:24.700)

But there's still, to push back on that, that same intuition from behavioral geneticists

Richard Haier (1:57:31.580)

would lead me to believe that there cannot possibly be a stable G factor because it's

Lex Fridman (1:57:37.340)

super complex.

Richard Haier (1:57:39.300)

Many of them would assert that as a logical outcome.

Lex Fridman (1:57:45.300)

But because I believe there is a stable G factor from lots of sources of data, not just

Richard Haier (1:57:51.820)

one study, but lots of sources of data over decades, I am more amenable to the idea that

Richard Haier (1:58:01.380)

whatever interactions between genes and environment exist, they can be explicated, they can be

Richard Haier (1:58:09.940)

studied, and that information can be used as a basis for molecular biology of intelligence.

Richard Haier (1:58:19.300)

Yeah, and we'll do this exact question because doesn't the stability of the G factor give

Lex Fridman (1:58:27.300)

you at least a hint that there is a biological basis for intelligence?

Richard Haier (1:58:33.780)

Yes, I think it's clear that the fact that an IQ score is correlated to things like thickness

Richard Haier (1:58:42.820)

of your cortex, that it's correlated to glucose metabolic rate in your brain, that identical

Lex Fridman (1:58:55.340)

twins reared apart are highly similar in their IQ scores.

Richard Haier (1:59:02.060)

These are all important observations that indicate, not just suggest, but indicate that

Lex Fridman (1:59:12.140)

there's a biological basis.

Lex Fridman (1:59:13.540)

And does anyone believe intelligence has nothing to do with the brain?

Lex Fridman (1:59:18.620)

I mean, it's so obvious.

Richard Haier (1:59:21.220)

Well indirectly definitely has to do with it, but the question is environment interacting

Lex Fridman (1:59:26.620)

with the brain or is it the actual raw hardware of the brain?

Richard Haier (1:59:34.920)

Well some would say that the raw hardware of the brain as it develops from conception

Richard Haier (1:59:45.500)

through adulthood, or at least through the childhood, that that so called hardware that

Richard Haier (1:59:52.700)

you are assuming is mostly genetic, in fact, is not as deterministic as you might think,

Lex Fridman (20:00.620)

It really, it's a fascinating group of people

Richard Haier (20:04.420)

who are studied.

Lex Fridman (20:08.100)

Not to get ahead of the story,

Lex Fridman (20:09.300)

but one of the most interesting things they found

Lex Fridman (20:12.540)

is a very strong relationship

Richard Haier (20:14.660)

between IQ measured at age 11 and mortality.

Lex Fridman (20:21.680)

So that, you know,

Richard Haier (20:24.380)

in the 70 years later, they looked at the survival rates

Lex Fridman (20:30.940)

and they could get death records from everybody.

Lex Fridman (20:33.300)

And Scotland has universal healthcare for everybody.

Lex Fridman (20:37.140)

And it turned out if you divide the people

Richard Haier (20:40.020)

by their age 11 IQ score into quartiles

Lex Fridman (20:44.020)

and then look at how many people are alive 70 years later,

Richard Haier (20:49.460)

the, I know this is in the book,

Lex Fridman (20:52.420)

I have the graph in the book,

Lex Fridman (20:54.660)

but there are essentially twice as many people alive

Lex Fridman (20:57.820)

in the highest IQ quartile than in the lowest IQ quartile.

Richard Haier (21:01.820)

It's true in men and women.

Lex Fridman (21:05.300)

Interesting.

Lex Fridman (21:06.940)

So it makes a big difference.

Lex Fridman (21:08.100)

Now, why this is the case is not so clear

Richard Haier (21:12.920)

since everyone had access to healthcare.

Lex Fridman (21:15.620)

Well, there's a lot, and we'll talk about it, you know,

Richard Haier (21:18.340)

just the sentences you used now

Lex Fridman (21:22.100)

could be explained by nature or nurture.

Richard Haier (21:25.780)

We don't know.

Lex Fridman (21:26.980)

Now, there's a lot of science that starts to then dig in

Lex Fridman (21:29.700)

and investigate that question.

Lex Fridman (21:31.700)

But let me linger on the IQ test.

Lex Fridman (21:33.700)

How are the test design, IQ test design, how do they work?

Lex Fridman (21:37.380)

Maybe some examples for people who are not aware.

Lex Fridman (21:39.980)

What makes a good IQ test question

Lex Fridman (21:44.020)

that sneaks up on this G factor measure?

Richard Haier (21:48.080)

Well, your question is interesting

Lex Fridman (21:49.740)

because you want me to give examples of items

Richard Haier (21:53.340)

that make good items.

Lex Fridman (21:55.220)

And what makes a good item is not so much its content,

Lex Fridman (21:59.420)

but its empirical relationship to the total score

Lex Fridman (22:03.180)

that turns out to be valid by other means.

Lex Fridman (22:07.740)

So for example, let me give you an odd example

Lex Fridman (22:12.500)

from personality testing.

Richard Haier (22:14.300)

Nice.

Lex Fridman (22:15.560)

So there's a personality test

Richard Haier (22:18.020)

called the Minnesota Multiphasic Personality Inventory, MMPI.

Lex Fridman (22:22.740)

Been around for decades.

Richard Haier (22:24.100)

I've heard about this test recently

Lex Fridman (22:25.980)

because of the Johnny Depp and Amber Heard trial.

Richard Haier (22:29.240)

I don't know if you've been paying attention to that.

Lex Fridman (22:31.180)

But they had psychologists.

Richard Haier (22:32.020)

I have not been paying attention to it.

Lex Fridman (22:33.540)

They had psychologists on the stand,

Lex Fridman (22:35.820)

and they were talking, apparently those psychologists did,

Lex Fridman (22:39.860)

again, I'm learning so much from this trial.

Richard Haier (22:42.020)

They did different battery of tests

Lex Fridman (22:45.820)

to diagnose personality disorders.

Richard Haier (22:50.420)

Apparently there's that systematic way of doing so,

Lex Fridman (22:53.300)

and the Minnesota one is one of the ones

Richard Haier (22:55.740)

that there's the most science on.

Lex Fridman (22:59.020)

There's a lot of great papers,

Richard Haier (23:00.440)

which were all continuously cited on the stand,

Lex Fridman (23:03.820)

which is fascinating to watch.

Richard Haier (23:05.060)

Sorry, a little bit of attention.

Lex Fridman (23:06.540)

It's okay.

Richard Haier (23:07.380)

I mean, this is interesting because you're right.

Lex Fridman (23:08.220)

It's been around for decades.

Richard Haier (23:09.540)

There's a lot of scientific research

Lex Fridman (23:11.220)

on the psychometric properties of the test,

Richard Haier (23:14.820)

including what it predicts with respect

Lex Fridman (23:18.000)

to different categories of personality disorder.

Lex Fridman (23:22.320)

But what I wanna mention is the content

Lex Fridman (23:24.840)

of the items on that test.

Richard Haier (23:26.860)

All of the items are essentially true false items.

Lex Fridman (23:32.440)

True or false, I prefer a shower to a bath.

Richard Haier (23:36.860)

True or false, I think Lincoln

Lex Fridman (23:39.900)

was a better president than Washington.

Lex Fridman (23:42.260)

But what of all these, what does that have to do?

Lex Fridman (23:47.560)

And the point is the content of these items,

Richard Haier (23:49.720)

nobody knows why these items in aggregate predict anything,

Lex Fridman (23:55.360)

but empirically they do.

Richard Haier (23:57.900)

It's a technique of choosing items for a test

Lex Fridman (24:01.880)

that is called dust bowl empiricism.

Richard Haier (24:05.800)

That the content doesn't matter,

Lex Fridman (24:07.440)

but for some reason when you get a criterion group

Richard Haier (24:10.640)

of people with this disorder and you compare them

Lex Fridman (24:13.600)

to people without that disorder,

Richard Haier (24:16.120)

these are the items that distinguish,

Lex Fridman (24:19.200)

irrespective of content.

Richard Haier (24:20.640)

It's a hard concept to grasp.

Lex Fridman (24:22.700)

Well, first of all, it's fascinating.

Lex Fridman (24:25.260)

But from, because I consider myself part psychologist

Lex Fridman (24:33.000)

because I love human robot interaction,

Lex Fridman (24:35.360)

and that's a problem.

Lex Fridman (24:36.600)

Half of that problem is a psychology problem

Richard Haier (24:39.600)

because there's a human.

Lex Fridman (24:41.800)

So designing these tests to get at the questions

Richard Haier (24:45.200)

is the fascinating part.

Lex Fridman (24:46.440)

Like how do you get to,

Lex Fridman (24:50.000)

like what does dust bowl empiricism refer to?

Lex Fridman (24:52.700)

Does it refer to the final result?

Richard Haier (24:57.460)

Yeah, so it's the test is dust bowl empiricism.

Lex Fridman (25:01.540)

But how do you arrive at the battery of questions?

Richard Haier (25:04.960)

I presume one of the things,

Lex Fridman (25:07.560)

now again, I'm going to the excellent testimony

Richard Haier (25:10.080)

in that trial, they explain it,

Lex Fridman (25:12.600)

because they also, they explain the tests.

Richard Haier (25:16.720)

That a bunch of the questions are kind of

Lex Fridman (25:20.320)

make you forget that you're taking a test.

Richard Haier (25:24.040)

Like it makes it very difficult for you

Lex Fridman (25:26.720)

to somehow figure out what you're supposed to answer.

Richard Haier (25:31.640)

Yes, it's called social desirability.

Lex Fridman (25:34.120)

But we're getting a little far afield

Richard Haier (25:35.580)

because I only wanted to give that example

Lex Fridman (25:37.400)

of dust bowl empiricism.

Richard Haier (25:40.500)

When we talk about the items on an IQ test,

Lex Fridman (25:44.400)

many of those items in the dust bowl empiricism method

Richard Haier (25:50.640)

have no face validity.

Lex Fridman (25:52.960)

In other words, they don't look like they measure anything.

Richard Haier (25:56.440)

Yes.

Lex Fridman (25:57.540)

Whereas most intelligence tests,

Richard Haier (25:59.920)

the items actually look like they're measuring

Lex Fridman (26:02.220)

some mental ability.

Lex Fridman (26:03.880)

So here's one of the.

Lex Fridman (26:05.320)

So you were bringing that up as an example

Richard Haier (26:07.000)

as what it is not.

Lex Fridman (26:08.200)

Yes.

Richard Haier (26:09.040)

Got it.

Lex Fridman (26:09.860)

Okay.

Lex Fridman (26:10.700)

So I don't want to go too far afield on it.

Lex Fridman (26:12.880)

Too far afield is actually one of the names of this podcast.

Lex Fridman (26:16.120)

So I should mention that.

Lex Fridman (26:19.080)

Far afield.

Richard Haier (26:20.000)

Far afield.

Lex Fridman (26:21.120)

Yeah, so anyway, sorry.

Lex Fridman (26:22.240)

So they feel the questions look like

Lex Fridman (26:25.000)

they pass the face validity test.

Lex Fridman (26:28.040)

And some more than others.

Lex Fridman (26:29.720)

So for example, let me give you a couple of things here.

Richard Haier (26:32.680)

If I, one of the subtests on a standard IQ test

Lex Fridman (26:37.320)

is general information.

Richard Haier (26:40.860)

Let me just think a little bit

Lex Fridman (26:41.880)

because I don't want to give you the actual item.

Lex Fridman (26:44.360)

But if I said, how far is it between Washington DC

Lex Fridman (26:49.820)

and Miami, Florida?

Richard Haier (26:52.320)

Within 500 miles plus or minus.

Lex Fridman (26:55.700)

Well, you know, it's not a fact most people memorize,

Lex Fridman (27:00.080)

but you know something about geography.

Lex Fridman (27:02.720)

You say, well, I flew there once.

Richard Haier (27:04.480)

I know planes fly for 500 miles.

Lex Fridman (27:06.600)

You know, you can kind of make an estimate.

Lex Fridman (27:10.360)

But it's also seems like it would be very cultural,

Lex Fridman (27:15.280)

you know, so there's that kind of general information.

Richard Haier (27:20.240)

Then there's vocabulary test.

Lex Fridman (27:22.520)

What does regatta mean?

Lex Fridman (27:27.520)

And I choose that word because that word was removed

Lex Fridman (27:31.320)

from the IQ test because people complained

Richard Haier (27:33.760)

that disadvantaged people would not know that word

Lex Fridman (27:38.280)

just from their everyday life.

