Marc Andreessen: How Risk Taking, Innovation & Artificial Intelligence Transform Human Experience

Andrew Huberman: [MUSIC PLAYING] Welcome to the Huberman Lab

podcast, where we discuss science and science-based tools for everyday life.

I'm Andrew Huberman, and I'm a professor of neurobiology and ophthalmology

at Stanford School of Medicine.

Today, my guest is Marc Andreessen.

Marc Andreessen is a software engineer and an investor in technology companies.

He co-founded and developed Mosaic, which was one of the

first widely used web browsers.

He also co-founded and developed Netscape, which was one of the

earliest widespread used web browsers.

And he co-founded and is a general partner at Andreessen Horowitz,

one of the most successful Silicon Valley venture capital firms.

All of that is to say that Mark Andreessen is one of the most successful

innovators and investors ever.

I was extremely excited to record this episode with Marc for several reasons.

First of all, he himself is an incredible innovator.

Second of all, he has an uncanny ability to spot the innovators of the future.

And third, Marc has shown over and over again the ability to understand

how technologies not yet even developed are going to impact the

way that humans interact at large.

Our conversation starts off by discussing what makes for an exceptional innovator,

as well as what sorts of environmental conditions make for exceptional

innovation and creativity more generally.

In that context, we talk about risk taking, not just in terms of risk taking

in one's profession, but about how some people, not all, but how some people who

are risk takers and innovators in the context of their work also seem to take

a lot of risks in their personal life and some of the consequences that can bring.

Then we discuss some of the most transformative technologies that are

now emerging, such as novel approaches to developing clean energy, as well

as AI or artificial intelligence.

With respect to AI, Marc shares his views as to why AI is likely to

greatly improve human experience, and we discuss the multiple roles

that AI is very likely to have in all of our lives in the near future.

Marc explains how not too long from now, all of us are very likely to have AI

assistants, for instance, assistants that give us highly informed health advice,

highly informed psychological advice.

Indeed, it is very likely that all of us will soon have AI assistants that govern

most, if not all, of our daily decisions.

And Marc explains how, if done correctly, this can be a tremendously

positive addition to our life.

In doing so, Marc provides a stark argument for those that argue that AI

is going to diminish human experience.

So if you're hearing about and or concerned about the ways that AI is

likely to destroy us today, you are going to hear about the many different

ways that AI technologies now in development are likely to enhance

our human experience at every level.

What you'll soon find is that while today's discussion does center around

technology and technology development, it is really a discussion about

human beings and human psychology.

So whether you have an interest in technology development and or AI,

I'm certain that you'll find today's discussion to be an important and

highly lucid view into what will soon be the future that we all live in.

Before we begin, I'd like to emphasize that this podcast is separate from my

teaching and research roles at Stanford.

It is, however, part of my desire a nd effort to bring zero cost

to consumer information about science and science-related

tools to the general public.

In keeping with that theme, I'd like to thank the sponsors of today's podcast.

Our first sponsor is LMNT.

LMNT is an electrolyte drink that has everything you need and nothing you don't.

That means plenty of the electrolytes, sodium, magnesium, and potassium in

the correct ratios, but no sugar.

The electrolytes and hydration are absolutely key for mental health,

physical health, and performance.

Even a slight degree of dehydration can impair our ability to think, our energy

levels and our physical performance.

LMNT makes it very easy to achieve proper hydration, and it does so by

including the three electrolytes in the exact ratios they need to be present.

I drink LMNT first thing in the morning when I wake up.

I usually mix it with about 16 to 32oz of water.

If I'm exercising, I'll drink one while I'm exercising, and I tend to

drink one after exercising as well.

Now, many people are scared off by the idea of ingesting sodium because obviously

we don't want to consume sodium in excess.

However, for people that have normal blood pressure, and especially for

people that are consuming very clean diets, that is consuming not so many

processed foods or highly processed foods, oftentimes we are not getting

enough sodium, magnesium and potassium, and we can suffer as a consequence.

And with LMNT , simply by mixing in water, it tastes delicious.

It's very easy to get that proper hydration.

If you'd like to try LMNT , you can go to drinklmnt, that's L-M-N-T,

.com/huberman to claim a free element sample pack with your purchase.

Again, that's drinklmnt.com/huberman.

Today's episode is also brought to us by Eight Sleep.

Eight Sleep makes smart mattress covers with cooling, heating

and sleep tracking capacity.

I've spoken many times before on this podcast about the fact that sleep, that

is getting a great night's sleep, is the foundation of all mental health,

physical health and performance.

When we're sleeping well, everything goes far better.

And when we are not sleeping well or enough, everything gets far

worse at the level of mental health, physical health and performance.

Now, one of the key things to getting a great night's sleep and waking up feeling

refreshed is that you have to control the temperature of your sleeping environment.

And that's because in order to fall and stay deeply asleep, you

need your core body temperature to drop by about one to three degrees.

And in order to wake up feeling refreshed and energized, you want

your core body temperature to increase by about one to three degrees.

With Eight Sleep , it's very easy to induce that drop in core body

temperature by cooling your mattress early and throughout the night and

warming your mattress toward morning.

I started sleeping on an Eight Sleep mattress cover a few years ago, and

it has completely transformed the quality of the sleep that I get.

So much so that I actually loathe traveling because I don't have my Eight

Sleep mattress cover when I travel.

If you'd like to try Eight Sleep , you can go to eightsleep.com/huberman and you'll

save up to $150 off their Pod 3 Cover.

Eight Sleep currently ships in the USA, Canada, UK, select

countries in t he EU and Australia.

Again, that's eightsleep.com/huberman.

And now for my discussion with Marc Andreessen.

Marc, welcome.

Marc Andreessen: Hey, thank you.

Andrew Huberman: Delighted to have you here and have so many

questions for you about innovation AI, your view of the landscape

of tech, and humanity in general.

I want to start off by talking about innovation from three

different perspectives.

There's the inner game, so to speak, or the psychology of the innovator,

or innovators, things like their propensity for engaging in conflict

or not, their propensity for having a dream or a vision, and in particular,

their innovation as it relates to some psychological trait or expression.

So we'll get to that in a moment.

The second component that I'm curious about is the outer landscape

around innovators, who they place themselves with, the sorts of

choices that they make and also the sorts of personal relationships

that they might have or not have.

And then the last component is this notion of the larger landscape that

they happen to find themselves in.

What time in history?

What's the geography?

Bay Area, New York, Dubai, etc.

So to start off, is there a common trait of innovators that you think

is absolutely essential as a seed to creating things that are really impactful?

Marc Andreessen: Yeah.

So I'm not a psychologist, but I've picked up some of the

concepts and some of the terms.

And so it was a great moment of delight in my life when I learned about the Big

Five personality traits, because I was like, aha, there's a way to actually

describe the answer to this question in at least reasonably scientific terms.

And so I think what you're looking for, when you're talking about real

innovators, like people who actually do really creative breakthrough work, I think

you're talking about a couple of things.

So one is very high in what's called trait openness, which is one of the

Big Five, which is basically just like, flat out open to new ideas.

And of course, the nature of trait openness is trait openness means

you're not just open to new ideas in one category, you're open to

many different kinds of new ideas.

And so we might talk about the fact that a lot of innovators also

are very creative people in other aspects of their lives, even outside

of their specific creative domain.

So that's important.

But of course, just being open is not sufficient, because if you're just

open, you could just be curious and explore and spend your entire life

reading and doing, talking to people and never actually create something.

So you also need a couple of other things.

You need a high level of conscientiousness, which is

another one of the Big Five.

You need somebody who's really willing to apply themselves, and in our world,

typically over a period of many years to be able to accomplish something great.

They typically work very hard.

That often gets obscured because the stories that end up getting told

about these people are, it's just like this kid, and he just had this idea,

and it was like a stroke of genius.

And it was like a moment in time and was just like, oh, he was so lucky.

And it's like, no, for most of these people, it's years and

years and years of applied effort.

And so you need somebody with an extreme, basically, willingness to defer

gratification and really apply themselves to a specific thing for a long time.

And of course, this is why there aren't very many of these people, there aren't

many people who are high in openness and high in conscientiousness because to a

certain extent, they're opposed traits.

And so you need somebody who has both of those.

Third is you need somebody high in disagreeableness, which

is the third of the Big Five.

So you need somebody who's just basically ornery, because if they're not ornery,

then they'll be talked out of their ideas by people who will be like, oh,

well, because the reaction, most people have new ideas is, oh, that's dumb.

And so somebody who's too agreeable will be easily dissuaded to not

pursue, not pulling the thread anymore.

So you need somebody highly disagreeable.

Again, the nature of disagreeableness is they tend to

be disagreeable about everything.

So they tend to be these very sort of iconoclastic kind of renegade characters.

And then there's just a table stakes component, which is they

just also need to be high IQ.

They just need to be really smart because it's hard to innovate in

any category if you can't synthesize large amounts of information quickly.

And so those are four basically high spikes, very rare traits that

basically have to come together.

You could probably also say they probably at some point need to be relatively low

on neuroticism, which is another of the Big Five, because if they're too neurotic,

they probably can't handle the stress.

Right. So it's kind of this dial in there.

And then, of course, if you're into the sort of science of the Big Five, basically

these are all people who are on the far outlying kind of point on the normal

distribution across all these traits.

And then that just gets you to, I think, the sort of hardest topic of

all around this whole concept, which there are very few of these people.

Andrew Huberman: Do you think they're born with these traits?

Marc Andreessen: Yeah, they're born with the traits.

And then, of course, the traits are not genetics, are not destiny, and so

the traits are not deterministic in the sense of that just because they

have those personality traits doesn't mean they're going to deliver great

creativity, but they need to have those properties because otherwise they're just

not either going to be able to do the work or they're not going to enjoy it.

Right.

I mean, look, a lot of these people are highly capable, competent people.

It's very easy for them to get, like, high paying jobs in traditional

institutions and get lots of traditional awards and end up with big paychecks.

And there's a lot of people at big institutions that you and I know well,

and I deal with many of these where people get paid a lot of money and

they get a lot of respect and they go for 20 years and it's great and they

never create anything new, right?

There's a lot of administrators, a lot of them end up in administrative

jobs, and that's fine, that's good.

The world needs that also, right?

The innovators can't run everything because the rate

of change would be too high.

Society, I think, probably wouldn't be able to handle it.

So you need some people who are on the other side who are going to kind of keep

the lights on and keep things running.

But there is this decision that people have to make, which is okay if I

have the sort of latent capability to do this, is this actually what

I want to spend my life doing?

And do I want to go through the stress and the pain and the trauma

and anxiety and the risk of failure?

And so, do I really want to?

Once in a while you run into somebody who's just like,

can't do it any other way.

They just have to.

Andrew Huberman: Who's an example of that?

Marc Andreessen: I mean, Elon's the paramount example of our time, and

I bring him up in part because he's such an obvious example, but in part

because he's talked about this in interviews where he basically says,

he's like, I can't turn it off.