Richard Haier (27:41.440)

Okay, here's another example

Lex Fridman (27:43.920)

from a different kind of subtest on.

Lex Fridman (27:46.520)

What's regatta, by the way?

Lex Fridman (27:48.200)

Regatta is a.

Richard Haier (27:50.400)

I think I'm disadvantaged.

Lex Fridman (27:51.640)

A sailing competition, a competition with boats.

Richard Haier (27:54.800)

Not necessarily sailing, but a competition with boats.

Lex Fridman (27:58.760)

Yep, yep, I'm probably disadvantaged in that way.

Richard Haier (28:02.200)

Okay, excellent, so that was removed anyway you were saying.

Lex Fridman (28:04.960)

Okay, so here's another subtest.

Richard Haier (28:07.840)

I'm gonna repeat a string of numbers,

Lex Fridman (28:09.840)

and when I'm done, I want you to repeat them back to me.

Lex Fridman (28:12.680)

Ready?

Lex Fridman (28:13.720)

Okay, seven, four, two, eight, one, six.

Richard Haier (28:21.540)

That's way too many.

Lex Fridman (28:22.560)

Seven, four, two, eight, one, six.

Richard Haier (28:25.160)

Okay, you get the idea.

Lex Fridman (28:26.280)

Now the actual test starts with a smaller number,

Richard Haier (28:30.800)

like two numbers, and then as people get it right,

Lex Fridman (28:33.500)

you keep going, adding to the string of numbers

Richard Haier (28:36.520)

until they can't do it anymore.

Lex Fridman (28:38.600)

Okay, but now try this.

Richard Haier (28:40.720)

I'm gonna say some numbers, and when I'm done,

Lex Fridman (28:43.760)

I want you to repeat them to me backwards.

Richard Haier (28:46.640)

I quit.

Lex Fridman (28:47.800)

Okay, now, so I gave you some examples

Richard Haier (28:51.520)

of the kind of items on an IQ test.

Lex Fridman (28:53.600)

General information, I can't even remember all,

Richard Haier (28:58.640)

general information, vocabulary, digit span forward

Lex Fridman (29:03.680)

and digit span backward.

Richard Haier (29:06.760)

Well, you said I can't even remember them.

Lex Fridman (29:08.960)

That's a good question for me.

Lex Fridman (29:11.480)

What does memory have to do with GFactor?

Lex Fridman (29:13.600)

Okay, well, let's hold on.

Richard Haier (29:15.200)

Okay, all right.

Lex Fridman (29:16.280)

Let's just talk about these examples.

Richard Haier (29:19.720)

Now, some of those items seem very cultural,

Lex Fridman (29:26.760)

and others seem less cultural.

Richard Haier (29:31.960)

Which ones do you think, scores on which subtest

Lex Fridman (29:35.400)

are most highly correlated with the GFactor?

Richard Haier (29:39.640)

Well, the intuitive answer is less cultural.

Lex Fridman (29:42.560)

Well, it turns out vocabulary is highly correlated,

Lex Fridman (29:49.860)

and it turns out that digit span backwards

Lex Fridman (29:54.220)

is highly correlated.

Lex Fridman (29:55.740)

How do you figure?

Lex Fridman (29:58.620)

Now you have decades of research to answer the question,

Lex Fridman (2:00:01.220)

but it is probabilistic and what affects the probabilities are things like in uterine environment

Lex Fridman (2:00:09.740)

and other factors like that, including chance.

Richard Haier (2:00:15.140)

That chance affects the way the neurons are connecting during gestation.

Lex Fridman (2:00:22.680)

It's not, hey, it's pre programmed.

Lex Fridman (2:00:26.820)

So there is push back on the concept that genes provide a blueprint, that it's a lot

Lex Fridman (2:00:35.500)

more fluid.

Richard Haier (2:00:36.500)

Well, but also, yeah, so there's a lot, a lot, a lot happens in the first few months

Lex Fridman (2:00:44.020)

of development.

Lex Fridman (2:00:47.340)

So in nine months inside the mother's body and in the few months afterwards, there's

Richard Haier (2:00:57.180)

a lot of fascinating stuff, like including chance and luck, like you said, how things

Richard Haier (2:01:02.120)

connect up.

Richard Haier (2:01:04.220)

The question is afterwards in your plasticity of the brain, how much adjustment there is

Richard Haier (2:01:08.460)

relative to the environment, how much that affects the G factor, but that's where the

Richard Haier (2:01:14.140)

whole conclusions of the studies that we've been talking about is that seems to have less

Lex Fridman (2:01:19.140)

and less and less of an effect as pretty quickly.

Richard Haier (2:01:23.620)

As yes, and I do think there is more of a genetic, by my view, and I'm not an expert

Richard Haier (2:01:30.060)

on this, I mean, genetics is a highly technical and complex subject.

Richard Haier (2:01:34.180)

I am not a geneticist, not a behavioral geneticist, but my reading of this, my interpretation

Richard Haier (2:01:41.900)

of this is that there is a genetic blueprint, more or less, and that has a profound influence

Lex Fridman (2:01:50.880)

on your subsequent intellectual development, including the G factor.

Lex Fridman (2:01:56.700)

And that's not to say things can't happen to, I mean, if you think of that genes provide

Richard Haier (2:02:03.900)

a potential, fine, and then various variables impact that potential, and every parent of

Richard Haier (2:02:12.060)

a newborn, implicitly or explicitly, wants to maximize that potential.

Lex Fridman (2:02:19.520)

This is why you buy educational toys.

Richard Haier (2:02:21.980)

This is why you pay attention to organic baby food.

Richard Haier (2:02:25.660)

This is why you do all these things, because you want your baby to be as healthy and as

Richard Haier (2:02:31.620)

smart as possible, and every parent will say that.

Richard Haier (2:02:36.000)

Is there a case to be made, can you steel man the case, that genetics is a very tiny

Lex Fridman (2:02:45.220)

component of all of this, and the environment is essential?

Lex Fridman (2:02:49.420)

I don't think the data supports that genetics is a tiny component.

Richard Haier (2:02:53.460)

I think the data support the idea that the genetics is a very important, and I don't

Richard Haier (2:02:58.260)

say component, I say influence, a very important influence, and the environment is a lot less

Richard Haier (2:03:05.740)

than people believe.

Lex Fridman (2:03:07.900)

Most people believe environment plays a big role.

Richard Haier (2:03:10.700)

I'm not so sure.

Lex Fridman (2:03:11.700)

I guess what I'm asking you is, can you see where what you just said, it might be wrong?

Lex Fridman (2:03:19.420)

Can you imagine a world, and what kind of evidence would you need to see to say, you

Lex Fridman (2:03:27.140)

know what, the intuition, the studies so far, like reversing the directions.

Lex Fridman (2:03:31.700)

So one of the cool things we have now more and more is we're getting more and more data,

Lex Fridman (2:03:36.140)

and the rate of the data is escalating because of the digital world.

Lex Fridman (2:03:41.860)

So when you start to look at a very large scale of data, both on the biology side and

Richard Haier (2:03:48.700)

the social side, we might be discovering some very counterintuitive things about society.

Richard Haier (2:03:53.860)

We might see the edge cases that reveal that if we actually scale those edge cases and

Richard Haier (2:04:00.300)

they become like the norm, that we'll have a complete shift in our, like you'll see G

Richard Haier (2:04:08.080)

factor be able to be modified throughout life in the teens and in later life.

Lex Fridman (2:04:15.300)

So is it any case you can make or for where your current intuitions are wrong?

Richard Haier (2:04:20.700)

Yes, and it's a good question because I think everyone should always be asked what evidence

Lex Fridman (2:04:25.300)

would change your mind.

Richard Haier (2:04:28.020)

It's certainly not only a fair question, it is really the key question for anybody working

Lex Fridman (2:04:32.860)

on any aspect of science.

Richard Haier (2:04:36.580)

I think that if environment was very important, we would have seen it clearly by now.

Richard Haier (2:04:45.900)

It would have been obvious that school interventions, compensatory education, early childhood education,

Richard Haier (2:04:53.660)

all these things that have been earnestly tried in well funded, well designed studies

Lex Fridman (2:04:59.340)

would show some effect, and they don't.

Lex Fridman (2:05:02.940)

What if the school, the way we've tried school, compensatory school sucks and we need to do

Lex Fridman (2:05:08.260)

better?

Richard Haier (2:05:09.260)

That's what everybody said at the beginning.

Lex Fridman (2:05:10.260)

That's what everybody said to Jensen.

Richard Haier (2:05:11.760)

He said, well, maybe we need to start earlier.

Lex Fridman (2:05:15.720)

Maybe we need not do prekindergarten, but pre, prekindergarten.

Richard Haier (2:05:20.320)

It's always an infinite, well, maybe we didn't get it right.

Lex Fridman (2:05:24.620)

But after decades of trying, 50 years, 50 or 60 years of trying, surely something would

Richard Haier (2:05:33.580)

have worked to the point where you could actually see a result and not need a probability level

Lex Fridman (2:05:39.700)

at 0.05 on some means.

Lex Fridman (2:05:42.540)

So that's why I, that's the kind of evidence that would change my mind.

Richard Haier (2:05:49.380)

Population level interventions like schooling that you would see like this actually has

Richard Haier (2:05:56.940)

an effect.

Lex Fridman (2:05:57.940)

Yes.

Lex Fridman (2:05:58.940)

And when you take adopted kids and they grow up in another family and you find out when

Richard Haier (2:06:04.940)

those adopted kids are adults, their IQ scores don't correlate with the IQ scores of their

Richard Haier (2:06:09.980)

adoptive parents, but they do correlate with their IQ scores of their biological parents

Lex Fridman (2:06:15.820)

whom they've never met.

Richard Haier (2:06:18.860)

I mean, these are important, these are powerful observations.

Lex Fridman (2:06:22.100)

And it would be convincing to you if the reverse was true.

Richard Haier (2:06:26.500)

Yes.

Lex Fridman (2:06:27.500)

That would be more.

Lex Fridman (2:06:28.500)

And there is some data on adoption that indicates that the adopted children are moving a little

Lex Fridman (2:06:35.180)

bit more toward their adoptive parents.

Lex Fridman (2:06:40.220)

But it's to me the overwhelming, I have this concept called the weight of evidence where

Lex Fridman (2:06:47.500)

I don't interpret any one study too much.

Richard Haier (2:06:50.500)

The weight of evidence tells me genes are important.

Lex Fridman (2:06:53.540)

But what does that mean?

Lex Fridman (2:06:54.780)

What does it mean that genes are important?

Richard Haier (2:06:57.100)

Knowing that gene expression, genes don't express themselves in a vacuum, they express

Richard Haier (2:07:03.180)

themselves in an environment.

Lex Fridman (2:07:05.900)

So the environment has to have something to do with it, especially if the best genetic

Richard Haier (2:07:10.300)

estimates of the amount of variants are around 50 or even if it's as high as 80%, it still

Lex Fridman (2:07:17.740)

leaves 20% of non genetic.

Richard Haier (2:07:21.540)

Now maybe that is all luck.

Lex Fridman (2:07:24.300)

Maybe that's all chance.

Richard Haier (2:07:25.580)

I could believe that, I could easily believe that.

Lex Fridman (2:07:31.320)

But I do think after 50 years of trying various interventions and nothing works, including

Richard Haier (2:07:39.060)

memory training, including listening to Mozart, including playing computer games, none of

Lex Fridman (2:07:44.340)

that has shown any impact on intelligence test scores.

Lex Fridman (2:07:49.060)

Is there data on the intelligence, the IQ of parents as it relates to the children?

Richard Haier (2:07:57.140)

Yes, and there is some genetic evidence of an interaction between the parents IQ and

Richard Haier (2:08:07.020)

the environment.

Richard Haier (2:08:08.760)

High IQ parents provide an enriched environment, which then can impact the child in addition

Richard Haier (2:08:17.020)

to the genes, it's that environment.

Lex Fridman (2:08:20.020)

So there are all these interactions that, think about the number of books in a household.

Lex Fridman (2:08:29.060)

This was a variable that's correlated with IQ and, well, why?

Richard Haier (2:08:37.140)

Especially if the kid never reads any of the books, it's because more intelligent people

Richard Haier (2:08:42.900)

have more books in their house.

Lex Fridman (2:08:45.620)

And if you're more intelligent and there's a genetic component to that, the child will

Richard Haier (2:08:52.500)

get those genes or some of those genes as well as the environment.