The ideas come, I have to pursue them, right?

It's why he's like running five companies at the same time

and, like working on a sixth.

It's just like he can't turn it off.

Look, there's a lot of other people who probably had the capability to do it,

who ended up talking themselves into or whatever events conspired to put them in

a position where they did something else.

Obviously, there are people who try to be creative, who

just don't have the capability.

And so, there's some venn diagram there of determinism through traits,

but also choices in life, and then also, of course, the situation in

which they're born, the context within which they grow up, culture, what their

parents expect of them, and so forth.

And so to kind of get all the way through this, you have to thread all

these needles kind of at the same time.

Andrew Huberman: Do you think there are folks out there that meet these criteria

who are disagreeable, but that can feign agreeableness, you know that can...?

[BOTH LAUGH] For those just listening, Marc just raised his right hand.

In other words, they can sort of, phrase that comes to mind maybe because

I can relate to it a little bit, they sneak up through the system, meaning

they behave ethically as it relates to the requirements of the system.

They're not breaking laws or breaking rules, in fact, quite the opposite,

they're paying attention to the rules and following the rules until

they get to a place where being disagreeable feels less threatening

to their overall sense of security.

Marc Andreessen: Yeah, I mean, look, the really highly competent people

don't have to break laws, right?

There was this myth that happened around the movie The Godfather , and

then there was this character, Meyer Lansky, who's like, ran basically

the Mafia 50, 60, 70 years ago.

And there was this great line of like, well, if Meyer Lansky had only applied

himself to running General Motors, he would have been the best CEO of all time.

It's like, no, not really, right?

The people who are great at running the big companies, they

don't have to be mob bosses.

They don't have to break laws.

They're smart and sophisticated enough to be able to work inside the system.

They don't need to take the easy out.

So, I don't think there's any implication that they have to break laws.

That said, they have to break norms, right?

And specifically, this is probably the thing that gets missed the most,

because the process of innovating, the process of creating something

new, once it works, the stories get retconned, as they say in comic books.

So the stories get adapted to where it's like it was inevitable all along.

Everybody always knew that this was a good idea.

The person has won all these awards, society embraced them.

And invariably, if you were with them when they were actually doing the work, or if

you actually get a couple of drinks into them and talk about it, it'd be like,

no, that's not how it happened at all.

They faced a wall of skepticism, just like a wall of basically

social, essentially denial.

No, this is not going to work.

No, I'm not going to join your lab.

No, I'm not going to come work for your company.

No, I'm not going to buy your product, right?

No, I'm not going to meet with you.

And so they get just like tremendous social resistance.

They're not getting positive feedback from their social network the way that

more agreeable people need to have, right?

And this is why agreeableness is a problem for innovation.

If you're agreeable, you're going to listen to the people around you.

They're going to tell you that new ideas are stupid, end of story.

You're not going to proceed.

And so I would put it more on like, they need to be able to deal with, they

need to be able to deal with social discomfort to the level of ostracism,

or at some point they're going to get shaken out and they're just going to quit.

Andrew Huberman: Do you think that people that meet these criteria

do best by banding with others that meet these criteria early?

Or is it important that they form this deep sense of self, like the ability

to cry oneself to sleep at night or lie in the fetal position, worrying

that things aren't going to work out and then still get up the next

morning and get right back out there.

Marc Andreessen: Right.

So, Sean Parker has the best line, by the way, on this.

He says being an entrepreneur or being a creator is like getting punched

in the face over and over again.

He said, eventually you start to like the taste of your own blood.

And I love that line because it makes everybody massively uncomfortable,

but it gives you a sense of how basically painful the process is.

If you talk to any entrepreneur who's been through it about that, they're like,

oh, yeah, that's exactly what it's like.

So, there is a big individual component to it.

But look, it can be very lonely, and especially very hard, I think, to do

this if nobody around you is trying to do anything even remotely similar.

And if you're getting just universally negative responses,

like very few people, I think very few people have the ego strength to

be able to survive that for years.

So I do think there's a huge advantage, and this is why you do see clusters.

There's a huge advantage to clustering.

Throughout history, you've had this clustering effect.

You had clustering of the great artists and sculptors in, you had the

clustering of the philosophers of Greece.

You had the clustering of tech people in Silicon Valley.

You have the clustering of know, arts, movie, TV people

in Los Angeles, and so forth.

And so, know, there's always a scene, right?

There's always, like a nexus and a place where people come

together for these kinds of things.

So, generally speaking, if somebody wants to work in tech, innovate in tech,

they're going to be much better off being around a lot of people who are trying to

do that kind of thing than they are in a place where nobody else is doing it.

Having said that, the clustering can have downsides, it can have side effects.

And you put any group of people together, and you do start to get

groupthink, even among people who are individually very disagreeable.

And so these same clusters where you get these very idiosyncratic

people, they do have fads and trends just like every place else.

And so they get wrapped up in their own social dynamics.

The good news is the social dynamic in those places is usually very forward

looking, and so it's usually like, I don't know, it's like a herd of iconoclasts

looking for the next big thing.

So iconoclasts, looking for the next big thing.

That's good.

The herd part.

That's what you've got to be careful of.

So even when you're in one of these environments, you have to

be careful that you're not getting sucked into the groupthink too much.

Andrew Huberman: When you say groupthink, do you mean excessive friction?

Do you do pressure testing each other's ideas to the point where

things just don't move forward?

Or are you talking about groupthink, where people start to form a consensus?

Or the self belief that, gosh, we are so strong because we are so different?

Can we better define groupthink?

Marc Andreessen: It's actually less either one of those things both happen.

Those are good.

Those are good.

The part of groupthink I'm talking about is just like, we all basically zero in, we

just end up zeroing in on the same ideas.

Right.

In Hollywood, there's this classic thing.

There are years where all of a sudden there's, like, a lot of volcano movies.

It's like, why are there all these volcano movies?

And it's just like, there was just something in the gestalt, right?

There was just something in the air.

Look, Silicon Valley has this.

There are moments in time where you'll have these.

It's like the old thing.

What's the difference between a fad and a trend?

Fad is the trend that doesn't last.

Right.

And so Silicon Valley is subject to both fads and trends, just like any place.

In other words, you take smart, disagreeable people, you cluster them

together, they will act like a herd.

They will end up thinking the same things unless they try very hard not to.

Andrew Huberman: You've talked about these personality traits of great innovators

before, and we're talking about them now.

You invest in innovators, you try and identify them, and you are one.

So you can recognize these traits here.

I'm making the presumption that you have these traits.

Indeed you do.

We'll just get that out of the way.

Have you observed people trying to feign these traits, and are there any

specific questions or behaviors that are a giveaway that they're pretending to

be the young Steve Jobs or that they're pretending to be the young Henry Ford?

Pick your list of other names that qualify as authentic, legitimate innovators.

We won't name names of people who have tried to disguise

themselves as true innovators.

But what are some of the litmus tests?

And I realize here that we don't want you to give these away to the

point where they lose their potency.

But if you could share a few of those.

Marc Andreessen: Good, we're actually a pretty open book on this.

First of all, yes, so there are people who definitely try to come in and basically

present as being something that they're not, and they've read all the books.

They will have listened to this interview.

They study everything and they construct a facade, and they come in

and present as something they're not.

I would say the amount of that varies exactly, correlated to the NASDAQ.

And so when stock prices are super low, you actually get the opposite.

When stock prices are super low, people get too demoralized.

And people who should be doing it basically give up because they

just think that the industry is over, the trend is over, whatever.

It's all hopeless.

And so you get this flushing thing.

So nobody ever shows up at a stock market low and says, like, I'm the

new next big thing and doesn't really want to do it because there are higher

status, the kinds of people who do the thing that you're talking about, they're

fundamentally oriented for social status.

They're trying to get the social status without actually the substance.

And there are always other places to go to get social status.

So after 2000, the joke was, when I got to Silicon Valley in

'93, '94, the Valley was dead.

We can talk about that.

By '98, it was roaring, and you had a lot of these people showing up, who

were, you basically had a lot of people showing up with these kind of stories.

2000, the market crashed.

By 2001, the joke was that there were these terms, B to C and B to B.

And in 1998, they meant B to C meant business to consumer and B to B

meant business to business, which is two different kinds of business

models for Internet companies.

By 2001, B to B meant back to banking and B to C meant back to consulting,

which is the high status people who, the people oriented to status, who showed up

to be in tech were like, yeah, screw it.

This is over.

Stick a fork in it.

I'm going to go back to Goldman Sachs or go back to McKinsey,

where I can be high status.

And so you get this flushing kind of effect that happens in a downturn.

That said, in a big upswing, yeah, you get a lot of people showing up with a

lot of kind of, let's say, public persona without the substance to back it up.

So the way we stress that you can actually say exactly how we test for this, because

the test exactly addresses the issue in a way that is impossible to fake.

And it's actually the same way homicide detectives try to find out if you've

actually, like, if you're innocent or whether you've killed somebody.

It's the same tactic, which is, you ask increasingly detailed questions, right?

And so the way the homicide cop does this is, what were you doing last night?

Oh, I was at a movie.

Which movie?

Which theater?

Okay, which seat did you sit in?

Okay, what was the end of the movie?

And you ask increasingly detailed questions and people have trouble.

At some point, people have trouble making up and things just fuzz

into just kind of obvious bullshit.

And basically fake founders basically have the same problem.

They're able to relay a conceptual theory of what they're doing that they've

kind of engineered, but as they get into the details, it just fuzzes out.

Whereas the true people that you want to back that can do it, basically what

you find is they've spent five or ten or 20 years obsessing on the details

of whatever it is they're about to do.

And they're so deep in the details that they know so much

more about it than you ever will.

And in fact, the best possible reaction is when they get mad, which

is also what the homicide cops say.

What you actually want is you want the emotional response of like, I can't

believe that you're asking me questions this detailed and specific and picky

and they kind of figure out what you're doing and then they get upset.

That's good, that's perfect, right?

But then they have to have proven themselves in the sense of,

they have to be able to answer the questions in great detail.

Andrew Huberman: Do you think that people that are able to answer those questions

in great detail have actually taken the time to systematically think through the

if-ands of all the possible implications of what they're going to do and they

have a specific vision in mind of how things need to turn out or will turn out?

Or do you think that they have a vision and it's a no matter

what, it will work out because the world will sort of bend around it?

I mean, in other words, do you think that they place their vision in

context or they simply have a vision and they have that tunnel vision of

that thing and that's going to be it?

Let's use you for an example with Netscape.

That's how I first came to know your name.

When you were conceiving Netscape, did you think, okay, there's this

search engine and this browser and it's going to be this thing that

looks this way and works this way and feels this way, did you think that?

And also think about that there was going to be a gallery of other search

engines and it would fit into that landscape of other search engines?