Lex Fridman (2:08:57.480)

But it's not the number of books in the house that actually directly impacts the child.

Lex Fridman (2:09:04.140)

So the two scenarios on this are you find that, and this was used to get rid of the

Richard Haier (2:09:11.620)

SAT test, oh, the SAT score is highly correlated with the social economic status of the parents.

Lex Fridman (2:09:18.300)

So all you're really measuring is how rich the parents are.

Lex Fridman (2:09:21.540)

Okay, well, why are the parents rich?

Lex Fridman (2:09:27.620)

And so the opposite kind of syllogism is that people who are very bright make more money,

Lex Fridman (2:09:37.940)

they can afford homes in better neighborhoods so their kids get better schools.

Richard Haier (2:09:44.540)

Now the kids grow up bright.

Lex Fridman (2:09:47.780)

Where in that chain of events does that come from?

Richard Haier (2:09:51.340)

Well, unless you have a genetically informative research design where you look at siblings

Richard Haier (2:09:58.720)

that have the same biological parents and so on, you can't really disentangle all that.

Richard Haier (2:10:05.940)

Most studies of social economic status and intelligence do not have a genetically informed

Lex Fridman (2:10:12.820)

design.

Lex Fridman (2:10:13.940)

So any conclusions they make about the causality of the social economic status being the cause

Lex Fridman (2:10:20.580)

of the IQ is a stretch.

Lex Fridman (2:10:25.540)

And where you do find genetically informative designs, you find most of the variance in

Lex Fridman (2:10:33.140)

your outcome measures are due to the genetic component.

Lex Fridman (2:10:38.980)

And sometimes the SES adds a little, but the weight of evidence is it doesn't add very

Lex Fridman (2:10:46.480)

much variance to predict what's going on beyond the genetic variance.

Lex Fridman (2:10:52.660)

So when you actually look at it in some, and there aren't that many studies that have genetically

Richard Haier (2:10:58.780)

informed designs, but when you do see those, the genes seem to have an advantage.

Richard Haier (2:11:05.060)

Sorry for the strange questions, but is there a connection between fertility or the number

Lex Fridman (2:11:13.620)

of kids that you have and G factor?

Lex Fridman (2:11:16.860)

So you know, the kind of conventional wisdom is people of maybe higher economic status

Lex Fridman (2:11:25.600)

or something like that are having fewer children.

Richard Haier (2:11:28.180)

I just loosely hear these kinds of things.

Lex Fridman (2:11:30.900)

Is there data that you're aware of in one direction or another on this?

Richard Haier (2:11:36.580)

Strange questions always get strange answers.

Lex Fridman (2:11:39.060)

Yes.

Richard Haier (2:11:40.060)

All right.

Lex Fridman (2:11:41.060)

Do you have a strange answer for that strange question?

Richard Haier (2:11:44.740)

The answer is there were some studies that indicated the more children in a family, the

Lex Fridman (2:11:54.540)

firstborn children would be more intelligent than the fourth or fifth or sixth.

Richard Haier (2:12:00.740)

It's not clear that those studies hold up over time.

Lex Fridman (2:12:05.500)

And of course what you see also is that families where there are multiple children, four, five,

Richard Haier (2:12:14.020)

six, seven, you know, really big families, the social economic status of those families

Lex Fridman (2:12:23.580)

usually in the modern age is not that high.

Richard Haier (2:12:28.700)

Maybe it used to be the aristocracy used to have a lot of kids, I'm not sure exactly.

Lex Fridman (2:12:33.500)

But there have been reports of correlations between IQ and fertility, but I'm not sure

Richard Haier (2:12:44.020)

that the data are very strong that the firstborn child is always the smartest.

Lex Fridman (2:12:50.500)

It seems like there's some data to that, but I'm not current on that.

Lex Fridman (2:12:54.980)

How would that be explained?

Lex Fridman (2:12:55.980)

That would be in a nurture.

Richard Haier (2:12:58.260)

Well, it could be nurture, it could be in uterine environment, I mean, and this is why

Richard Haier (2:13:08.500)

this, you know, like many areas of science, you said earlier that there are a lot of gray

Richard Haier (2:13:14.740)

areas and no definitive answers.

Richard Haier (2:13:21.660)

This is not uncommon in science that the closer you look at a problem, the more questions

Richard Haier (2:13:28.900)

you get, not the fewer questions, because the universe is complicated.

Lex Fridman (2:13:35.820)

And the idea that we have people on this planet who can study the first nanoseconds of the

Richard Haier (2:13:42.980)

Big Bang, that's pretty amazing.

Lex Fridman (2:13:48.260)

And I've always said that if they can study the first nanoseconds of the Big Bang, we

Richard Haier (2:13:53.340)

can certainly figure out something about intelligence that allows that.

Richard Haier (2:13:58.900)

I'm not sure what's more complicated, the human mind or the physics of the universe.

Richard Haier (2:14:06.020)

It's unclear to me.

Lex Fridman (2:14:08.220)

I think we overemphasize.

Richard Haier (2:14:09.220)

Well, that's a very humbling statement.

Richard Haier (2:14:13.220)

Maybe it's a very human centric, egotistical statement that our mind is somehow super complicated,

Lex Fridman (2:14:18.100)

but biology is a tricky one to unravel.

Lex Fridman (2:14:22.860)

Consciousness, what is that?

Richard Haier (2:14:27.100)

I've always believed that consciousness and intelligence are the two real fundamental

Lex Fridman (2:14:34.180)

problems of the human brain, and therefore I think they must be related.

Richard Haier (2:14:41.380)

Yeah, heart problems like walk together, holding hands kind of idea.

Richard Haier (2:14:49.060)

You may not know this, but I did some of the early research on anesthetic drugs with brain

Richard Haier (2:14:54.180)

imaging trying to answer the question, what part of the brain is the last to turn off

Lex Fridman (2:14:58.900)

when someone loses consciousness?

Lex Fridman (2:15:01.520)

And is that the first part of the brain to turn on when consciousness is regained?

Lex Fridman (2:15:07.060)

And I was working with an anesthesiologist named Mike Alkire, who was really brilliant

Richard Haier (2:15:11.060)

at this.

Richard Haier (2:15:12.060)

These were really the first studies of brain imaging using positron emission tomography

Richard Haier (2:15:18.300)

long before fMRI.

Lex Fridman (2:15:21.700)

And you would inject a radioactive sugar that labeled the brain, and the harder the brain

Richard Haier (2:15:28.220)

was working, the more sugar it would take up, and then you could make a picture of glucose

Lex Fridman (2:15:33.980)

use in the brain.

Lex Fridman (2:15:36.420)

And he was amazing.

Richard Haier (2:15:38.560)

He managed to do this in normal volunteers he brought in and anesthetized as if they

Richard Haier (2:15:44.860)

were going into surgery.

Richard Haier (2:15:49.060)

He managed all the human subjects requirements on this research, and he was brilliant at

Richard Haier (2:15:55.380)

this.

Lex Fridman (2:15:57.100)

And what we did is we had these normal volunteers come in on three occasions.

Richard Haier (2:16:05.140)

On one occasion, he gave them enough anesthetic drug so they were a little drowsy.

Lex Fridman (2:16:14.800)

And on another occasion, they came in and he fully anesthetized them.

Lex Fridman (2:16:20.780)

And he would say, Mike, can you hear me, and the person would say, uh, yeah.

Lex Fridman (2:16:31.820)

And then we would scan people under no anesthetic condition.

Lex Fridman (2:16:37.540)

So same person.

Lex Fridman (2:16:39.980)

And we were looking to see if we could see the part of the brain turn off.

Richard Haier (2:16:46.780)

He subsequently tried to do this with fMRI, which has a faster time resolution, and you

Richard Haier (2:16:51.980)

could do it in real time as the person went under and then regain consciousness where

Richard Haier (2:16:56.780)

you couldn't do that with PET.

Lex Fridman (2:16:57.980)

You had to have three different occasions.

Lex Fridman (2:17:00.780)

And the results were absolutely fascinating.

Richard Haier (2:17:03.380)

We did this with different anesthetic drugs, and different drugs impacted different parts

Richard Haier (2:17:08.600)

of the brain.

Lex Fridman (2:17:09.780)

So we were naturally looking for the common one, and it seemed to have something to do

Richard Haier (2:17:15.820)

with the thalamus.

Lex Fridman (2:17:18.300)

And consciousness, this was actual data on consciousness, actual consciousness.

Lex Fridman (2:17:25.740)

What part of the brain turns on?

Lex Fridman (2:17:28.420)

What part of the brain turns off?

Richard Haier (2:17:30.900)

It's not so clear.

Lex Fridman (2:17:33.220)

But maybe has something to do with the thalamus.

Richard Haier (2:17:35.940)

The sequence of events seemed to have the thalamus in it.

Lex Fridman (2:17:41.180)

Now here's the question.

Lex Fridman (2:17:42.700)

Are some people more conscious than others?

Lex Fridman (2:17:45.780)

Are there individual differences in consciousness?

Lex Fridman (2:17:49.660)

And I don't mean it in the psychedelic sense.

Lex Fridman (2:17:53.020)

I don't mean it in the political consciousness sense.

Richard Haier (2:17:55.780)

I just mean it in everyday life.

Lex Fridman (2:17:57.300)

Do some people go through everyday life more conscious than others?

Lex Fridman (2:18:01.900)

And are those the people we might actually label more intelligent?

Richard Haier (2:18:06.700)

Now the other thing I was looking for is whether the parts of the brain we were seeing in the

Richard Haier (2:18:11.700)

anesthesia studies were the same parts of the brain we were seeing in the intelligence

Lex Fridman (2:18:16.820)

studies.

Richard Haier (2:18:17.820)

Now, this was very complicated, expensive research.

Lex Fridman (2:18:22.580)

We didn't really have funding to do this.

Richard Haier (2:18:24.500)

We were trying to do it on the fly.

Lex Fridman (2:18:26.260)

I'm not sure anybody has pursued this.

Richard Haier (2:18:29.660)

I'm retired now.

Lex Fridman (2:18:31.740)

He's gone on to other things.

Lex Fridman (2:18:34.100)

But I think it's an area of research that would be fascinating to see the parts, a lot

Lex Fridman (2:18:41.020)

more imaging studies now of consciousness.

Richard Haier (2:18:43.140)

I'm just not up on them.

Lex Fridman (2:18:45.260)

But basically the question is which imaging, so newer imaging studies to see in high resolution,

Richard Haier (2:18:52.460)

spatial and temporal way, which part of the brain lights up when you're doing intelligence

Richard Haier (2:18:59.260)

tasks and which parts of the brain lights up when you're doing consciousness tasks and

Richard Haier (2:19:03.700)

see the interplay between them, try to infer, that's the challenge of neuroscience, without

Richard Haier (2:19:09.460)

understanding deeply, looking from the outside, try to infer something about how the whole

Richard Haier (2:19:18.260)

thing works.

Lex Fridman (2:19:19.260)

Well, imagine this.

Richard Haier (2:19:21.220)

Here's a simple question.

Richard Haier (2:19:23.160)

Does it take more anesthetic drug to have a person lose consciousness if their IQ is

Lex Fridman (2:19:33.460)

140 than a person with an IQ of 70?

Lex Fridman (2:19:39.060)

That's an interesting way to study it.

Richard Haier (2:19:40.660)

Yeah.

Lex Fridman (2:19:41.660)

I mean, if the answer to that is a stable yes, that's very interesting.

Lex Fridman (2:19:48.060)

So I tried to find out and I went to some anesthesiology textbooks about how you dose

Lex Fridman (2:19:55.700)

and they dose by weight.

Lex Fridman (2:19:59.620)

And what I also learned, this is a little bit off subject, anesthesiologists are never

Lex Fridman (2:20:07.440)

sure how deep you are.

Lex Fridman (2:20:10.980)

And they usually tell by poking you with a needle and if you don't jump, they tell the

Lex Fridman (2:20:14.780)

surgeon to go ahead.

Richard Haier (2:20:17.100)

I'm not sure that's literally true, but it's...

Lex Fridman (2:20:20.460)

Well, it might be very difficult to know precisely how deep you are.

Richard Haier (2:20:26.380)

It has to do with the same kind of measurements that you were doing with the consciousness.

Lex Fridman (2:20:31.860)

It's difficult to know.

Lex Fridman (2:20:34.740)

So I don't lose my train of thought.

Lex Fridman (2:20:35.940)

I couldn't find in the textbooks anything about dosing by intelligence.