Or were you just projecting your vision of this thing as this

unique and special brainchild?

Marc Andreessen: Let me give the general answer, and then we can

talk about the specific example.

So the general answer is what?

Entrepreneurship, creativity, innovation is what economists call

decision making under uncertainty.

In both parts, those are important decision making.

Like, you're going to make a ton of decisions because you have to

decide what to do, what not to do.

And then uncertainty, which is like, the world's a complicated place.

And in mathematical terms, the world is a complex adaptive

system with feedback loops.

And Isaac Asimov wrote in his novels, he wrote about this field

called psychohistory, which is the idea that there's like a

supercomputer that can predict the future of human affairs, right?

And it's like, we don't have that.

[LAUGHS] Not yet.

Andrew Huberman: [LAUGHS] Not yet.

We'll get to that later.

Marc Andreessen: We certainly don't have that yet.

And so you're just dealing, you know, military commanders call

this the fog of war, right?

You're just dealing with a situation where the number of

variables are just off the charts.

It's all these other people who are inherently unpredictable, making all

these decisions in different directions.

And then the whole system is combinatorial, which is these

people are colliding with each other, influencing their decisions.

And so, I mean, look, the most straightforward kind of way to

think about this is, it's amazing.

Like, anybody who believes in economic central planning, it always

blows my mind because it's just like, try opening a restaurant.

Try just opening a restaurant on the corner down here.

And like 50/50 odds, the restaurant is going to work.

And all you have to do to run a restaurant is have a thing and serve food.

And it's like most restaurants fail, right?

People who run restaurants are pretty smart.

They usually think about these things very hard, and they all want to

succeed, and it's hard to do that.

And so to start a tech company or to start an artistic movement or to fight

a war, you're just going into this, basically conceptual battleground or

in military terms, real battleground, where there's just like incredible levels

of complexity, branching future paths, and so there's nothing predictable.

And so what we look for is basically the really good innovators.

They've got a drive to basically be able to cope with that and deal with that.

And they basically do that in two steps.

So one is they try to pre-plan as much as they possibly can and we

call that the process of navigating the, what we call the idea maze.

And so the idea maze basically is, I've got this general idea, and it might be

the Internet is going to work or search or whatever, and then it's like, okay, in

their head, they have thought through of like, okay, if I do it this way, that way,

this third way, here's what will happen.

Then I have to do that, then I have to do this, then I have to

bring in somebody to do that.

Here's the technical challenge I'm going to hit.

And they got in their heads as best anybody could, they've got as

complete a sort of a map of possible futures as they could possibly have.

And this is where I say, when you ask them increasingly detailed questions, that's

what you're trying to kind of get them to kind of chart out, is, okay, how far ahead

have you thought, and how much are you anticipating all of the different twists

and turns that this is going to take?

Okay, so then they start on day one, and then, of course, what

happens is now they're in it, now they're in the fog of war, right?

They're in future uncertainty.

And now that idea maze is maybe not helpful practically, but now they're

going to be basically constructing it on the fly, day by day, as they

learn and discover new things and as the world changes around them.

And of course, it's a feedback loop, because if their thing starts to

work, it's going to change the world.

And then the fact the world is changing is going to cause

their plan to change as well.

And so, yeah, the great ones, basically, the great ones

course correct every single day.

They take stock of what they've learned.

They modify the plan.

The great ones tend to think in terms of hypotheses, right?

Like a scientific sort of mentality, which is they tend to think,

okay, I'm going to try this.

I'm going to go into the world, I'm going to announce that I'm doing this for sure.

I'm going to say, this is my plan.

I'm going to tell all my employees that, and I'm going to tell all my

investors that, and I'm going to put a stake in there, and it's my plan,

and then I'm going to try it, and even though I sound like I have complete

certainty, I know that I need to test to find out whether it's going to work.

And if it's not, then I have to go back to all those same people and

have to say, well, actually, we're not going left, we're going right.

And they have to run that loop thousands of times to get through the other side.

And this led to the creation of this great term pivot, which has been very helpful

in our industry because the word, when I was young, the word we used was fuck

up, and pivot sounds like so much better, sounds like so much more professional.

But, yeah, you make mistakes.

It's just too complicated to understand.

You course correct, you adjust, you evolve.

Often these things, at least in business, the businesses that end up working

really well tend to be different than the original plan, but that's part of

the process of a really smart founder basically working their way through

reality as they're executing their plan.

Andrew Huberman: The way you're describing this has parallels to

a lot of models in biology and the practice of science, random walks,

but that aren't truly random, pseudo-random walks in biology, etc.

But one thing that is becoming clear from the way you're describing

this is that I could imagine a great risk to early success.

So, for instance, somebody develops a product, people are excited by it,

they start to implement that product, but then the landscape changes, and

they don't learn how to pivot to use the less profane version of it.

They don't learn how to do that.

In other words, and I think of everything these days, or most everything, in

terms of reward schedules and dopamine reward schedules, because that is

the universal currency of reward.

And so when you talk about the Sean Parker quote of learning to enjoy

the taste of one's own blood, that is very different than learning to

enjoy the taste of success, right?

It's about internalizing success as a process of being self

determined and less agreeable, etc.

In other words, building up of those five traits becomes the source of dopamine,

perhaps in a way that's highly adaptive.

So on the outside, we just see the product, the end product, the iPhone,

the MacBook, the Netscape, etc.

But I have to presume, and I'm not a psychologist, but I have done

neurophysiology and I've studied the dopamine system enough to know that

what's being rewarded in the context of what you're describing sounds

to be a reinforcement of those five traits, rather than, oh, it's going

to be this particular product, or the company is going to look this way, or

the logo is going to be this or that.

That all seems like the peripheral to what's really going on, that great

innovators are really in the process of establishing neural circuitry

that is all about reinforcing the me and the process of being.

Marc Andreessen: So this is like extrinsic versus intrinsic motivation.

So, the Steve Jobs kind of Zen version of this, right?

Or the sort of hippie version of this was the journey is the reward.

He always told his employees that.

It's like, look, everybody thinks in terms of these big public markers,

like the stock price or the IPO or the product launch or whatever.

He's like, no, it's actually the process itself is the point.

Right to your point, if you have that mentality, then that's an intrinsic

motivation, not an extrinsic motivation.

And so that's the kind of intrinsic motivation that can

keep you going for a long time.

Another way to think about it is competing against yourself, right?

It's like, can I get better at doing this?

And can I prove to myself that I can get better?

There's also a big social component to this, and this is one of the

reasons why Silicon Valley punches so far above its weight as a place.

There's a psychological component which also goes to the comparison set.

So a phenomenon that we've observed over time is the leading tech company

in any city will aspire to be as large as the previous leading tech company in

that city, but often not larger, right?

Because they have a model of success.

And as long as they beat that level of success, they've kind of checked

the box like they've made it.

But then, in contrast, you're in Silicon Valley, and you look around

and it's just like Facebook and Cisco and Oracle and Hewlett Packard and--

Andrew Huberman: --Gladiators--

Marc Andreessen: --Yeah.

And you're just, like, looking at these giants.

Many of them are still, Mark Zuckerberg, still going to work every day.

And so these people are, like, the role models are, like, alive.

They're, like, right there, and it's so clear how much better they are and how

much bigger their accomplishments are.

And so what we find is young founders in that environment

have much greater aspirations.

Because, again, at that point, maybe it's the social status, maybe there's

an extrinsic component to that, or maybe it helps calibrate that internal

system to basically say, actually, no, the opportunity here is not to build

what you may call a local maximum form of success, but let's build to a

global maximum form of success, which is something as big as we possibly can.

Ultimately, the great ones are probably driven more internally than

externally when it comes down to it.

And that is where you get this phenomenon where you get people who are extremely

successful and extremely wealthy who very easily could punch out and

move to Fiji and just call it, and they're still working 16 hour days.

Obviously something explains that that has nothing to do with external rewards,

and I think it's an internal thing.

Andrew Huberman: As many of you know, I've been taking AG1 daily

since 2012, so I'm delighted that they're sponsoring the podcast.

AG1 is a vitamin mineral probiotic drink that's designed to meet all of

your foundational nutrition needs.

Now, of course, I try to get enough servings of vitamins and minerals

through whole food sources that include vegetables and fruits every day.

But oftentimes I simply can't get enough servings.

But with AG1, I'm sure to get enough vitamins and minerals

and the probiotics that I need.

And it also contains adaptogens to help buffer stress.

Simply put, I always feel better when I take AG1.

I have more focus and energy, and I sleep better.

And it also happens to taste great.

For all these reasons, whenever I'm asked if you could take just

one supplement, what would it be?

I answer AG1.

If you'd like to try AG1, go to drinkag1.com/huberman

to claim a special offer.

They'll give you five free travel packs plus a year's supply of Vitamin D3K2.

Again, that's drinkag1.com/huberman.

I've heard you talk a lot about the inner landscape, the inner psychology

of these folks, and I appreciate that.

We're going even deeper into that today.

And we will talk about the landscape around whether or not Silicon Valley

or New York, whether or not there are specific cities that are ideal

for certain types of pursuits.

I think there was an article written by Paul Graham some years ago, about the

conversations that you overhear in a city will tell you everything you need to know

about whether or not you belong there in terms of your professional pursuits.

Some of that's changed over time, and now we should probably add Austin to the

mix because it was written some time ago.

In any event, I want to return to that, but I want to focus on an aspect

of this intrinsic versus extrinsic motivators in terms of something

that's a bit more cryptic, which is one's personal relationships.

If I think about the catalog of innovators in Silicon Valley, some of them, like

Steve Jobs, had complicated personal lives, romantic personal lives early

on, and it sounds like he worked it out.

I don't know.

I wasn't their couple's therapist.

But when he died, he was in a marriage that for all the world

seemed like a happy marriage.

You also have examples of innovators who have had many partners, many

children with other partners.

Elon comes to mind.

I don't think I'm disclosing anything that isn't already obvious.

Those could have been happy relationships and just had many of them.

But the reason I'm asking this is you can imagine that for the innovator,

the person with these traits, who's trying to build up this thing, whatever

it is, that having someone, or several people in some cases, who just truly

believe in you when the rest of the world may not believe in you yet or

at all, could be immensely powerful.

And we have examples from cults that embody this.

We have examples from politics.

We have examples from tech innovation and science.

And I've always been fascinated by this because I feel like it's the more

cryptic and yet very potent form of allowing someone to build themselves up.

It's a combination of inner psychology and extrinsic motivation.

Because obviously, if that person were to die or leave them or cheat

on them or pair up with some other innovator, which we've seen several

times recently and in the past, it can be devastating to that person.

But what are your thoughts on the role of personal, and in particular,

romantic relationship as it relates to people having an idea and their

feeling that they can really bring that idea to fruition in the world?

Marc Andreessen: So it's a real mixed bag.

You have lots of examples in all directions, and I

think it's something like.

Something like the following.