Richard Haier (2:20:40.980)

I asked my friend, the anesthesiologist, he said, no, he doesn't know.

Richard Haier (2:20:45.620)

I said, can we do a chart review and look at people using their years of education as

Lex Fridman (2:20:52.220)

a proxy for IQ?

Lex Fridman (2:20:54.420)

Because if someone's gone to graduate school, that tells you something.

Richard Haier (2:20:58.340)

You can make some inference as opposed to someone who didn't graduate high school.

Lex Fridman (2:21:02.780)

Can we do a chart review?

Lex Fridman (2:21:03.980)

And he says, no, they never really put down the exact dose.

Lex Fridman (2:21:08.460)

And no, he said, no.

Lex Fridman (2:21:10.620)

So to this day, the simple question, does it take more anesthetic drug to put someone

Lex Fridman (2:21:18.820)

under if they have a high IQ or less, or less?

Richard Haier (2:21:23.440)

It could go either way.

Richard Haier (2:21:24.440)

Because by the way, our early PET scan studies of intelligence found the unexpected result

Richard Haier (2:21:33.780)

of an inverse correlation between glucose metabolic rate and intelligence.

Lex Fridman (2:21:38.540)

It wasn't how much a brain area lit up.

Lex Fridman (2:21:43.020)

How much it lit up was negatively correlated to how well they did on the test, which led

Lex Fridman (2:21:48.380)

to the brain efficiency hypothesis, which is still being studied today.

Lex Fridman (2:21:54.140)

And there's more and more evidence that the efficiency of brain information processing

Lex Fridman (2:22:00.280)

is more related to intelligence than just more activity.

Richard Haier (2:22:08.780)

Yeah, and it'll be interesting, again, this is the total hypothesis, how much in the relationship

Richard Haier (2:22:14.180)

between intelligence and consciousness, it's not obvious that those two, if there's correlation,

Richard Haier (2:22:22.060)

they could be inversely correlated.

Lex Fridman (2:22:23.860)

Wouldn't that be funny?

Richard Haier (2:22:26.060)

If you, the consciousness factor, the C factor plus the G factor equals one.

Richard Haier (2:22:38.340)

It's a nice trade off, you get a trade off, how deeply you experience the world versus

Lex Fridman (2:22:43.420)

how deeply you're able to reason through the world.

Lex Fridman (2:22:48.140)

What a great hypothesis.

Richard Haier (2:22:51.060)

Certainly somebody listening to this can do this study.

Richard Haier (2:22:54.140)

Even if it's the aliens analyzing humans a few centuries from now, let me ask you from

Richard Haier (2:22:59.740)

an AI perspective, I don't know how much you've thought about machines, but there's the famous

Richard Haier (2:23:08.600)

Turing test, test of intelligence for machines, which is a beautiful, almost like a cute formulation

Richard Haier (2:23:17.220)

of intelligence that Alan Turing proposed.

Richard Haier (2:23:24.220)

Basically conversation being, if you can fool a human to think that a machine is a human

Richard Haier (2:23:33.420)

that passes the test, I suppose you could do a similar thing for humans.

Richard Haier (2:23:40.620)

If I can fool you that I'm intelligent, then that's a good test of intelligence.

Richard Haier (2:23:48.700)

You're talking to two people, and the test is saying who has a higher IQ.

Richard Haier (2:24:02.500)

It's an interesting test, because maybe charisma can be very useful there, and you're only

Richard Haier (2:24:07.960)

allowed to use conversation, which is the formulation of the Turing test.

Lex Fridman (2:24:11.340)

Anyway, all that to say is what are good tests of intelligence for machines?

Lex Fridman (2:24:18.700)

What do you think it takes to achieve human level intelligence for machines?

Richard Haier (2:24:23.700)

I have thought a little bit about this, but every time I think about these things, I rapidly

Richard Haier (2:24:30.760)

reach the limits of my knowledge and imagination.

Richard Haier (2:24:37.580)

When Alexa first came out, and I think there was a competing one, well, there was Siri

Richard Haier (2:24:47.460)

with Apple, and Google had Alexa.

Lex Fridman (2:24:50.100)

No, no, Amazon had Alexa.

Richard Haier (2:24:52.660)

Amazon had Alexa.

Lex Fridman (2:24:53.660)

Google has Google Home.

Richard Haier (2:24:54.660)

Google has something.

Richard Haier (2:24:55.660)

I proposed to one of my colleagues that he buy one of these, one of each, and then ask

Richard Haier (2:25:04.020)

it questions from the IQ test.

Lex Fridman (2:25:09.020)

But it became apparent that they all searched the internet, so they all can find answers

Richard Haier (2:25:15.860)

to questions like how far is it between Washington and Miami, and repeat after me.

Richard Haier (2:25:22.540)

Now, I don't know if you said to Alexa, I'm going to repeat these numbers backwards to

Richard Haier (2:25:29.620)

me.

Lex Fridman (2:25:30.620)

I don't know what would happen.

Richard Haier (2:25:31.620)

I've never done it.

Lex Fridman (2:25:33.620)

So one answer to your question is you're going to try it right now.

Richard Haier (2:25:38.060)

Let's try it.

Lex Fridman (2:25:39.060)

No.

Richard Haier (2:25:40.060)

Let's try it.

Lex Fridman (2:25:41.060)

No, no, no.

Richard Haier (2:25:42.060)

Yes, Siri.

Lex Fridman (2:25:43.060)

So it would actually probably go to Google search, and it will be all confusing kind

Richard Haier (2:25:47.500)

of stuff.

Lex Fridman (2:25:49.500)

It would fail.

Richard Haier (2:25:50.500)

Well, then I guess there was a test that it would fail.

Richard Haier (2:25:53.980)

Well, but that's not, that has to do more with the language of communication versus

Richard Haier (2:26:02.560)

the content.

Lex Fridman (2:26:03.740)

So if you did an IQ test to a person who doesn't speak English, and the test was administered

Richard Haier (2:26:09.300)

in English, that's not really the test of...

Lex Fridman (2:26:11.660)

Well, let's think about the computers that beat the Jeopardy champions.

Richard Haier (2:26:15.740)

Yeah, so that, because I happen to know how those are programmed, those are very hard

Lex Fridman (2:26:21.760)

coded, and there's definitely a lack of intelligence there.

Richard Haier (2:26:25.660)

There's something like IQ tests, there's a guy, an artificial intelligence researcher,

Richard Haier (2:26:36.060)

Francois Chollet, he's at Google, he's one of the seminal people in machine learning.

Richard Haier (2:26:40.980)

He also, as a fun aside thing, developed an IQ test for machines.

Lex Fridman (2:26:45.860)

Oh, I haven't heard that.

Richard Haier (2:26:47.660)

I'd just like to know about that.

Lex Fridman (2:26:49.580)

I'll actually email you this, because it'd be very interesting for you.

Richard Haier (2:26:53.340)

It doesn't get much attention, because people don't know what to do with it, but it deserves

Richard Haier (2:26:59.900)

a lot of attention, which is, it basically does a pattern type of tests, where you have

Richard Haier (2:27:06.180)

to do, you know, one standard one is, you're given three things, and you have to do a fourth

Lex Fridman (2:27:12.740)

one, that kind of thing, so you have to understand the pattern here.

Lex Fridman (2:27:17.120)

And for that, it really simplifies to, so the interesting thing is, he's trying not

Lex Fridman (2:27:28.740)

to achieve high IQ, he's trying to achieve like, pretty low bar for IQ.

Richard Haier (2:27:35.300)

Things that are kind of trivial for humans, and they're actually really tough for machines.

Richard Haier (2:27:42.060)

It's just seeing, playing with these concepts of symmetry, of counting, like if I give you

Richard Haier (2:27:49.420)

one object, two objects, three objects, you'll know the last one is four objects, you can

Richard Haier (2:27:54.980)

like count them, you can cluster objects together, it's both visually and conceptually, we can

Richard Haier (2:28:01.300)

do all these things with our mind, that we take for granted, the objectness of things.

Richard Haier (2:28:07.420)

You can like, figure out what spatially is an object and isn't, and we can play with

Richard Haier (2:28:14.740)

those ideas, and machines really struggle with that, so he really cleanly formulated

Richard Haier (2:28:21.540)

these IQ tests, I wonder what like, that would equate to for humans with IQ, but it'd be

Richard Haier (2:28:27.340)

a very low IQ, but that's exactly the kind of formulation, like okay, we want to be able

Richard Haier (2:28:33.940)

to solve this, how do we solve this, and he does it as a challenge, and nobody's been

Richard Haier (2:28:38.060)

able to, it's similar to the Alexa prize, which is Amazon is hosting a conversational

Richard Haier (2:28:44.260)

challenge, nobody's been able to do well on his, but that's an interesting, those kinds

Richard Haier (2:28:51.440)

of tests are interesting, because we take for granted all the ability of the human mind

Richard Haier (2:28:58.500)

to play with concepts, and to formulate concepts out of novel things, so like, things we've

Richard Haier (2:29:08.620)

never seen before, we're able to use that, I mean that's, I've talked to a few people

Richard Haier (2:29:14.440)

that design IQ tests, sort of online, they write IQ tests, and I was trying to get some

Richard Haier (2:29:20.540)

questions from them, and they spoke to the fact that we can't really share questions

Richard Haier (2:29:25.180)

with you, because part of the, like first of all, it's really hard work to come up with

Richard Haier (2:29:31.980)

questions, it's really, really hard work, it takes a lot of research, but it also takes

Richard Haier (2:29:37.240)

a lot, it's novelty generating, you're constantly coming up with really new things, and part

Richard Haier (2:29:46.060)

of the point is that they're not supposed to be public, they're supposed to be new

Richard Haier (2:29:51.380)

to you when you look at them, it's interesting that the novelty is fundamental to the hardness

Richard Haier (2:29:56.700)

of the problem, at least a part of what makes the problem hard is that you've never seen

Lex Fridman (2:30:02.320)

it before.

Richard Haier (2:30:03.320)

Right, that's called fluid intelligence, as opposed to what's called crystallized intelligence,

Richard Haier (2:30:08.860)

which is your knowledge of facts, you know things, but can you use those things to solve

Richard Haier (2:30:16.200)

a problem, those are two different things.

Lex Fridman (2:30:19.500)

Do you think we'll be able to, because we spoke, I don't want to miss opportunity to

Richard Haier (2:30:25.180)

talk about this, we spoke about the neurobiology, about the molecular biology of intelligence,

Lex Fridman (2:30:30.300)

do you think one day we'll be able to modify the biology of, or the genetics of a person

Richard Haier (2:30:39.460)

to modify their intelligence, to increase their intelligence, we started this conversation

Lex Fridman (2:30:45.040)

by talking about a pill you could take, do you think that such a pill would exist?

Richard Haier (2:30:49.020)

Metaphorically, I do, and I am supremely confident that it's possible because I am supremely

Richard Haier (2:30:57.900)

ignorant of the complexities of neurobiology, and so I have written that the nightmares

Richard Haier (2:31:07.220)

of neurobiologists, understanding the complexities, this cascade of events that happens at the

Lex Fridman (2:31:15.860)

synaptic level, that these nightmares are what fuel some people to solve.

Lex Fridman (2:31:25.060)

So some people, you have to be undaunted, I mean yeah, this is not easy, look we're

Richard Haier (2:31:31.520)

still trying to figure out cancer, it was only recently that they figured out why aspirin

Richard Haier (2:31:38.660)

works, you know, these are not easy problems, but I also have the perspective of the history

Richard Haier (2:31:47.980)

of science, is the history of solving problems that are extraordinarily complex.

Lex Fridman (2:31:57.580)

And seem impossible at the time.

Lex Fridman (2:31:58.740)

And seem impossible at the time.

Lex Fridman (2:32:01.140)

And so one of the things you look at, at companies like Neuralink, you have brain computer interfaces,

Richard Haier (2:32:08.460)

you start to delve into the human mind and start to talk about machines measuring but

Richard Haier (2:32:12.500)

also sending signals to the human mind, and you start to wonder what that has, what impact

Lex Fridman (2:32:19.180)

that has on the G factor.

Richard Haier (2:32:23.700)

Modifying in small ways or in large ways the functioning, the mechanical, electrical, chemical

Lex Fridman (2:32:32.420)

functioning of the brain.

Richard Haier (2:32:34.140)

I look at everything about the brain, there are different levels of explanation.