So first, we talked about the personality traits of these people.

They tend to be highly disagreeable.

Andrew Huberman: Doesn't foster a good romantic relationship.

Marc Andreessen: Highly disagreeable people can be

difficult to be in a relationship.

[LAUGHS]

Andrew Huberman: [LAUGHS] I may have heard of that once or twice before.

A friend may have given me that example.

Marc Andreessen: Yeah.

Right.

And maybe you just need to find the right person who compliments that

and is willing to, there's a lot of relationships where it's always this

question about relationships, right?

Which is, do you want to have the same personality growth profile, the

same behavioral traits, basically, as your partner, or do you actually

want to have, is it an opposite thing?

I'm sure you've seen this.

There are relationships where you'll have somebody who's highly disagreeable,

who's paired with somebody who's highly agreeable, and it actually works out great

because one person just gets to be on their soapbox all the time, and the other

person is just like, okay, it's fine.

Right? It's fine.

It's good.

You put two disagreeable people together, maybe sparks fly and they

have great conversations all the time, and maybe they come to hate each other.

Anyway, so these people, if you're going to be with one of these

people, you're fishing out of the disagreeable end of the pond.

And again, when I say disagreeable, I don't mean these are normal distributions.

I don't mean, like 60% disagreeable or 80% disagreeable.

The people we're talking about are 99.99% disagreeable.

So these are ordinary people.

So part of it's that.

And then, of course, they have the other personality traits.

They're super conscientious.

They're super driven.

As a consequence, they tend to work really hard.

They tend to not have a lot of time for family vacations or other things.

Then they don't enjoy them if they're forced to go on them.

And so, again, that kind of thing can fray at a relationship.

So there's a fair amount in there that's loaded.

Like, somebody who's going to partner with one of these people

needs to be signed up for the ride.

And that's a hard thing.

That's a hard thing to do.

Or you need a true partnership of two of these, which is also hard to do.

So I think that's part of it.

And then, look, I think a big part of it is people achieve a certain level of

success, and either in their own minds or publicly, and then they start to be

able to get away with things, right?

And they start to be able to.

It's like, well, okay, now we're rich and successful and famous, and now I

deserve, and this is where you get into...

I view this now in the realm of personal choice.

You get into this thing where people start to think that they deserve things,

and so they start to behave in very bad ways, and then they blow up their

personal worlds as a consequence.

And maybe they regret it later, and maybe they don't.

Right?

It's always a question.

I think there's that.

And then, I don't know, maybe the other part of it is that some people just

need more emotional support than others.

And I don't know that that's a big, I don't know that that tilts either way.

I know some of these people who have great, loving relationships and seem

to draw very much on having this kind of firm foundation to rely upon.

And then I know other people who are just like, their personal lives

are just a continuous train wreck.

And it doesn't seem to matter, like, professionally, they just

keep doing what they're doing.

And maybe we could talk here about whatever is the personality

trait for risk taking.

Some people are so incredibly risk prone that they need to take risk in

all aspects of their lives at all times.

And if part of their life gets stable, they find a way to blow it up.

And that's some of these people you could describe in those terms also.

Andrew Huberman: Yeah, let's talk about that.

Because I think risk taking and sensation seeking is something that

fascinates me for my own reasons and in my observations of others.

Does it dovetail with these five traits in a way that can really serve innovation,

in ways that can benefit everybody?

The reason I say to benefit everybody is because there is a view of how

we're painting this picture of the innovator as this really cruel person.

But oftentimes, what we're talking about are innovations that make the

world far better for billions of people.

Marc Andreessen: Yeah, that's right.

And by the way, everything we're talking about also is not just in

tech or science or in business.

Everything we're also talking about is true for the arts.

The history of artistic expression.

You have people with all these same kinds of traits.

Andrew Huberman: Well, I was thinking about Picasso and his regular turnover

of lovers and partners, and he was very open about the fact that it was one of the

sources of his productivity, creativity.

He wasn't shy about that.

I suppose if he were alive today, it might be a little bit different.

He might be judged a little differently.

Marc Andreessen: Or that was his story for behaving in a pattern

that was very awful for the people around him, and he didn't care.

Andrew Huberman: Right, maybe they left him?

Marc Andreessen: Yeah. Who knows?

Right?

Puts and takes to all this, but no.

Okay, so I have a theory.

So here's a theory.

This is one of these, I keep a list of things that will get me

kicked out of a dinner party and topics at any given point in time.

Andrew Huberman: Do you read it before you go in?

Marc Andreessen: Yeah.

On auto recall, so that I can get out of these things.

Here's the thing that can get me kicked out of a dinner

party, especially these days.

So think of the kind of person where it's very clear that they're super high, to

your point, this is somebody who's super high output in whatever domain they're in.

They've done things that have fundamentally changed the world.

They've brought new, whether it's businesses or technologies or works

of art, entire schools of creative expression, in some cases to the world.

And then at a certain point, they blow themselves to smithereens, right?

And they do that either through a massive financial scandal.

They do that through a massive personal breakdown.

They do that through some sort of public expression that causes

them a huge amount of problems.

They say the wrong thing, maybe not once, but several hundred times,

and blow themselves to smithereens.

There's this moral arc that people kind of want to apply, which it's

like the Icarus flying too close to the sun and he had it coming and he

needed to keep his ego under control.

And you get kind of this judgment that applies.

So I have a different theory on this.

So the term I use to describe these people, and by the way, a lot of other

people who don't actually blow themselves up but get close to it, which is a

whole 'nother set of people, I call them martyrs to civilizational progress.

We're backwards, civilizational progress.

So look, the only way civilization gets moved forward is when people

like this do something new.

Because civilization as a whole does not do new things.

Groups of people do not do new things.

These things don't happen automatically.

By default nothing changes.

The only way civilizational change on any of these axes ever happens is because one

of these people stands up and says, no, I'm going to do something different than

what everybody else has ever done before.

So, this is progress, like, this is actually how it happens.

Sometimes they get lionized or awarded.

Sometimes they get crucified.

Sometimes the crucifixion is literal.

Sometimes it's just symbolic.

But they are those kinds of people, and then martyrs when they go down in

flames and again, this is where it really screws the people's moral judgments

because everybody wants to have the sort of super clear story of like, okay, he

did a bad thing and he was punished.

And I'm like, no, he was the kind of person who was going to do great things

and also was going to take on a level of risk and take on a level of sort of

extreme behavior such that he was going to expose himself to flying too close to

the sun, wings melt and crash to ground.

But it's a package deal.

The reason you have the Picasso's and the Beethovens and all these

people is because they're willing to take these extreme level of risks.

They are that creative and original, not just in their art or their business,

but in everything else that they do that they will set themselves up

to be able to fail psychologically.

A psychologist would probably, or psychiatrist would probably say maybe.

To what extent do they actually have a death wish at some point.

Do they want to punish themselves?

Do they want to fail?

That I don't know.

But you see this.

They deliberately move themselves too close to the sun, and you can see it

when it's happening, because if they get too far away from the sun, they

deliberately move back towards it.

Right.

They come right back, and they want the risk anyway.

So martyrs to civilizational progress.

This is how progress happens.

When these people crash and burn, the natural inclination

is to judge them morally.

I tend to think we should basically say, look, and I don't even know if this

means, like, giving them a moral pass or whatever, but it's like, look, this

is how civilization progresses, and we need to at least understand that there's

a self sacrificial aspect to this that may be tragic and often is tragic, but

it is quite literally self sacrificial.

Andrew Huberman: Are there any examples of great innovators who were able to

compartmentalize their risk taking to such a degree that they had what seemed

to be a morally impeccable life in every domain except in their business pursuits?

Marc Andreessen: Yeah, that's right.

So some people are very highly controlled like that.

Some people are able to very narrowly, and I don't really want to set myself

an example on a lot of this, but I will tell you as an example, I will

never use debt in business, number one.

Number two, I have the most placid personal life you can imagine.

Number three, I'm the last person in the world who is ever

going to do an extreme sport.

I mean, I'm not even going to go in the sauna on the ice bath.

I'm not doing any of this.

I'm not tele skiing.

Andrew Huberman: No obligation.

Marc Andreessen: I'm not on the Titan.

I'm not going down to see the Titanic.

Goodness, you weren't doing any of this.

I'm not doing any of this stuff.

I have no interest.

I don't play golf.

I don't ski.

I have no interest in any of this stuff, right?

And I know people like this, right, who are very high achievers.

It's just like, yeah, they're completely segmented.

They're extreme risk takers.

In business, they're completely buttoned down on the personal side, they're

completely buttoned down financially.

They're scrupulous with following every rule and law you can possibly imagine,

but they're still fantastic innovators.

And then I know many others who are just like their life is on fire all

the time, in every possible way.

And whenever it looks like the fire is turning into embers, they figure out

a way to relight the fire, and they just really want to live on the edge.

And so I think that's an independent variable.

And again, I would apply the same thing.

I think the same thing applies to the arts.

Classical music as an example.

I think Bach was, as an example, one of the best musicians of all

time, had just a completely sedate personal life, never had any aberrant

behavior at all in his personal life.

Family man, tons of kids, apparently pillar of the community.

Right.

And so if Bach could be Bach and yet not burn his way through 300 mistresses

or whatever, maybe you can, too.

Andrew Huberman: So in thinking about these two different categories of

innovators, those that take on tremendous risk in all domains of their life and

those that take on tremendous risk in a very compartmentalized way, I don't

know what the percentages are, but I have to wonder if in this modern age of

the public being far less forgivable, what I'm referring to is cancel culture.

Do you think that we are limiting the number of innovations in total

by just simply frightening or eliminating an enormous category of

innovators because they don't have the confidence or the means or the

strategies in place to regulate?

So they're just either bowing out or they're getting crossed off,

they're getting canceled one by one.

Marc Andreessen: So do you think the public is less tolerant than

they used to be or more tolerant?

Andrew Huberman: Well, the systems that, I'm not going to be careful here.

I think the large institution systems are not tolerant of what the public

tells them they shouldn't be tolerant of.

And so if there's enough noise, there's enough noise in the mob.

I think institutions bow out.

And here I'm referring not just to, they essentially say, okay,

let the cancellation proceed.

Maybe they're the gavel that comes down, but they're not the

lever that got the thing going.

And so I'm not just thinking about universities.

I'm also thinking about advertisers.

I'm thinking about the big movie houses that cancel a film that a

given actor might be in because they had something in their personal life

that's still getting worked out.

I'm thinking about people who are in a legal process that's not

yet resolved, but the public has decided they're a bad person, etc.

Marc Andreessen: My question is, are we really talking about the public?

I agree with your question, and I'm going to come back to it, but I'm

going to examine one part of your question, which is, is this really

the public we're talking about.

And I would just say Exhibit A is who is the current frontrunner for

the Republican nomination today?

The public, at least on one side of the political aisle, seems very on board.

Number two, like, look, there's a certain musician who flew too close to

the sun, blew himself to smithereens.

He's still hitting all time highs on music streams every month.

The public seems fine.

I would argue the public is actually more open to these things than

it actually maybe ever has been.

And we could talk about why that's the case.

I think it's a differentiation, and this is what your question was

aiming at, but it's a differentiation between the public and the elites.

My view is everything that you just described is an elite phenomenon.

And actually, the public is very much not on board with it.

So what's actually happening is what's happened is the public

and the elites have gapped out.

The public is more forgiving of what previously might have been considered kind

of aberant and extreme behavior, right?

F.

Scott Fitzgerald, "there are no second acts in American lives"

turns out was completely wrong.

Turns out there are second acts, third acts, fourth acts.

Apparently you can have an unlimited number of acts.

The public is actually up for it.

Yeah.

Andrew Huberman: I mean, I think of somebody like Mike Tyson, right?

I feel like his life exemplifies everything.

That's amazing and great and also terrible about America.

Marc Andreessen: If we took Mike Tyson to dinner tonight at any restaurant anywhere

in the United States, what would happen?

Andrew Huberman: He would be loved.

Marc Andreessen: Oh, he would be like, the outpouring of enthusiasm and

passion and love would be incredible.

It would be unbelievable.

This is a great example.

And again, I'm not even going to draw more.

I'm not even going to say I agree with that or disagree with that.

I think we all intuitively know that the public is just like, 100%, absolutely.

He's a legend. He's a living legend.

He's like a cultural touchstone.

Absolutely.

And you see it when he shows up in movies, right?

I don't remember the, I mean, the big breakthrough where I figured this out

with respect to him because I don't really follow sports, but when he showed up in

that, it was that first Hangover movie, and he shows up and I was in a theater

and the audience just goes, bananas crazy.

They're so excited to see him.

Andrew Huberman: He evokes delight.

I always say that Mike Tyson is the only person I'm aware of that can

wear a shirt with his own name on it, and it somehow doesn't seem wrong.

In fact, it just kind of makes you like him more.

His ego feels very contoured in a way that he knows who he is and who he was, and yet

there's a humbleness woven in, maybe as a consequence of all that he's been through.

I don't know.

But, yeah, people love Mike.

Marc Andreessen: Public loves him now.

Exactly.

Now, if he shows up to lecture at Harvard, right, I think you're probably

going to get a different reaction?

[LAUGHS] Andrew Huberman: I don't know.

I don't know!

You know, the guy who wrote The Wire gave a talk at Harvard, and it sounded

to me, based on his report of that, which is very interesting, in fact,

that people adore people who are connected to everybody in that way.

I feel like everybody loves Mike.

From above his status, the sides below his status, he occupies

this halo of love and adoration.

Marc Andreessen: Okay.

Andrew Huberman: All right.

Marc Andreessen: Yeah.

Look, the other side of this is the elites, and you kind of alluded

to this, of the institution.

So basically, it's like the people who are at least nominally in charge or

feel like that they should be in charge.

Andrew Huberman: I want to make sure we define elite.

So you're not necessarily talking about people who are wealthy.

You're talking about people who have authority within institutions.

Marc Andreessen: So the ultimate definition of an elite is

who can get who fired, right.

That's the ultimate test.

Who can get who fired, boycotted, blacklisted, ostracized.

Like when push, prosecuted, jailed, like when push comes to shove.

I think that's always the question, who can destroy whose career?

And of course, you'll notice that that is heavily asymmetric

when these fights play out.

Like, it's very clear which side can get the other side fired and which side can't.

And so, yeah, so, look, I think we live in a period of time where

the elites have gotten to be extreme in a number of dimensions.

I think it's characterized by, for sure, extreme groupthink, extreme

sanctimony, extreme moral, I would say dudgeon, this weird sort of modern

puritanism, and then an extreme sort of morality of punishment and terror

against their perceived enemies.

But I want to go through that because I actually think that's

a very different phenomenon.

I think what's happening at the elites is very different than what's

happening in the population at large.

And then, of course, I think there's a feedback loop in there, which

is, I think the population at large is not on board with that program.

Right.

I think the elites are aware that the population is not

on board WIth that program.

I think they judge the population negatively as a consequence, that causes

the elites to harden their own positions.

That causes them to be even more alienating to the population.

And so they're in sort of an oppositional negative feedback loop.

But again, it's a sort of question, okay, who can get who fired?

And so elites are really good at getting normal people fired.

Ostracized, banned, hit pieces in the press, like, whatever.

For normal people to get elites fired, they have to really band together, right.

And really mount a serious challenge, which mostly doesn't happen, but might

be starting to happen in some cases.

Andrew Huberman: Do you think this power of the elites over, stemmed

from social media sort of going against its original purpose?

I mean, when you think social media, you think you're giving each

and every person their own little reality TV show, their own voice.

And yet we've seen a dramatic uptick in the number of cancellations and

firings related to immoral behavior based on things that were either

done or amplified on social media.

It's almost as if the public is holding the wrong end of the knife.

Marc Andreessen: Yeah, so the way I describe it, I use these two terms,

and they're somewhat interchangeable, but elites and institutions.

And then they're somewhat interchangeable because who runs the institutions?

The elites, right?

And so it's sort of a self reinforcing thing.

And institutions of all kinds.

Institutions, everything from the government, bureaucracies, companies,

nonprofits, foundations, NGOs, tech companies, on and on and on.

Like people who are in charge of big complexes and that carry a lot of,

basically, power and influence and capability and money as a consequence

of their positional authority.

So the head of a giant foundation may never have done anything in their

life that would cause somebody to have a high opinion of them as a person.

But they're in charge of this gigantic multi billion dollar

complex and have all this power.

And so that's just defined terms, at least in institutions.

So, it's actually interesting.

Gallup has been doing polls on the following on the question of trust

in institutions, which is sort of therefore a proxy for trust in elites,

basically since the early 1970s.

And they do this across all the categories of big institutions, basically everyone.

I just talked about a bunch of others.

Big business, small business, banks, newspapers, broadcast

television, the military, police.

So they've got like 30 categories or something.

And basically what you see is almost all the categories basically started in

the early 70s at like 60 or 70% trust.

And now almost across the board, they've just had a complete,

basically linear slide down for 50 years, basically my whole life.

And they're now bottoming out.

Congress and journalists bottom out at like 10%.

The two groups everybody hates are Congress and journalists.

And then it's like a lot of other big institutions are

like, in their 20s, 30s, 40s.

Actually, big business actually scores fairly high.

Tech actually scores quite high.

The military scores quite high.

But basically everything else has really caved in.

This is sort of my fundamental challenge to everybody who basically says, and

you didn't do this, but you'll hear the simple form of this, which is social

media caused the current trouble.

And let's call this an example, collapse in faith in institutions and elites.

Let's call that part of the current trouble.

Everybody's like, well, social media caused that.

I was like, well, no, social media, social media is new, right?

In the last...

social media is effectively new, practically speaking, since 2010,

2012 is when it really took off.

And so, if the trend started in the early 1970s and has been continuous, then

we're dealing with something broader.

Martin Gurri wrote, I think, the best book on this called the Revolt of the Public

, where he goes through this in detail.

He does say that social media had a lot to do with what's

happened in the last decade.

But he says, yeah, if you go back, you look further, it was

basically two things coinciding.

One was just a general change in the media environment.

And in particular, the 1970s is when you started to, and especially in the 1980s,

is when you started to get specifically talk radio, which was a new outlet.

And then you also got cable television.

And then you also, by the way, it's actually interesting in that you had

paperback books, which was another one of these, which was an outlet.

So you had like a fracturing in the media landscape that started in the

50s through the, then, of course, the Internet blew it wide open.

Having said that, if the elites and the institutions were fantastic,

you would know it more than ever.

Information is more accessible.

And so the other thing that he says, and I agree with, is the public is

not being tricked into thinking the elites and institutions are bad.

They're learning that they're bad, and therefore, the mystery of the Gallup

poll is why those numbers aren't all just zero, which is arguably, in a

lot of cases, where they should be.

Andrew Huberman: I think one reason that--

Marc Andreessen: --By the way, he thinks this is bad.

So he and I have a different view.

So here's where he and I disagree.

He thinks this is bad.

So he basically says, you can't replace elites with nothing.

You can't replace institutions with nothing, because what you're just left

with is just going to be wreckage.

You're going to be left with a completely, basically atomized, out of control

society that has no ability to marshal any sort of activity in any direction.

It's just going to be a dog eat dog awful world.

I have a very different view on that which we can talk about.

Andrew Huberman: Yeah, I'd love to hear your views on that.

I'd like to take a quick break and acknowledge our sponsor, InsideTracker.

InsideTracker is a personalized nutrition platform that analyzes

data from your blood and DNA to help you better understand your body and

help you meet your health goals.

I'm a big believer in getting regular blood work done for the simple reason

that many of the factors that impact your immediate and long term health can only

be analyzed from a quality blood test.

However, with a lot of blood tests out there, you get information back

about blood lipids, about hormones and so on, but you don't know

what to do with that information.

With InsideTracker, they have a personalized platform that makes it very

easy to understand your data, that is, to understand what those lipids, what

those hormone levels, etc., mean, and behavioral supplement, nutrition and

other protocols to adjust those numbers to bring them into the ranges that are ideal

for your immediate and long term health.

InsideTracker's ultimate plan now includes measures of both APOB and of Insulin,

which are key indicators of cardiovascular health and energy regulation.

If you'd like to try InsideTracker, you can visit insidetracker.com/huberman to

get 20% off any of InsideTracker's plans.

Again, that's insidetracker.com/huberman to get 20% off.

The quick question I was going to ask before we go there is, I think that

one reason that I and many other people sort of reflexively assume that social

media caused the demise of our faith and institutions is, well, first of all, I

wasn't aware of this lack of correlation between the decline in faith in

institutions and the rise of social media.

But secondarily that we've seen some movements that have essentially

rooted themselves in tweets, in comments, in posts that get amplified,

and those tweets and comments and posts come from everyday people.

In fact, I can't name one person who initiated a given cancellation or

movement because it was the sort of dogpiling or mob adding-on to some

person that was essentially anonymous.

So I think that for many of us, we have the, to use neuroscience language,

as sort of a bottom up perspective, oh, someone sees something in their

daily life or experiences something in their daily life, and they tweet about

it or they comment about it or they post about it, and then enough people

dogpile on the accused that it picks up force, and then the elites feel

compelled, obligated to cancel somebody.

That tends to be the narrative.

And so I think the logical conclusion is, oh, social media

allows for this to happen.

Whereas normally someone would just be standing on the corner shouting

or calling lawyers that don't have faith in them, and you've got the Erin

Brockovich model that turns into a movie.

But that's a rare case of this lone woman who's got this idea in mind about how a

big institution is doing wrong or somebody is doing wrong in the world and then can

leverage the big institution, excuse me.