Richard Haier (2:32:39.860)

On one hand you have a behavioral level, but then you have brain circuitry, and then you

Richard Haier (2:32:46.420)

have neurons, and then you have dendrites, and then you have synapses, and then you have

Richard Haier (2:32:57.380)

the neurotransmitters, and the presynaptic and the postsynaptic terminals, and then you

Richard Haier (2:33:06.580)

have all the things that influence neurotransmitters, and then you have the individual differences

Lex Fridman (2:33:13.980)

among people.

Richard Haier (2:33:15.460)

Yeah, it's complicated, but 51 million people in the United States have IQs under 85 and

Lex Fridman (2:33:27.540)

struggle with everyday life.

Lex Fridman (2:33:32.540)

Shouldn't that motivate people to take a look at this?

Richard Haier (2:33:37.140)

Yeah, but I just want to linger one more time that you have to remember that the science

Richard Haier (2:33:46.220)

of intelligence, the measure of intelligence is only a part of the human condition.

Richard Haier (2:33:54.740)

The thing that makes life beautiful and the creation of beautiful things in this world

Richard Haier (2:33:59.260)

is perhaps loosely correlated, but is not dependent entirely on intelligence.

Lex Fridman (2:34:08.860)

Absolutely, I certainly agree with that.

Lex Fridman (2:34:12.220)

So for anyone sort of listening, I'm still not convinced that more intelligence is always

Lex Fridman (2:34:22.660)

better if you want to create beauty in this world.

Richard Haier (2:34:26.060)

I don't know.

Richard Haier (2:34:27.060)

Well, I didn't say more intelligence is always better if you want to create beauty.

Richard Haier (2:34:31.340)

I just said all things being equal, more is better than less.

Lex Fridman (2:34:36.460)

That's all I mean.

Richard Haier (2:34:37.460)

Yeah, but that's sort of that I just want to sort of say because a lot to me, one of

Richard Haier (2:34:42.980)

the things that makes life great is the opportunity to create beautiful things, and so I just

Richard Haier (2:34:50.380)

want to sort of empower people to do that no matter what some IQ test says.

Richard Haier (2:34:56.380)

At the population level, we do need to look at IQ tests to help people and to also inspire

Richard Haier (2:35:02.420)

us to take on some of these extremely difficult scientific questions.

Lex Fridman (2:35:07.900)

Do you have advice for young people in high school, in college, whether they're thinking

Lex Fridman (2:35:16.860)

about career or they're thinking about a life they can be proud of?

Richard Haier (2:35:20.740)

Is there advice you can give whether they want to pursue psychology or biology or engineering

Lex Fridman (2:35:29.760)

or they want to be artists and musicians and poets?

Richard Haier (2:35:33.700)

I can't advise anybody on that level of what their passion is, but I can say if you're

Richard Haier (2:35:43.320)

interested in psychology or if you're interested in science and the science around the big

Richard Haier (2:35:52.840)

questions of consciousness and intelligence and psychiatric illness, we haven't really

Richard Haier (2:36:00.640)

talked about brain illnesses and what we might learn from.

Richard Haier (2:36:07.020)

If you are trying to develop a drug to treat Alzheimer's disease, you are trying to develop

Richard Haier (2:36:12.540)

a drug to impact learning and memory, which are core to intelligence.

Lex Fridman (2:36:20.420)

So it could well be that the so called IQ pill will come from a pharmaceutical company

Richard Haier (2:36:26.100)

trying to develop a drug for Alzheimer's disease.

Richard Haier (2:36:29.460)

Because that's exactly what you're trying to do, right, yeah, just like you said.

Lex Fridman (2:36:33.540)

What will that drug do in a college student that doesn't have Alzheimer's disease?

Lex Fridman (2:36:38.460)

So I would encourage people who are interested in psychology, who are interested in science

Richard Haier (2:36:47.740)

to pursue a scientific career and address the big questions.

Lex Fridman (2:36:54.300)

And the most important thing I can tell you if you're going to be in kind of a research

Richard Haier (2:37:03.080)

environment is you got to follow the data where the data take you.

Lex Fridman (2:37:07.420)

You can't decide in advance where you want the data to go.

Lex Fridman (2:37:10.700)

And if the data take you to places that you don't have the technical expertise to follow,

Richard Haier (2:37:16.300)

like you know, I would like to understand more about molecular biology, but I'm not

Richard Haier (2:37:21.940)

going to become a molecular biologist now.

Lex Fridman (2:37:24.940)

But I know people who are, and my job is to get them interested to take their expertise

Richard Haier (2:37:31.140)

into this direction.

Lex Fridman (2:37:33.640)

And that it's not so easy.

Lex Fridman (2:37:36.580)

And if the data takes you to a place that's controversial, that's counterintuitive in

Richard Haier (2:37:41.820)

this world, no, I would say it's probably a good idea to still push forward boldly,

Lex Fridman (2:37:52.460)

but to communicate the interpretation of the results with skill, with compassion, with

Richard Haier (2:38:01.380)

the greater breadth of understanding of humanity, not just the science, of the impact of the

Richard Haier (2:38:07.700)

results.

Richard Haier (2:38:08.900)

One famous psychologist wrote about this issue that somehow a balance has to be found between

Richard Haier (2:38:16.340)

pursuing the science and communicating it with respect to people's sensitivities, the

Lex Fridman (2:38:22.260)

legitimate sensitivities, somehow.

Richard Haier (2:38:26.020)

He didn't say how.

Lex Fridman (2:38:27.020)

Somehow.

Richard Haier (2:38:28.020)

Somehow.

Lex Fridman (2:38:29.020)

And this is...

Richard Haier (2:38:30.020)

This sense, somehow, and balance is left up to the interpretation of the reader.

Richard Haier (2:38:37.500)

Let me ask you, you said big questions, the biggest, or one of the biggest, we already

Richard Haier (2:38:44.220)

talked about consciousness and intelligence, one of the most fascinating, one of the biggest

Lex Fridman (2:38:48.060)

questions.

Lex Fridman (2:38:49.060)

But let's talk about the why.

Lex Fridman (2:38:51.420)

Why are we here?

Lex Fridman (2:38:53.420)

What's the meaning of life?

Lex Fridman (2:38:54.420)

I'm not going to tell you.

Lex Fridman (2:38:55.420)

You know you're not going to tell me?

Lex Fridman (2:38:56.420)

This is very...

Richard Haier (2:39:00.940)

I'm going to have to wait for your next book.

Lex Fridman (2:39:03.420)

The meaning of life.

Richard Haier (2:39:07.260)

We do the best we can to get through the day.

Lex Fridman (2:39:12.700)

And then there's just a finite number of the days.

Lex Fridman (2:39:16.140)

Are you afraid of the finiteness of it?

Lex Fridman (2:39:17.820)

I think about it more and more as I get older.

Richard Haier (2:39:21.140)

Yeah, I do.

Lex Fridman (2:39:23.940)

And it's one of these human things, that it is finite, we all know it.

Richard Haier (2:39:30.900)

Most of us deny it and don't want to think about it.

Richard Haier (2:39:35.980)

Sometimes you think about it in terms of estate planning, you try to do the rational thing.

Richard Haier (2:39:42.460)

Sometimes it makes you work harder because you know your time is more and more limited

Lex Fridman (2:39:46.900)

and you want to get things done.

Richard Haier (2:39:50.740)

I don't know where I am on that.

Lex Fridman (2:39:53.700)

It is just one of those things that's always in the back of my mind.

Lex Fridman (2:40:00.860)

And I don't think that's uncommon.

Lex Fridman (2:40:02.540)

Well it's just like G factor and intelligence, it's a hard truth that's there.

Lex Fridman (2:40:09.420)

And sometimes you kind of walk past it and you don't want to look at it, but it's still

Lex Fridman (2:40:15.100)

there.

Richard Haier (2:40:16.100)

Yeah.

Lex Fridman (2:40:17.100)

Yes, you can't escape it.

Lex Fridman (2:40:20.180)

And the thing about the G factor and intelligence is everybody knows this is true on a personal

Lex Fridman (2:40:28.220)

daily basis.

Richard Haier (2:40:31.020)

Even if you think back to when you were in school, you know who the smart kids were.

Richard Haier (2:40:39.060)

When you are on the phone talking to a customer service representative, that in response to

Richard Haier (2:40:44.820)

your detailed question is reading a script back to you and you get furious at this.

Lex Fridman (2:40:52.860)

Have you ever called this person a moron or wanted to call this person a moron?

Richard Haier (2:40:56.920)

You're not listening to me.

Richard Haier (2:40:58.820)

Everybody has had the experience of dealing with people who they think are not at their

Richard Haier (2:41:03.380)

level.

Lex Fridman (2:41:05.260)

It's just common because that's the way human beings are.

Richard Haier (2:41:09.460)

That's the way life is.

Lex Fridman (2:41:11.140)

But we also have a poor estimation of our own intelligence.

Richard Haier (2:41:16.380)

We have a poor, and we're not always a great, our judgment of human character of other people

Lex Fridman (2:41:22.980)

is not as good as a battery of tests.

Richard Haier (2:41:29.100)

That's where bias comes in.

Lex Fridman (2:41:31.780)

That's where our history, our emotions, all of that comes in.

Richard Haier (2:41:35.660)

So, you know, people on the internet, you know, there's such a thing as the internet

Lex Fridman (2:41:39.860)

and people on the internet will call each other dumb all the time.

Richard Haier (2:41:45.940)

You know, that's the worry here is that we give up on people.

Richard Haier (2:41:53.060)

We put them in a bin just because of one interaction or some small number of interactions as if

Richard Haier (2:42:00.140)

that's it.

Lex Fridman (2:42:01.140)

They're hopeless.

Richard Haier (2:42:02.140)

That's just in their genetics.

Lex Fridman (2:42:03.640)

But I think no matter what the science here says, once again, that does not mean we should

Richard Haier (2:42:11.140)

not have compassion for our fellow man.

Lex Fridman (2:42:15.020)

That's exactly what the science does say.

Richard Haier (2:42:17.700)

It's not opposite of what the science says.

Richard Haier (2:42:22.540)

Everything I know about psychology, everything I've learned about intelligence, everything

Richard Haier (2:42:30.120)

points to the inexorable conclusion that you have to treat people as individuals respectfully

Lex Fridman (2:42:38.220)

and with compassion.

Richard Haier (2:42:40.060)

Because through no fault of their own, some people are not as capable as others.

Lex Fridman (2:42:46.780)

And you want to turn a blind eye to it, you want to come up with theories about why that

Richard Haier (2:42:52.220)

might be true, fine.

Lex Fridman (2:42:54.300)

I would like to fix some of it as best I can.

Lex Fridman (2:42:58.060)

And everybody is deserving of love.

Lex Fridman (2:43:01.860)

Richard, this is a good way to end it, I think.

Richard Haier (2:43:05.340)

I'm just getting warmed up here.

Lex Fridman (2:43:07.500)

I know.

Richard Haier (2:43:08.500)

I know you can go for another many hours, but to respect your extremely valuable time,

Lex Fridman (2:43:15.220)

this is an amazing conversation.

Richard Haier (2:43:16.780)

Thank you for the teaching company, the lectures you've given with the New York Science of

Lex Fridman (2:43:23.900)

Intelligence.

Richard Haier (2:43:24.900)

Thank you for everything you're doing, it's a difficult topic, it's a topic that's controversial

Lex Fridman (2:43:29.700)

and sensitive to people and to push forward boldly and in that nuanced way, just thank

Richard Haier (2:43:35.820)

you for everything you do.

Lex Fridman (2:43:37.060)

And thank you for asking the big questions of intelligence, of consciousness.

Richard Haier (2:43:42.260)

Well thank you for asking me.

Lex Fridman (2:43:43.500)

I mean, there's nothing like good conversation on these topics.

Richard Haier (2:43:47.660)

Thanks for listening to this conversation with Richard Haier.

Lex Fridman (2:43:50.100)

To support this podcast, please check out our sponsors in the description.

Lex Fridman (2:43:54.540)

And now, let me leave you with some words from Albert Einstein.

Lex Fridman (2:43:57.620)

It is not that I'm so smart, but I stay with the questions much longer.

Richard Haier (2:44:04.460)

Thank you for listening and hope to see you next time.

Lex Fridman (30:03.300)

how do you figure?