But the way that you describe it is that the elites are leading this shift.

So what is the role of the public in it?

Just to give it a concrete example, if, for instance, no one tweeted or

commented on me, too, or no one tweeted or commented about some ill behavior

of some, I don't know, university faculty member or business person,

would the elite have come down on them?

Marc Andreessen: Anyway, what's happening?

Based on what I've seen over the years, there is so much astroturfing right now.

There are entire categories of people who are paid to do this.

Some of them we call journalists, some of them we call activists,

some of them we call NGO nonprofit.

Some of them we call university professors, some of them we

call grad students, whatever, they're paid to do this.

I don't know if you've ever looked into the misinformation industrial complex?

There's this whole universe of basically these funded groups

that basically do misinformation.

And they're constantly mounting these kinds of attacks.

They're constantly trying to gin up this kind of basically panic

to cause somebody to get fired.

Andrew Huberman: So it's not a grassroots--

Marc Andreessen: --No. It's the opposite of grassroots.

No.

Almost always going to trace these things back.

It was a journalist, it was an activist, it was a public figure of some kind.

These are entrepreneurs in a sort of a weird way.

Basically their job, mission calling, is all wrapped up together

like they're true believers, but they're also getting paid to do it.

And there's a giant funding, I mean, there's a very large funding

complex for this coming from certain high profile people who put

huge amounts of money into this.

Andrew Huberman: Is this well known?

Marc Andreessen: Yes.

Well, it is in my world.

So this is what the social media companies have been on the receiving

end of for the last decade.

It's basically a political media activism complex with very deep pockets behind it.

And you've got people who basically, literally have people who sit all day

and watch the TV network on the other side or watch the Twitter feeds on the

other side, and they basically wait.

It's like every politician, this has been the case for a long time now.

Every politician who goes out and gives stump speeches, you'll see there's always

somebody in the crowd with a camcorder or now with a phone recording them.

And that's somebody from the other campaign who's paid somebody to

just be there and record every single thing the politician says.

So that when a Mitt Romney says, whatever, the 47% thing, they've

got it on tape, and then they clip it, and they try to make it viral.

And again, look, these people believe what they're doing.

I'm not saying it's even dishonest.

Like, these people believe what they're doing.

They think they're fighting a holy war.

They think they're protecting democracy.

They think they're protecting civilization.

They think they're protecting whatever it is they're protecting.

And then they know how to use the tools, and so they know how

to try to gin up the outrage.

And then, by the way, sometimes it works in social cascades.

Sometimes it works, sometimes it doesn't.

Sometimes they cascade, sometimes they don't.

But if you follow these people on Twitter, this is what they do every day.

They're constantly trying to, like, light this fire.

Andrew Huberman: I assume that it was really bottom up, but it sounds like

it's sort of middle level, and that it captures the elites, and then the

thing takes on a life of its own.

Marc Andreessen: By the way, it also intersects with the trust and safety

groups at the social media firms who are responsible for figuring out who gets

promoted and who gets banned across this.

And you'll notice one large social media company has recently changed

hands and has implemented a different kind of set of trust and safety.

And all of a sudden, a different kind of boycott movement has

all of a sudden started to work that wasn't working before that.

And another kind of boycott movement is not working as well anymore.

And so, for sure, there's an intermediation happening.

Look, the stuff that's happening in the world today is being intermediated

through social media, because social media is the defining media of our time.

But there are people who know how to do this and do this for a living.

No, I view very much the cancellation wave, like, this whole thing, it's an

elite phenomenon, and when it appears to be a grassroots thing, it's either

grassroots among the elites, which is possible because there's a fairly

large number of people who are signed up for that particular crusade, but

there's also a lot of astroturfing that's taking place inside that.

The question is, okay, at what point does the population at

large get pulled into this?

And maybe there are movements, certain points in time where they

do get pulled in, and then maybe later they get disillusioned.

And so then there's some question there.

And then there's another question of like, well, if the population at

large is going to decide what these movements are, are they going to be the

same movements that the elites want?

And how are the elites going to react when the population

actually fully expresses itself?

Like I said, there's a feedback loop between these where the more extreme

the elites get, they tend to push the population to more extreme views

on the other side and vice versa.

So it ping pongs back and forth.

And so, yeah, this is our world.

Andrew Huberman: Yeah, this explains a lot.

Marc Andreessen: I want to make sure that Schellenberger, Matt Taibbi, a bunch

of these guys have done a lot of work.

If you just look into what's called the misinformation industrial complex,

you'll find a network of money and power that is really quite amazing.

Andrew Huberman: I've seen more and more Schellenberger showing up.

Marc Andreessen: Right.

And he's just, look, he's just on this stuff.

He, and just, they're literally just like tracking money.

It's very clear how the money flows, including a remarkable amount of money

out of the government, which is, of course, in theory, very concerning.

Andrew Huberman: Very interesting.

Marc Andreessen: The government should not be funding programs that take

away people's constitutional rights.

And yet somehow that is what's been happening.

Andrew Huberman: Very interesting.

I want to make sure that I hear your ideas about why the decline

in confidence in institutions is not necessarily problematic.

Is this going to be a total destruction, burning down of the

forest that will lead to new life?

Is that your view?

Marc Andreessen: Well, so this is the thing.

And look, there's a question if you're, there's a couple of questions in here,

which is like, how bad is it really?

How bad are they?

Right. And I think they're pretty bad.

A lot of them are actually pretty bad.

So that's one big question.

And then, yeah, look, the other question is like, okay, if the institution has gone

bad or a group of elites have gone bad, it's this wonderful word, reform, right?

Can they be reformed?

And everybody always wants to reform everything, and yet somehow nothing

ever quite ever gets reformed.

And so people are trying to reform housing policy in the Bay Area for

decades, and we're not building.

We're building fewer houses than ever before.

So somehow reform movements seem to lead to just more bad stuff.

But anyway, yeah.

So if you have an existing institution, can it be reformed?

Can it be fixed from the inside?

What's happened in universities?

There are professors at Stanford as an example, who very much

think that they can fix Stanford.

Like, I don't know what you think.

It doesn't seem like it's going in productive directions right now.

Andrew Huberman: Well, I mean, there are many things about Stanford

that function extremely well.

It's a big institution.

It's certainly got its issues like any other place.

They're also my employer, Marc's giving me some interesting looks.

He wants me to get a little more vocal.

Marc Andreessen: I didn't mean to put you on the spot.

Yeah.

Andrew Huberman: I mean, one of the things about being a researcher

at a big institution like Stanford is, well, first of all, it meets

the criteria that you described.

Know, you look to the left, you look to the right or anywhere above or

below you, and you have excellence.

Right?

I mean, I've got a Nobel Prize winner below me whose daddy also won

a Nobel Prize, and his scientific offspring is likely to win.

I mean, it inspires you to do bigger things than one

ordinarily would, no matter what.

So there's that, and that's great.

And that persists.

There's all the bureaucratic red tape about trying to get things done and how

to implement decisions is very hard, and there are a lot of reasons for that.

And then, of course, there are the things that many people are aware of.

There are public accusations about people in positions of great leadership,

and that's getting played out.

And the whole thing becomes kind of overwhelming and a little bit

opaque when you're just trying to run your lab or live your life.

And so I think one of the reasons for this lack of reform that you're

referring to is because there's no position of reformer, right?

So deans are dealing with a lot of issues.

Provosts are dealing with a lot of issues.

Presidents are dealing with a lot of issues, and then some in some cases.

And so we don't have a dedicated role of reformer, someone to go in and

say, listen, there's just a lot of fat on this and we need to trim it

or we need to create this or do that.

There just isn't a system to do that.

And that's, I think in part, because universities are built on old systems,

and it's like the New York subway.

It's amazing i t still works as well as it does, and yet it's

got a ton of problems also.

Marc Andreessen: So the point, we could debate the university specifically,

but the point is like, look, if you do think institutions are going bad, and

then you have to make it number one.

You have to figure out if you think institutions are going bad.

The population largely does think that at the very least, the people

who run institutions ought to really think hard about what that means.

Andrew Huberman: But people still strive to go to these places.

And I still hear from people who, for instance, did not go to

college, are talking about how a university degree is useless.

They'll tell you how proud they are that their son or daughter is going

to Stanford or is going to UCLA or is going to Urbana Champaign.

I mean, it's almost like, to me, that's always the most shocking contradiction,

is like, these institutions don't matter.

But then when people want to hold up a card that says why their

kid is great, it's not about how many pushups they can do or that

they started their own business.

Most of the time it's they're going to this university.

And I think, well, what's going on here?

Marc Andreessen: So do you think the median voter in the United States

can have their kid go to Stanford?

Andrew Huberman: No.

Marc Andreessen: Do you think the median voter in the United States

could have their kid admitted to Stanford, even with a perfect SAT?

Andrew Huberman: No, no.

In this day and age, the competition is so fierce that it requires more.

Marc Andreessen: Yeah.

So first of all, again, we're dealing here.

Yes.

We're dealing with a small number of very elite institutions.

People may admire them or not.

Most people have no connectivity to them whatsoever.

In the statistics, in the polling, universities are not doing well.

The population at large, yeah, they may have fantasies about their

kid going to Stanford, but the reality of it is they have a very

collapsing view of these institutions.

So anyway, this actually goes straight to the question of alternatives then, right?

Which is like, okay, if you believe that there's collapsing faith in the

institutions, if you believe that it is merited, at least in some ways, if

you believe that reform is effectively impossible, then you are faced...

We could debate each of those, but the population at large

seems to believe a lot of that.

Then there's a question of like, okay, can it be replaced?

And if so, are you better off replacing these things basically,

while the old things still exist?

Or do you actually need to basically clear the field to be

able to have the new thing exist?

The universities are a great case study of this because of

how student loans work, right?

And the way student loans work is to be an actual competitive university

and compete, you need to have access to federal student lending.

Because if you don't, everybody has to pay out of pocket.

And it's completely out of reach for anybody other than a certain class of

either extremely rich or foreign students.

So you need access to a federal student loan facility.

To get access to a federal student loan facility, you need

to be an accredited university.

Guess who runs the accreditation council?

Andrew Huberman: I don't know.

Marc Andreessen: The existing universities, right?

So it's a self laundering machine.

Like they decide who the new universities are.

Guess how many new universities get accredited, each year to be able...

Andrew Huberman: Zero.

Marc Andreessen: Zero, right?

And so as long as that system is in place, and as long as they have the government

wired the way that they do, and as long as they control who gets access to

federal student loan funding, of course there's not going to be any competition.

Of course there can't be a new institution that's going to be able to get to scale.

It's not, not possible.

And so if you actually wanted to create a new system that was better

in, you know, I would argue dozens or hundreds of ways, it could obviously be

better if you were starting it today.

It probably can't be done as long as the existing institutions are actually intact.

And this is my counter to Martin, which is like, yeah, look, if we're going

to tear down the old, there may be a period of disruption before we get to

the new, but we're never going to get to the new if we don't tear down the old.

Andrew Huberman: When you say counter to Martin, you're talking about

the author of Revolt of the Public

? Marc Andreessen: Yeah, Martin Gurri.

What Martin Gurri says is like, look, he said basically as follows, the elites

deserve contempt, but the only thing worse than these elites that deserve

contempt would be no elites at all.

And he basically says on the other side of the destruction of the elites

and the institutions is nihilism.

You're basically left with nothing.

And by the way, there is a nihilistic streak.

I mean, there's a nihilistic streak in the culture and the politics today.

There are people who basically would just say, yeah, just tear

the whole system down without any particular plan for what follows.

And so I think he makes a good point and that you want to be careful that you

actually have a plan on the other side that you think is actually achievable.

But again, the counterargument to that is if you're not willing

to actually tear down the old, you're not going to get to the new.

Now, what's interesting, of course, is this is what happens

every day in business, right?

So the entire way, how do you know that the capitalist system works?

The way that you know is that the old companies, when they're no longer like

the best at what they do, they get torn down and then they ultimately die and

they get replaced by better companies.

Andrew Huberman: Yeah, I haven't seen a Sears in a while.

Marc Andreessen: Exactly.

And we know what's so interesting is we know in capitalism, in a

market economy, we know that's the sign of health, that's the sign of

how the system is working properly.

And in fact, we get actually judged by antitrust authorities

in the government on that basis.

It's like the best defense against antitrust charges is no, people

are coming to kill us and they're doing a really good job of it.

That's how we know we're doing our job.

And in fact, in business we are specifically, it is specifically

illegal for companies in the same industry to get together and plot

and conspire and plan and have things like these accreditation bureaus.

If I created the equivalent in my companies of the kind of accreditation

bureau that the universities have, I'd get sent straight to federal prison

and a trust violation Sherman Act.

Straight to prison.

People have been sent to prison for that.

So in the business world, we know that you want everything

subject to market competition.

We know that you want creative destruction.

We know that you want replacement of the old with superior new.

It's just once we get outside of business, we're like, oh, we don't want any of that.

We want basically stagnation and log rolling and basically institutional

incestuous, like entanglements and conflicts of interest as

far as the eye can see, and then we're surprised by the results.

Andrew Huberman: So let's play it out as a bit of a thought experiment.

So let's say that one small banding together of people who want to start a new

university where there is free exchange of open ideas, where unless somebody has

egregious behavior, violent behavior, truly sexually inappropriate behavior

against somebody that is committing a crime, they're allowed to be there.

They're allowed to be a student or a faculty member or administrator.

And let's just say this accreditation bureau allowed student loans for

this one particular university.

Or let's say that there was an independent source of funding for that university

such that students could just apply there.

They didn't need to be part of this elite, accredited group, which sounds

very mafia-like, frankly, not necessarily violent, but certainly coercive in

the way that it walls people out.

Let's say that then there were 20 or 30 of those or 40 of those.

Do you think that over time, that model would overtake the existing model?

Marc Andreessen: Isn't it interesting that those don't exist?

Remember Sherlock Holmes, The Dog that Didn't Bark

? Andrew Huberman: It is interesting that they don't exist.

Marc Andreessen: Right. So there's two possibilities.

One is like, nobody wants that, which I don't believe.

And then the other is like, the system is wired in a way that

will just simply not allow it.

And you did a hypothetical in which the system would allow it.

And my response to that is, no, of course the system won't allow that.

Andrew Huberman: Or the people that band together have enough money or get enough

resources to say, look, we can afford to give loans to 10,000 students per year.

10,000 isn't a trivial number when thinking about the size of a university.

And most of them hopefully will graduate in four years and there'll be a turnover.

Do you think that the great future innovators would tend to orient toward

that model more than they currently do toward the traditional model?

What I'm trying to get back to here is how do you think that the current model

thwarts innovation, as well as maybe some ways that it still supports innovation?

Certainly cancellation and the risk of cancellation from the way that we framed

it earlier, is going to discourage risk takers of the category of risk takers

that take risk in every domain that really like to fly close to the sun

and sometimes into the sun or are--

Marc Andreessen: --Doing research that is just not politically palatable.

Andrew Huberman: Right, that we can't even talk about on this podcast, probably

without causing a distraction of what we're actually trying to talk about.

Marc Andreessen: That gives up the whole game right there.

Exactly.

Andrew Huberman: I keep a file, and it's a written file because I'm afraid

to put it into electronic form of all the things that I'm afraid to talk

about publicly because I come from a lineage of advisors where all three died

young, and I figure, if nothing else, I'll die, and then I'll make it into

the world and let's say 510 years, 20 years, and if not, I know a certainty

I'm going to die at some point, and then we'll see where all those issues stand.

In any event--

Marc Andreessen: --is that list getting l onger over time or shorter?

Andrew Huberman: Oh, it's definitely getting longer.

Marc Andreessen: Isn't that interesting?

Andrew Huberman: Yeah, it's getting much longer.

I mean, there are just so many issues that I would love to explore on this

podcast with experts and that I can't explore, just even if I had a panel

of them, because of the way that things get soundbited and segmented

out and taken out of context, it's like the whole conversation is lost.

And so, unfortunately, there are an immense number of equally interesting

conversations that I'm excited to have, but it is a little disturbing.

Marc Andreessen: Do you remember Lysenkoism?

Andrew Huberman: No.

Marc Andreessen: Famous in the history of the Soviet Union.

This is the famous thing.

So there was a geneticist named Lysenko.

Andrew Huberman: That's why it sounds familiar, but I'm not calling to--

Marc Andreessen: --Well, he was the guy who did communist genetics, the field

of genetics, the Soviets did not approve of the field of genetics because, of

course, they believed in the creation of the new man and total equality,

and genetics did not support that.

And so if you were doing traditional genetics, you were going to know, at

the very least be fired, if not killed.

And so this guy Lysenko stood up and said, oh, I've got Marxist genetics, right?

I've got, like a whole new field of genetics that basically

is politically compliant.

And then they actually implemented that in the agriculture

system of the Soviet Union.

And it's the origin of one of the big reasons that the Soviet Union

actually fell, which was they ultimately couldn't feed themselves.

Andrew Huberman: So create a new notion of biology as it relates to genetics.

Marc Andreessen: Politically correct biology, right?

They not only created it, they taught it, they mandated it, they required it, and

then they implemented it in agriculture.

Andrew Huberman: Interesting.

Marc Andreessen: I never understood.

There was a bunch of things in history I never understood until the

last decade, and that's one of them.

Andrew Huberman: Well, I censor myself at the level of deleting certain things,

but I don't contort what I do talk about.

So I tend to like to play on lush, open fields.

Just makes my life a lot easier.

Marc Andreessen: But this goes to the rot.

This goes to the rot, and I'll come back to your question, but this goes

to the rot in the existing system, which is, by the way, I'm no different.

I'm just like you.

Like, I'm trying not to light myself on fire either.

But the rot in the existing system, and by system, I mean the institutions

and the elites, the rot is that the set of things that are no longer allowed.

I mean, that list is obviously expanding over time, and that's real, historically

speaking, that doesn't end in good places.

Andrew Huberman: Is this group of a particular generation that

we can look forward to the time when they eventually die off.

Marc Andreessen: It's a third of the Boomers plus the Millennials.

Andrew Huberman: So, got a while.

Marc Andreessen: Good news, bad news.

Gen X is weird, right?

I'm Gen X.

Gen X is weird because we kind of slipped in the middle.

We were kind of the, I don't know how to describe it.

We were the kind of non-political generation kind of sandwiched between

the Boomers and the Millennials.

Gen Z is a very, I think, open question right now which way they go.

I could imagine them being actually much more intense than the

Millennials on all these issues.

I could also imagine them reacting to the Millennials

and being far more open minded.

Andrew Huberman: We don't know which way it's going to go.

Marc Andreessen: Yeah, it's going to go.

It might be different groups of them.

Andrew Huberman: I'm Gen X also, I'm 47, you're...?

Marc Andreessen: 52.

Andrew Huberman: So I grew up with some John Hughes films and so where the

jocks and the hippies and the punks, and were all divided and they were

all segmented, but then it all sort of mishmashed together a few years later.

And I think that had a lot to do with, like you said, the sort of

apolitical aspect of our generation.

Marc Andreessen: The Gen X just knew the Boomers were nuts, right?

Like, one of the great sitcoms of the era was Family Ties , right?

With the character Michael P.

Keaton.

And he was just like, this guy is just like, yeah, my Boomer

hippie parents are crazy.

I'm just going to go into business and actually do something productive.

There was something iconic about that character in our culture.

And people like me were like, yeah, obviously you go into business, you

don't go into political activism.

And then it's just like, man, that came whipping back around

with the next generation.

So just to touch real quick on the university thing.

So, look, there are people trying to do, and I'm actually going to do a

thing this afternoon with the University of Austin, which is one of these.

And so there are people trying to do new universities.

Like, I would say it's certainly possible.

I hope they succeed.

I'm pulling for them.

I think it'd be great.

I think it'd be great if there w ere a lot more of them.

Andrew Huberman: Who founded this university?

Marc Andreessen: This is a whole group of people.

I don't want to freelance on that because I don't know originally who the idea was--

Andrew Huberman: --University of Austin, not UT Austin.

Marc Andreessen: Yeah. So this is not UT Austin.

It's called the University of Austin.

Or they call it.

I think it's UATX?

And it's a lot of very sharp people associated with it.

They're going to try, very much exactly like what you described.

They're going to try to do a new one.

I would just tell you the wall of opposition that they're

up against is profound.

And part of it is economic, which is can they ever get access

to federal student lending?

And I hope that they can, but it seems nearly inconceivable the

way the system is rigged today.

And then the other is just like they already have come under, I mean,

anybody who publicly associates with them who is in traditional academia

immediately gets lit on fire, and there's, you know, cancellation campaigns.

So they're up against a wall of social ostracism.

Andrew Huberman: Wow.

Marc Andreessen: They're up against a wall of press attacks.

They're up against a wall of people just like doing the thing, pouncing on,

anytime anybody says anything, they're going to try to burn the place down.

Andrew Huberman: This reminds me of Jerry Springer episodes and Geraldo

Rivera episodes where it's like if a teen listened to Danzig or Marilyn

Manson type music or Metallica, that they were considered a devil worshiper.

Now we just laugh, right?

We're like, that's crazy, right?

People listen to music with all sorts of lyrics and ideas and looks.

That's crazy.

But there were people legitimately sent to prison.

I think it was a West Memphis three, right?

These kids out in West Memphis that looked different, acted different,

were accused of murders that eventually was made clear they clearly didn't

commit, but they were in prison because of the music they listened to.

I mean, this sounds very similar to that.

And I remember seeing bumpersickers, Free the West Memphis Three!