Richard Haier (30:04.580)

Right, so now there's good research that gives you

Lex Fridman (30:08.980)

intuition about what kind of questions get at it,

Richard Haier (30:11.940)

just like there's something I've done,

Lex Fridman (30:18.100)

I've actually used for research in semi autonomous vehicle,

Richard Haier (30:21.980)

like whether humans are paying attention,

Lex Fridman (30:24.360)

there's a body of literature that does end back test,

Richard Haier (30:28.460)

for example, we have to put workload on the brain

Lex Fridman (30:35.620)

to do recall, memory recall, and that helps you

Richard Haier (30:38.900)

kind of put some work onto the brain

Lex Fridman (30:42.080)

while the person is doing some other task,

Lex Fridman (30:44.260)

and does some interesting research with that.

Lex Fridman (30:47.700)

But that's loading the memory,

Lex Fridman (30:48.940)

so there's like research around stably

Lex Fridman (30:52.260)

what that means about the human mind,

Lex Fridman (30:54.100)

and here you're saying recall backwards

Lex Fridman (30:58.420)

is a good protector.

Richard Haier (31:00.020)

It's a transformation.

Lex Fridman (31:01.740)

Yeah, so you have to do some,

Richard Haier (31:05.700)

like you have to load that into your brain,

Lex Fridman (31:07.820)

and not just remember it, but do something with it.

Richard Haier (31:11.220)

Right, here's another example of a different kind of test

Lex Fridman (31:14.460)

called the Hick paradigm, and it's not verbal at all.

Richard Haier (31:18.360)

It's a little box, and there are a series of lights

Lex Fridman (31:23.140)

arranged in a semi circle at the top of the box,

Lex Fridman (31:27.420)

and then there's a home button that you press,

Lex Fridman (31:31.300)

and when one of the lights goes on,

Richard Haier (31:34.740)

there's a button next to each of those lights,

Lex Fridman (31:37.780)

you take your finger off the home button,

Lex Fridman (31:39.940)

and you just press the button

Lex Fridman (31:41.980)

next to the light that goes on,

Lex Fridman (31:44.100)

and so it's a very simple reaction time.

Lex Fridman (31:46.500)

Light goes on, as quick as you can, you press the button,

Lex Fridman (31:49.220)

and you get a reaction time

Lex Fridman (31:50.540)

from the moment you lift your finger off the button

Richard Haier (31:53.700)

to when you press the button where the light is.

Lex Fridman (31:58.620)

That reaction time doesn't really correlate

Richard Haier (32:02.160)

with IQ very much, but if you change the instructions,

Lex Fridman (32:07.160)

and you say three lights are gonna come on simultaneously,

Richard Haier (32:13.120)

I want you to press the button next to the light

Lex Fridman (32:15.600)

that's furthest from the other two.

Lex Fridman (32:19.000)

So maybe lights one and two go on,

Lex Fridman (32:21.260)

and light six goes on simultaneously.

Richard Haier (32:24.320)

You take your finger off,

Lex Fridman (32:25.560)

and you would press the button by light six.

Richard Haier (32:28.880)

That's that reaction time to a more complex task.

Lex Fridman (32:34.360)

It's not really hard.

Richard Haier (32:36.360)

Almost everybody gets it all right,

Lex Fridman (32:38.360)

but your reaction time to that

Richard Haier (32:41.040)

is highly correlated with the G factor.

Lex Fridman (32:43.760)

This is fascinating.

Lex Fridman (32:45.040)

So reaction time, so there's a temporal aspect to this.

Lex Fridman (32:48.560)

So what role does time?

Richard Haier (32:50.040)

Speed of processing.

Lex Fridman (32:50.880)

It's the speed of processing.

Richard Haier (32:53.020)

Is this also true for ones that take longer,

Lex Fridman (32:55.740)

like five, 10, 30 seconds?

Lex Fridman (32:58.640)

Is time part of the measure with some of these things?

Lex Fridman (33:01.200)

Yes, and that is why some of the best IQ tests

Richard Haier (33:05.640)

have a time limit, because if you have no time limit,

Lex Fridman (33:10.640)

people can do better,

Lex Fridman (33:12.960)

but it doesn't distinguish among people that well.

Lex Fridman (33:17.760)

So that adding the time element is important.

Lex Fridman (33:21.440)

So speed of information processing,

Lex Fridman (33:25.020)

and reaction time is a measure

Richard Haier (33:26.680)

of speed of information processing,

Lex Fridman (33:29.320)

turns out to be related to the G factor.

Lex Fridman (33:31.920)

But the G factor only accounts for maybe half

Lex Fridman (33:35.040)

or some amount on the test performance.

Richard Haier (33:37.540)

For example, I get pretty bad test anxiety.

Lex Fridman (33:42.040)

Like I was never, I mean,

Richard Haier (33:46.280)

I just don't enjoy tests.

Lex Fridman (33:47.800)

I enjoy going back into my cave and working.

Richard Haier (33:51.280)

Like I've always enjoyed homework way more than tests,

Lex Fridman (33:56.080)

no matter how hard the homework is,

Richard Haier (33:57.800)

because I can go back to the cave

Lex Fridman (33:59.360)

and hide away and think deeply.

Richard Haier (34:00.760)

There's something about being watched

Lex Fridman (34:02.700)

and having a time limit that really makes me anxious,

Lex Fridman (34:06.040)

and I can just see the mind not operating optimally at all.

Lex Fridman (34:10.200)

But you're saying underneath there,

Richard Haier (34:11.640)

there's still a G factor, there's still.

Lex Fridman (34:13.720)

No question, there's no question.

Richard Haier (34:16.400)

Boy.

Lex Fridman (34:17.240)

And if you get anxious taking the test,

Richard Haier (34:19.280)

many people say, oh, I didn't do well,

Lex Fridman (34:20.880)

because I'm anxious.

Richard Haier (34:23.560)

I hear that a lot.

Lex Fridman (34:24.760)

Say, well, fine, if you're really anxious during the test,

Richard Haier (34:28.360)

the score will be a bad estimate of your G factor.

Lex Fridman (34:32.040)

It doesn't mean the G factor isn't there.

Richard Haier (34:34.040)

That's right.

Lex Fridman (34:34.880)

And by the way, standardized tests like the SAT,

Richard Haier (34:40.600)

they're essentially intelligence tests.

Lex Fridman (34:43.060)

They are highly G loaded.

Richard Haier (34:45.240)

Now, the people who make the SAT don't wanna mention that.

Lex Fridman (34:50.680)

They have enough trouble justifying standardized testing,

Lex Fridman (34:54.020)

but to call it an intelligence test

Lex Fridman (34:56.020)

is really beyond the pale.

Lex Fridman (34:58.440)

But in fact, it's so highly correlated,

Lex Fridman (35:00.720)

because it's a reasoning test.

Richard Haier (35:03.200)

SAT is a reasoning test,

Lex Fridman (35:04.880)

a verbal reasoning, mathematical reasoning.

Lex Fridman (35:08.240)

And if it's a reasoning test, it has to be related to G.

Lex Fridman (35:14.040)

But if people go in and take a standardized test,

Richard Haier (35:17.560)

whether it's an IQ test or the SAT,

Lex Fridman (35:20.120)

and they happen to be sick that day with 102 fever,

Richard Haier (35:24.600)

the score is not going to be a good estimate of their G.

Lex Fridman (35:29.600)

If they retake the test when they're not anxious

Richard Haier (35:33.080)

or less anxious or don't have a fever,

Lex Fridman (35:36.640)

the score will go up, and that will be a better estimate.

Lex Fridman (35:39.960)

But you can't say their G factor increased

Lex Fridman (35:43.120)

between the two tests.

Richard Haier (35:45.160)

Well, it's interesting.

Lex Fridman (35:46.480)

So the question is how wide of a battery of tests

Lex Fridman (35:50.000)

is required to estimate the G factor well?

Lex Fridman (35:53.380)

Because I'll give you as my personal example,

Richard Haier (35:55.180)

I took the SAT in, I think it was called the ACT,

Lex Fridman (35:58.720)

where I was two, also, I took SAT many times.

Richard Haier (36:02.880)

Every single time, I got it perfect on math.

Lex Fridman (36:05.480)

And verbal, the time limit on the verbal

Richard Haier (36:08.760)

made me very anxious.

Lex Fridman (36:10.960)

I did not, I mean, part of it,

Richard Haier (36:12.480)

I didn't speak English very well.

Lex Fridman (36:14.200)

But honestly, it was like you're supposed to remember stuff,

Lex Fridman (36:17.440)

and I was so anxious.

Lex Fridman (36:18.760)

And as I'm reading, I'm sweating, I can't,

Richard Haier (36:21.640)

you know that feeling you have when you're reading a book

Lex Fridman (36:26.640)

and you just read a page and you know nothing

Richard Haier (36:30.100)

about what you've read because you zoned out.

Lex Fridman (36:32.600)

That's the same feeling of like, I can't, I have to,

Richard Haier (36:36.440)

you're like, nope, read and understand.

Lex Fridman (36:39.480)

And that anxiety is like, and you start seeing

Richard Haier (36:43.080)

like the typography versus the content of the words.

Lex Fridman (36:47.120)

Like that was, I don't, it's interesting

Richard Haier (36:50.300)

because I know that what they're measuring,

Lex Fridman (36:55.300)

I could see being correlated with something.

Lex Fridman (36:58.780)

But that anxiety or some aspect of the performance

Lex Fridman (37:04.660)

sure plays a factor.

Lex Fridman (37:07.020)

And I wonder how you sneak up in a stable way.

Lex Fridman (37:10.420)

I mean, this is a broader discussion

Richard Haier (37:11.900)

about like standardized testing, how you sneak up,

Lex Fridman (37:16.580)

how you get at the fact that I'm super anxious

Lex Fridman (37:19.860)

and still nevertheless measure some aspect

Lex Fridman (37:22.260)

of my ontology.

Richard Haier (37:23.100)

I wonder, I don't know.

Lex Fridman (37:24.860)

I don't know if you can say to that,

Richard Haier (37:26.640)

that time limit sure is a pain.

Lex Fridman (37:28.540)

Well, let me say this.

Richard Haier (37:30.540)

There are two ways to approach the very real problem

Lex Fridman (37:34.160)

that you say that some people just get anxious

Richard Haier (37:36.880)

or not good test takers.

Lex Fridman (37:38.940)

By the way, part of testing is you know the answer,

Richard Haier (37:45.620)

you can figure out the answer or you can't.

Lex Fridman (37:49.980)

If you don't know the answer, there are many reasons

Richard Haier (37:52.900)

you don't know the answer at that particular moment.

Lex Fridman (37:55.240)

You may have learned it once and forgotten it.

Richard Haier (37:58.460)

It may be on the tip of your tongue

Lex Fridman (38:00.580)

and you just can't get it

Richard Haier (38:01.860)

because you're anxious about the time limit.

Lex Fridman (38:03.820)

You may never have learned it.

Richard Haier (38:05.860)

You may have been exposed to it,

Lex Fridman (38:08.660)

but it was too complicated and you couldn't learn it.

Richard Haier (38:11.420)

I mean, there are all kinds of reasons here.

Lex Fridman (38:13.940)

But for an individual to interpret your scores

Richard Haier (38:18.820)

as an individual, whoever is interpreting the score

Lex Fridman (38:23.340)

has to take into account various things

Richard Haier (38:26.260)

that would affect your individual score.

Lex Fridman (38:29.220)

And that's why decisions about college admission

Richard Haier (38:32.740)

or anything else where tests are used

Lex Fridman (38:35.720)

are hardly ever the only criterion to make a decision.

Lex Fridman (38:42.340)

And I think people are, college admissions

Lex Fridman (38:45.160)

letting go of that very much.

Richard Haier (38:46.860)

Oh yes, yeah.

Lex Fridman (38:48.100)

But what does that even mean?

Richard Haier (38:51.060)

Because is it possible to design standardized tests

Lex Fridman (38:55.540)

that do get, that are useful to college admissions?

Richard Haier (38:58.420)

Well, they already exist.

Lex Fridman (38:59.740)

The SAT is highly correlated with many aspects

Richard Haier (39:03.820)

of success at college.

Lex Fridman (39:05.220)

Here's the problem.

Lex Fridman (39:06.380)

So maybe you could speak to this.

Lex Fridman (39:09.220)

The correlation across the population versus individuals.

Lex Fridman (39:13.360)

So our criminal justice system is designed to make sure,

Lex Fridman (39:23.420)

wow, it's still, there's tragic cases

Richard Haier (39:27.300)

where innocent people go to jail,

Lex Fridman (39:29.700)

but you try to avoid that.

Lex Fridman (39:31.340)

And the same way with testing,

Lex Fridman (39:34.460)

it just, it would suck for an SAT to miss genius.

Richard Haier (39:38.780)

Yes, and it's possible, but it's statistically unlikely.

Lex Fridman (39:43.260)

So it really comes down to which piece of information

Richard Haier (39:51.980)

maximizes your decision making ability.

Lex Fridman (39:58.640)

So if you just use high school grades, it's okay.

Lex Fridman (40:05.300)

But you will miss some people

Lex Fridman (40:07.100)

who just don't do well in high school,

Lex Fridman (40:09.140)

but who are actually pretty smart,

Lex Fridman (40:11.340)

smart enough to be bored silly in high school,

Lex Fridman (40:13.980)

and they don't care,

Lex Fridman (40:14.940)

and their high school GPA isn't that good.

Lex Fridman (40:17.780)

So you will miss them in the same sense

Lex Fridman (40:21.380)

that somebody who could be very able and ready for college

Richard Haier (40:25.980)

just doesn't do well on their SAT.

Lex Fridman (40:28.340)

This is why you make decisions

Richard Haier (40:31.660)

with taking in a variety of information.

Lex Fridman (40:36.140)

The other thing I wanted to say,

Richard Haier (40:38.080)

I talked about when you make a decision for an individual,

Lex Fridman (40:43.840)

statistically for groups,

Richard Haier (40:46.720)

there are many people who have a disparity

Lex Fridman (40:50.060)

between their math score and their verbal score.

Richard Haier (40:53.080)

That disparity, or the other way around,

Lex Fridman (40:55.640)

that disparity is called tilt.

Richard Haier (40:58.680)

The score is tilted one way or the other.

Lex Fridman (41:01.760)

And that tilt has been studied empirically

Richard Haier (41:05.080)

to see what that predicts.

Lex Fridman (41:07.320)

And in fact, you can't make predictions

Richard Haier (41:09.440)

about college success based on tilt.

Lex Fridman (41:14.880)

And mathematics is a good example.

Richard Haier (41:16.760)

There are many people,

Lex Fridman (41:18.320)

especially non native speakers of English

Richard Haier (41:20.680)

who come to this country,

Lex Fridman (41:22.360)

take the SATs, do very well on the math

Lex Fridman (41:24.940)

and not so well on the verbal.

Lex Fridman (41:26.840)

Well, if they're applying to a math program,

Richard Haier (41:31.220)

the professors there who are making the decision

Lex Fridman (41:33.840)

or the admissions officers

Richard Haier (41:35.400)

don't wait so much to score on verbal,

Lex Fridman (41:39.600)

especially if it's a non native speaker.

Richard Haier (41:42.080)

Well, so yeah, you have to try to,

Lex Fridman (41:44.480)

in the admission process, bring in the context.

Lex Fridman (41:47.800)

But non native isn't really the problem.

Lex Fridman (41:50.760)

I mean, that was part of the problem for me.

Lex Fridman (41:53.720)

But it's the anxiety was, which it's interesting.

Lex Fridman (41:57.980)

It's interesting.

Richard Haier (41:59.000)

Oh boy, reducing yourself down to numbers.

Lex Fridman (42:06.560)

But it's still true.

Richard Haier (42:07.880)

It's still the truth.

Lex Fridman (42:09.240)

It's a painful truth.

Richard Haier (42:10.760)

That same anxiety that led me to be,

Lex Fridman (42:16.080)

to struggle with the SAT verbal tests

Richard Haier (42:20.600)

is still within me in all ways of life.

Lex Fridman (42:24.640)

So maybe that's not anxiety.

Richard Haier (42:26.880)

Maybe that's something, like personality

Lex Fridman (42:30.600)

is also pretty stable.

Richard Haier (42:32.380)

Personality is stable.

Lex Fridman (42:34.480)

Personality does impact the way you navigate life.

Richard Haier (42:41.040)

There's no question.

Lex Fridman (42:42.460)

Yeah, and we should say that the G factor in intelligence

Richard Haier (42:45.680)

is not just about some kind of number on a paper.

Lex Fridman (42:50.400)

It also has to do with how you navigate life.

Lex Fridman (42:53.940)

How easy life is for you in this very complicated world.

Lex Fridman (43:00.880)

So personality's all tied into that

Richard Haier (43:02.880)

in some deep fundamental way.

Lex Fridman (43:05.880)

But now you've hit the key point

Richard Haier (43:07.720)

about why we even want to study intelligence.

Lex Fridman (43:11.320)

And personality, I think, to a lesser extent.

Lex Fridman (43:13.360)

But that's my interest, is more on intelligence.

Lex Fridman (43:17.480)

I went to graduate school and wanted to study personality,

Lex Fridman (43:20.140)

but that's kind of another story

Lex Fridman (43:22.640)

how I got kind of shifted from personality research

Richard Haier (43:25.240)

over to intelligence research.

Lex Fridman (43:27.440)

Because it's not just a number.

Richard Haier (43:30.040)

Intelligence is not just an IQ score.

Lex Fridman (43:32.560)

It's not just an SAT score.

Richard Haier (43:34.740)

It's what those numbers reflect about your ability

Lex Fridman (43:39.740)

to navigate everyday life.

Richard Haier (43:43.000)

It has been said that life is one long intelligence test.

Lex Fridman (43:48.000)

And who can't relate to that?

Lex Fridman (43:55.400)

And if you doubt, see, another problem here

Lex Fridman (43:58.440)

is a lot of critics of intelligence research,

Richard Haier (44:00.800)

intelligence testing, tend to be academics

Lex Fridman (44:04.000)

who, by and large, are pretty smart people.

Lex Fridman (44:07.240)

And pretty smart people, by and large,

Lex Fridman (44:10.040)

have enormous difficulty understanding

Lex Fridman (44:12.900)

what the world is like for people with IQs of 80 or 75.

Lex Fridman (44:18.960)

It is a completely different everyday experience.

Richard Haier (44:23.040)

Even IQ scores of 85, 90.

Lex Fridman (44:27.900)

You know, there's a popular television program, Judge Judy.

Richard Haier (44:32.040)

Judge Judy deals with everyday people

Lex Fridman (44:35.340)

with everyday problems, and you can see the full range

Richard Haier (44:39.440)

of problem solving ability demonstrated there.

Lex Fridman (44:43.240)

And sometimes she does it for laughs,

Lex Fridman (44:45.300)

but it really isn't funny because people who are,

Lex Fridman (44:52.460)

there are people who are very limited

Richard Haier (44:54.500)

in their life navigation, let alone success,

Lex Fridman (45:00.040)

by not having good reasoning skills, which cannot be taught.

Richard Haier (45:05.040)

We know this, by the way, because there are many efforts.

Lex Fridman (45:07.720)

You know, the United States military,

Richard Haier (45:09.320)

which excels at training people,

Lex Fridman (45:12.160)

I mean, I don't know that there's a better organization

Richard Haier (45:14.560)

in the world for training diverse people,

Lex Fridman (45:18.360)

and they won't take people with IQs under,

Richard Haier (45:20.760)

I think, 83 is the cutoff, because they have found

Lex Fridman (45:24.960)

they are unable to train people with lower IQs

Richard Haier (45:30.520)

to do jobs in the military.

Lex Fridman (45:32.520)

So one of the things that G Factor has to do is learning.

Richard Haier (45:35.720)

Absolutely, some people learn faster than others.

Lex Fridman (45:40.720)

Some people learn more than others.

Richard Haier (45:43.520)

Now, faster, by the way, is not necessarily better,

Lex Fridman (45:47.320)

as long as you get to the same place eventually.

Richard Haier (45:51.720)

But, you know, there are professional schools

Lex Fridman (45:54.320)

that want students who can learn the fastest

Richard Haier (45:57.640)

because they can learn more or learn better.

Lex Fridman (46:01.520)

Or learn deeper, or all kinds of ideas

Richard Haier (46:06.880)

about why you select people with the highest scores.

Lex Fridman (46:09.560)

And there's nothing funnier, by the way,

Richard Haier (46:12.640)

to listen to a bunch of academics

Lex Fridman (46:15.120)

complain about the concept of intelligence

Lex Fridman (46:17.680)

and intelligence testing, and then you go

Lex Fridman (46:20.120)

to a faculty meeting where they're discussing

Richard Haier (46:22.480)

who to hire among the applicants.

Lex Fridman (46:24.760)

And all they talk about is how smart the person is.

Richard Haier (46:28.400)

We'll get to that, we'll sneak up to that in different ways,

Lex Fridman (46:31.200)

but there's something about reducing a person

Richard Haier (46:33.000)

to a number that in part is grounded

Lex Fridman (46:35.280)

to the person's genetics that makes people very uncomfortable.

Lex Fridman (46:38.800)

But nobody does that.

Lex Fridman (46:40.480)

Nobody in the field actually does that.

Richard Haier (46:43.760)

That is a worry that is a worry like,

Lex Fridman (46:54.040)

well, I don't wanna call it a conspiracy theory.

Richard Haier (46:55.880)

I mean, it's a legitimate worry,

Lex Fridman (46:58.360)

but it just doesn't happen.

Richard Haier (47:01.400)

Now, I had a professor in graduate school

Lex Fridman (47:03.600)

who was the only person I ever knew

Richard Haier (47:05.880)

who considered the students only by their test scores.

Lex Fridman (47:12.680)

And later in his life, he kind of backed off that.

Richard Haier (47:19.480)

Let me ask you this, so we'll jump around,

Lex Fridman (47:21.560)

I'll come back to it, but I tend to,

Richard Haier (47:26.560)

I've had like political discussions with people

Lex Fridman (47:29.520)

and actually my friend Michael Malice, he's an anarchist.

Richard Haier (47:36.240)

I disagree with him on basically everything

Lex Fridman (47:39.200)

except the fact that love is a beautiful thing in this world.

Lex Fridman (47:47.360)

And he says this test about left versus right,

Lex Fridman (47:50.600)

whatever, it doesn't matter what the test is,

Lex Fridman (47:52.280)

but he believes, the question is,

Lex Fridman (47:54.920)

do you believe that some people are better than others?

Richard Haier (48:00.360)

Question is ambiguous.

Lex Fridman (48:03.880)

Do you believe some people are better than others?

Lex Fridman (48:06.120)

And to me, sort of the immediate answer is no.

Lex Fridman (48:11.720)

It's a poetic question, it's an ambiguous question, right?

Richard Haier (48:15.680)

Like people wanna maybe the temptation

Lex Fridman (48:19.520)

to ask better at what, better at like sports and so on.

Richard Haier (48:23.440)

No, to me, I stand with the sort of defining documents

Lex Fridman (48:28.440)

of this country, which is all men are created equal.

Richard Haier (48:32.280)

There's a basic humanity.

Lex Fridman (48:34.320)

And there's something about tests of intelligence.

Richard Haier (48:39.840)

Just knowing that some people are different,

Lex Fridman (48:43.380)

like the science of intelligence that shows

Richard Haier (48:45.400)

that some people are genetically

Lex Fridman (48:49.200)

in some stable way across a lifetime,

Richard Haier (48:52.400)

have a greater intelligence than others,

Lex Fridman (48:56.160)

makes people feel like some people are better than others.

Lex Fridman (49:01.200)

And that makes them very uncomfortable.

Lex Fridman (49:03.600)

And I, maybe you can speak to that.

Richard Haier (49:06.960)

The fact that some people are more intelligent than others

Lex Fridman (49:09.560)

in a way that's, cannot be compensated

Richard Haier (49:14.560)

through education, through anything you do in life.

Lex Fridman (49:22.600)

What do we do with that?

Richard Haier (49:24.480)

Okay, there's a lot there.

Lex Fridman (49:26.800)

We haven't really talked about the genetics of it yet.

Lex Fridman (49:29.960)

But you are correct in that it is my interpretation

Lex Fridman (49:35.240)

of the data that genetics has a very important influence

Richard Haier (49:39.720)

on the G factor.

Lex Fridman (49:41.380)

And this is controversial, and we can talk about it,

Lex Fridman (49:44.440)

but if you think that genetics,

Lex Fridman (49:47.040)

that genes are deterministic, are always deterministic,

Richard Haier (49:50.800)

that leads to kind of the worry that you expressed.

Lex Fridman (49:55.320)

But we know now in the 21st century

Richard Haier (49:58.600)

that many genes are not deterministic,

Lex Fridman (50:00.920)

that are probabilistic,

Richard Haier (50:02.800)

meaning their gene expression can be influenced.

Lex Fridman (50:09.880)

Now, whether they're influenced only

Richard Haier (50:11.320)

by other biological variables or other genetic variables

Lex Fridman (50:16.440)

or environmental or cultural variables,

Richard Haier (50:19.160)

that's where the controversy comes in.

Lex Fridman (50:23.040)

And we can discuss that in more detail if you like.

Lex Fridman (50:27.200)

But to go to the question about better, are people better?

Lex Fridman (50:31.700)

There's zero evidence that smart people are better

Richard Haier (50:36.700)

with respect to important aspects of life,

Lex Fridman (50:43.300)

like honesty, even likability.

Richard Haier (50:47.460)

I'm sure you know many very intelligent people

Lex Fridman (50:50.340)

who are not terribly likable or terribly kind

Richard Haier (50:53.420)

or terribly honest.

Lex Fridman (50:55.480)

Is there something to be said?

Lex Fridman (50:56.760)

So one of the things I've recently reread

Lex Fridman (50:59.820)

for the second time, I guess that's what the word reread

Richard Haier (51:03.860)

means, the rise and fall of the Third Reich,

Lex Fridman (51:08.980)

which is, I think, the best telling

Richard Haier (51:12.060)

of the rise and fall of Hitler.

Lex Fridman (51:14.700)

And one of the interesting things about the people that,

Lex Fridman (51:20.900)

how should I say it?

Lex Fridman (51:27.260)

Justified or maybe propped up the ideas

Richard Haier (51:32.260)

that Hitler put forward is the fact

Lex Fridman (51:35.940)

that they were extremely intelligent.

Richard Haier (51:38.460)

They were the intellectual class.

Lex Fridman (51:41.780)

They were like, it was obvious that they thought

Richard Haier (51:46.260)

very deeply and rationally about the world.

Lex Fridman (51:49.540)

So what I would like to say is one of the things

Richard Haier (51:52.660)

that shows to me is some of the worst atrocities

Lex Fridman (51:56.100)

in the history of humanity have been committed

Richard Haier (51:58.860)

by very intelligent people.

Lex Fridman (52:00.780)

So that means that intelligence

Richard Haier (52:04.620)

doesn't make you a good person.

Lex Fridman (52:06.420)

I wonder if there's a G factor for intelligence.

Richard Haier (52:12.820)

I wonder if there's a G factor for goodness.

Lex Fridman (52:16.860)

The Nietzschean good and evil,

Richard Haier (52:19.180)

of course that's probably harder to measure

Lex Fridman (52:21.740)

because it's such a subjective thing

Lex Fridman (52:23.240)

what it means to be good.

Lex Fridman (52:25.340)

And even the idea of evil is a deeply uncomfortable thing

Lex Fridman (52:29.380)

because how do we know?

Lex Fridman (52:31.340)

But it's independent, whatever it is,

Richard Haier (52:33.500)

it's independent of intelligence.

Lex Fridman (52:35.900)

So I agree with you about that.

Lex Fridman (52:37.980)

But let me say this.

Lex Fridman (52:39.340)

I have also asserted my belief

Richard Haier (52:44.100)

that more intelligence is better than less.

Lex Fridman (52:49.220)

That doesn't mean more intelligent people are better people

Lex Fridman (52:54.400)

but all things being equal,

Lex Fridman (52:55.860)

would you like to be smarter or less smart?

Lex Fridman (52:58.740)

So if I had a pill, I have two pills.

Lex Fridman (53:01.260)

I said, this one will make you smarter,

Richard Haier (53:02.880)

this one will make you dumber.

Lex Fridman (53:04.500)

Which one would you like?

Richard Haier (53:06.460)

Are there any circumstances

Lex Fridman (53:07.940)

under which you would choose to be dumber?

Richard Haier (53:09.860)

Well, let me ask you this.

Lex Fridman (53:11.560)

That's a very nuanced and interesting question.

Lex Fridman (53:16.040)

There's been books written about this, right?

Lex Fridman (53:19.680)

Now we'll return to the hard questions,

Richard Haier (53:21.680)

the interesting questions,

Lex Fridman (53:22.760)

but let me ask about human happiness.

Lex Fridman (53:25.780)

Does intelligence lead to happiness?

Lex Fridman (53:29.180)

No.

Richard Haier (53:31.820)

So, okay, so back to the pill then.

Lex Fridman (53:34.980)

So when would you take the pill?

Lex Fridman (53:38.980)

So you said IQ 80, 90, 100, 110,

Lex Fridman (53:44.860)

you start going through the quartiles

Lex Fridman (53:46.580)

and is it obvious?

Lex Fridman (53:50.220)

Isn't there diminishing returns

Lex Fridman (53:54.500)

and then it starts becoming negative?

Lex Fridman (53:57.420)

This is an empirical question.

Lex Fridman (54:00.420)

And so that I have advocated in many forums

Lex Fridman (54:06.160)

more research on enhancing the G factor.

Richard Haier (54:11.820)

Right now there have been many claims

Lex Fridman (54:14.420)

about enhancing intelligence with,

Richard Haier (54:17.320)

you mentioned the NBAC training,

Lex Fridman (54:19.100)

it was a big deal a few years ago, it doesn't work.

Richard Haier (54:22.640)

Data is very clear, it does not work.

Lex Fridman (54:25.740)

Or doing like memory tests, like training and so on.

Richard Haier (54:28.580)

Yeah, it may give you a better memory in the short run,

Lex Fridman (54:32.700)

but it doesn't impact your G factor.

Richard Haier (54:38.140)

It was very popular a couple of decades ago

Lex Fridman (54:40.940)

that the idea that listening to Mozart

Richard Haier (54:44.560)

could make you more intelligent.

Lex Fridman (54:46.580)

There was a paper published on this

Richard Haier (54:48.180)

with somebody I knew published this paper,

Lex Fridman (54:50.300)

and intelligence researchers never believed it for a second.

Richard Haier (54:54.660)

Been hundreds of studies, all the meta analyses,

Lex Fridman (54:57.820)

all the summaries and so on,

Richard Haier (54:59.420)

show that there's nothing to it, nothing to it at all.

Lex Fridman (55:05.900)

But wouldn't it be something,

Richard Haier (55:08.780)

wouldn't it be world shaking

Lex Fridman (55:12.820)

if you could take the normal distribution of intelligence,

Richard Haier (55:15.940)

which we haven't really talked about yet,

Lex Fridman (55:17.500)

but IQ scores and the G factor

Richard Haier (55:20.400)

is thought to be a normal distribution,

Lex Fridman (55:23.940)

and shift it to the right so that everybody is smarter?

Richard Haier (55:30.140)

Even a half a standard deviation would be world shaking,

Lex Fridman (55:35.340)

because there are many social problems,

Richard Haier (55:38.700)

many, many social problems that are exacerbated

Lex Fridman (55:43.260)

by people with lower ability to reason stuff out

Lex Fridman (55:48.700)

and navigate everyday life.

Lex Fridman (55:51.860)

So.

Richard Haier (55:52.700)

I wonder if there's a threshold.

Lex Fridman (55:53.780)

So maybe I would push back and say universal shifting

Richard Haier (55:59.420)

of the normal distribution

Lex Fridman (56:02.240)

may not be the optimal way of shifting.

Richard Haier (56:05.060)

Maybe it's better to,

Lex Fridman (56:07.340)

whatever the asymmetric kind of distributions

Richard Haier (56:10.620)

is like really pushing the lower up

Lex Fridman (56:13.380)

versus trying to make the people

Richard Haier (56:17.580)

at the average more intelligent.

Lex Fridman (56:19.660)

So you're saying that if in fact

Richard Haier (56:21.060)

there was some way to increase G,

Lex Fridman (56:23.540)

let's just call it metaphorically a pill, an IQ pill,

Richard Haier (56:27.800)

we should only give it to people at the lower end.

Lex Fridman (56:30.580)

No, it's just intuitively I can see

Richard Haier (56:34.940)

that life becomes easier at the lower end if it's increased.

Lex Fridman (56:39.820)

It becomes less and less,

Richard Haier (56:41.820)

it is an empirical scientific question,

Lex Fridman (56:43.580)

but it becomes less and less obvious to me

Richard Haier (56:46.180)

that more intelligence is better.

Lex Fridman (56:50.500)

At the high end, not because it would make life easier,

Lex Fridman (56:56.500)

but it would make whatever problems you're working on

Lex Fridman (57:00.980)

more solvable.

Lex Fridman (57:02.700)

And if you are working on artificial intelligence,

Lex Fridman (57:06.740)

there's a tremendous potential for that to improve society.

Richard Haier (57:13.300)

I understand.

Lex Fridman (57:14.580)

So at whatever problems you're working on, yes.

Lex Fridman (57:19.000)

But there's also the problem of the human condition.

Lex Fridman (57:21.720)

There's love, there's fear,

Lex Fridman (57:24.340)

and all of those beautiful things

Lex Fridman (57:26.700)

that sometimes if you're good at solving problems,

Richard Haier (57:29.860)

you're going to create more problems for yourself.

Lex Fridman (57:32.100)

It's, I'm not exactly sure.

Lex Fridman (57:34.460)

So ignorance is bliss is a thing.

Lex Fridman (57:37.220)

So there might be a place,

Richard Haier (57:38.400)

there might be a sweet spot of intelligence

Lex Fridman (57:40.960)

given your environment, given your personality,

Richard Haier (57:43.680)

all of those kinds of things.

Lex Fridman (57:45.060)

And that becomes less beautifully complicated

Richard Haier (57:48.140)

the more and more intelligent you become.

Lex Fridman (57:50.500)

But that's a question for literature,

Richard Haier (57:53.160)

not for science perhaps.

Lex Fridman (57:54.700)

Well, imagine this.

Richard Haier (57:56.180)

Imagine there was an IQ pill

Lex Fridman (57:58.380)

and it was developed by a private company

Lex Fridman (58:01.580)

and they are willing to sell it to you.

Lex Fridman (58:05.060)

And whatever price they put on it,

Richard Haier (58:07.900)

you are willing to pay it

Lex Fridman (58:09.500)

because you would like to be smarter.

Lex Fridman (58:11.780)

But just before they give you a pill,

Lex Fridman (58:14.360)

they give you a disclaimer form to sign.

Richard Haier (58:18.380)

Yes.

Lex Fridman (58:20.020)

Don't hold us,

Richard Haier (58:22.620)

you understand that this pill has no guarantee

Lex Fridman (58:25.180)

that your life is going to be better

Lex Fridman (58:26.980)

and in fact it could be worse.

Lex Fridman (58:28.920)

Well, yes, that's how lawyers work.

Lex Fridman (58:32.200)

But I would love for science to answer the question

Lex Fridman (58:35.160)

to try to predict if your life

Richard Haier (58:36.640)

is going to be better or worse

Lex Fridman (58:38.680)

when you become more or less intelligent.

Richard Haier (58:41.200)

It's a fascinating question

Lex Fridman (58:43.260)

about what is the sweet spot for the human condition.

Richard Haier (58:47.600)

Some of the things we see as bugs

Lex Fridman (58:49.820)

might be actually features,

Richard Haier (58:51.900)

may be crucial to our overall happiness

Lex Fridman (58:55.360)

as our limitations might lead to more happiness than less.

Lex Fridman (58:59.120)

But again, more intelligence is better at the lower end.

Lex Fridman (59:02.720)

That's more, that's something that's less arguable

Lex Fridman (59:06.480)

and fascinating if possible to increase.

Lex Fridman (59:10.080)

But you know, there's virtually no research

Richard Haier (59:12.960)

that's based on a neuroscience approach

Lex Fridman (59:15.320)

to solving that problem.

Richard Haier (59:17.520)

All the solutions that have been proposed

Lex Fridman (59:20.800)

to solve that problem or to ameliorate that problem

Richard Haier (59:25.400)

are essentially based on the blank slate assumption

Lex Fridman (59:29.680)

that enriching the environment, removing barriers,

Richard Haier (59:34.680)

all good things by the way,

Lex Fridman (59:36.000)

I'm not against any of those things.

Lex Fridman (59:38.160)

But there's no empirical evidence

Lex Fridman (59:39.700)

that they're going to improve the general reasoning ability

Richard Haier (59:45.260)

or make people more employable.

Lex Fridman (59:47.920)

Have you read Flowers of Algernon? Yes.

Richard Haier (59:51.840)

That's to the question of intelligence and happiness.

Lex Fridman (59:56.240)

There are many profound aspects of that story.

Richard Haier (59:59.440)

It was a film that was very good.