And I thought this was some crazy thing.

And you look into it and this isn't, it's a little bit niche,

but these are real lives.

And there was an active witch hunt for people that looked

different and acted different.

And yet now we're sort of in this inverted world where on the one hand we're all

told that we can express ourselves however we want, but on the other

hand, you can't get a bunch of people together to take classes where they learn

biology and sociology and econ in Texas.

Wild.

Marc Andreessen: Yes.

Well, so the simple explanation is this is Puritanism, right?

So this is the original American Puritanism that just works

itself out through the system in different ways at different times.

There's a religious phenomenon in America called the Great Awakenings.

There will be these periods in American history where there's

basically religiosity fades and then there will be this snapback

effect where you'll have basically this frenzy basically, of religion.

In the old days, it would have been tent revivals and people speaking

in tongues and all this stuff.

And then in the modern world, it's of the form that we're living through right now.

And so, yeah, it's just basically these waves of sort of American religious, and

remember, religion in our time, religious impulses in our time don't get expressed

because we live in more advanced times.

We live in scientifically informed times.

And so religious impulses in our time don't show up as overtly religious.

They show up in a secularized form, which, of course, conveniently, is

therefore not subject to the First Amendment separation of church and state.

As long as the church is secular, there's no problem.

But we're acting out these kind of religious scripts over and over

again, and we're in the middle of another religious frenzy.

Andrew Huberman: There's a phrase that I hear a lot, and I don't

necessarily believe it, but I want your thoughts on it, which is,

"the pendulum always swings back."

Marc Andreessen: Yeah, not quite.

[LAUGHS]

Andrew Huberman: So that's how I feel, too, because--

Marc Andreessen: --Boy, that would be great.

Andrew Huberman: Take any number of things that we've talked about, and,

gosh, it's so crazy the way things have gone with institutions, or it's

so crazy the way things have gone with social media, or it's so crazy, fill

in the blank and people will say, well, the pendulum always swings back like

it's the stock market or something.

After every crash, there'll be an eventual boom and vice versa.

Marc Andreessen: By the way, that's not true either.

Most stock markets we have are, of course, survivorship.

It's all survivorship.

Everything is survivor.

Everything you just said is obviously survivorship bias.

Right.

So if you look globally, most stock markets, over time crash

and burn and never recover.

The American stock market hasn't always recovered.

Andrew Huberman: I was referring to the American stock market.

Marc Andreessen: Globally, b ut the reason everybody refers to the

American stock market is because it's the one that doesn't do

that, the other 200 or whatever, crash and burn and never recover.

Let's go check in on the Argentina stock market right now.

I don't think it's coming back anytime soon.

Andrew Huberman: My father is Argentine and immigrated to the US in the 1960s,

so he would definitely agree with you.

Marc Andreessen: Yeah.

When their stocks crash, they don't come back.

And then Lysenkoism, like, the Soviet Union never recovered from

Lysenkoism, it never came back.

It led to the end of the country, you know, literally.

The things that took down the Soviet Union were oil and wheat.

And the wheat thing, you can trace the crisis back to Lysenkoism.

No, look, pendulum swings back is true only in the cases where the

pendulum swings back, everybody just conveniently forgets all the other

circumstances where that doesn't happen.

One of the things people, you see this in business also, people have a really

hard time confronting really bad news.

I don't know if you've noticed that.

I think every doctor who's listening right now is like, yeah, no shit.

But have you seen in business, there are situations, that Star

Trek , remember Star Trek ? The Kobayashi Maru simulator, right?

So the big lesson to become a Star Trek captain is you had to go through the

simulation called the Kobayashi Maru, and the point was, there's no way to win.

It's a no win scenario.

And then it turned out like, Captain Kirk was the only

person to ever win the scenario.

And the way that he did it was he went in ahead of time and hacked the simulator.

It was the only way to actually get through.

And then there was a debate whether to fire him or make him a captain.

So they made him a captain.

You know, the problem is, in real life, you do get the

Kobayashi Maru on a regular basis.

Like, there are actual no win situations that you can't work your way out of.

And as a leader, you can't ever cop to that, right?

Because you have to carry things forward, and you have to look for

every possible choice you can.

But every once in a while, you do run into a situation where

it's really not recoverable.

And at least I've found people just cannot cope with that.

What happens is they basically, then they basically just exclude it from

their memory that it ever happened.

Andrew Huberman: I'm glad you brought up simulators, because I want to make sure

that we talk about the new and emerging landscape of AI artificial intelligence.

And I could try and smooth our conversation of a moment ago with this

one by creating some clever segue, but I'm not going to, except I'm going to ask, is

there a possibility that AI is going to remedy some of what we're talking about?

Let's make sure that we earmark that for discussion a little bit later.

But first off, because some of the listeners of this podcast

might not be as familiar with AI as perhaps they should be.

We've all heard about artificial intelligence.

People hear about machine learning, etc.

But it'd be great if you could define for us what AI is.

People almost immediately hear AI and think, okay, robots taking over.

I'm going to wake up, and I'm going to be strapped to the bed and my organs

are going to be pulled out of me.

The robots are going to be in my bank account.

They're going to kill all my children and dystopia for most.

Clearly, that's not the way it's going to go if you believe that machines

can augment human intelligence, and human intelligence is a good thing.

So tell us what AI is and where you think it can take us, both good and bad.

Marc Andreessen: So, there was a big debate when the computer was first

invented, which is in the 1930s, 1940s, people like Alan Turing and

John von Neumann and these people.

And the big debate at the time was because they knew they wanted to build computers.

They had the basic idea, and there had been, like, calculating machines before

that, and there had been these looms that you basically programmed to punch cards.

And so there was a prehistory to computers that had to do with building sort of

increasingly complex calculating machines.

So they were kind of on a track, but they knew they were going to

be able to build, they called it a general purpose computer that could

basically, you could program, in the way that you program computers today.

But they had a big debate early on, which is, should the fundamental

architecture of the computer be based on either A, like calculating machines,

like cache registers and looms and other things like that, or should it

be based on a model of the human brain?

And they actually had this idea of computers modeled on the human

brain back then, and this is this concept of so called neural networks.

And it's actually fairly astonishing from a research standpoint.

The original paper on neural networks actually was published in 1943.

So they didn't have our level of neuroscience, but they actually knew

about the neuron, and they actually had a theory of neurons interconnecting

and synapses and information processing in the brain even back then.

And a lot of people at the time basically said, you know what?

We should basically have the computer from the start be modeled after the

human brain, because if the computer could do everything that the human

brain can do, that would be the best possible general purpose computer.

And then you could have it do jobs, and you could have it create

art, and you could have it do all kinds of things like humans can do.

It turns out that didn't happen.

In our world, what happened instead was the industry went in the other direction.

It went basically in the model of the calculating machine or the cash register.

And I think, practically speaking, that kind of had to be the case, because

that was actually the technology that was practical at the time.

But that's the path and so what we all have experiences with, up to and including

the iPhone in our pocket, is computers built on that basically calculating

machine model, not the human brain model.

And so what that means is computers, as we have come to understand

them, they're basically like mathematical savants at best.

So they're really good at doing lots of mathematical calculations.

They're really good at executing these extremely detailed computer programs.

They're hyper literal.

One of the things you learn early when you're a programmer is, as the

human programmer, you have to get every single instruction you give

the computer correct because it will do exactly what you tell it to do.

And bugs in computer programs are always a mistake on the part of the programmer.

Interesting.

You never blame the computer.

You always blame the programmer because that's the nature of the

thing that you're dealing with.

Andrew Huberman: One downscore off and the whole thing--

Marc Andreessen: --Yeah, and it's the programmer's fault.

And if you talk to any programmer, they'll agree with this.

They'll be like, yeah, if there's a problem, it's my fault.

I did it.

I can't blame the computer.

The computer has no judgment.

It has no ability to interpret, synthesize, develop an independent

understanding of anything.

It's literally just doing what I tell it to do step by step.

So for 80 years we've had this, just this very kind of hyper

literal kind of model computers.

Technically, these are what are called von Neumann machines, based after

the mathematician John von Neumann.

They run in that way, and they've been very successful and very important,

and our world has been shaped by them.

But there was always this other idea out there, which is, okay, how about

a completely different approach, which is based much more on how the

human brain operates, or at least our kind of best understanding of

how the human brain operates, right?

Because those aren't the same thing.

It basically says, okay, what if you could have a computer

instead of being hyper literal?

What if you could have it actually be conceptual and creative and

able to synthesize information and able to draw judgments and able

to behave in ways that are not deterministic but are rather creative?

And the applications for this, of course, are endless.

And so, for example, the self-driving car, the only way that you cannot

program a computer with rules to make it a self-driving car, you

have to do what Tesla and Waymo and these other companies have done.

Now you have to use, right, you have to use this other architecture,

and you have to basically teach them how to recognize objects in

images at high speeds, basically the same way the human brain does.

And so those are so called neural networks running inside.

Andrew Huberman: So, essentially, let the machine operate based on priors.

We almost clipped a boulder going up this particular drive, and so therefore,

this shape that previously the machine didn't recognize as a boulder, it now

introduces to its catalog of boulders.

Is that a good example?

Marc Andreessen: Let's even make it even starker for a self-driving car.

There's something in the road.

Is it a small child or a plastic shopping bag being blown by the wind?

Very important difference.

If it's a shopping bag, you definitely want to go straight through it, because

if you deviate off course, you're going to make a fast, it's the same

challenge we have when we're driving.

You don't want to swerve to avoid a shopping bag because you might hit

something that you didn't see on the side.

But if it's a small child for sure you want to swerve, right?

But in that moment, small children come in different shapes and descriptions and

are wearing different kinds of clothes.

Andrew Huberman: They might tumble onto the road the same way a bag would tumble.

Marc Andreessen: Yeah, they might look like they're tumbling.

And by the way, they might be wearing a Halloween mask.

Right.

They might not have a recognizable human face.

It might be a kid with one leg.

You definitely want to not hit those.

This is what basically we figured out is you can't apply the rules based

approach of a Von Neumann machine to basically real life and expect the

computer to be in any way understanding or resilient, to change to basically

things happening in real life.

And this is why there's always been such a stark divide between what the

machine can do and what the human can do.

And so, basically, what's happened is in the last decade, that second type

of computer, the neural network based computer, has started to actually work.

It started to work, actually, first, interestingly, in vision, recognizing

objects and images, which is why the self-driving car is starting to work.

Andrew Huberman: Face recognition.

Marc Andreessen: Face recognition.

Andrew Huberman: I mean, when I started off in visual neuroscience,

which is really my original home in neuroscience, the idea that a computer

or a camera could do face recognition better than a human was like a very

low probability event based on the technology we had at the time, based

on the understanding of the face recognition cells and the fusiform gyrus.

Now, you would be smartest to put all your money on the machine.

You want to find faces in airports, even with masks on and at profile

versus straight on, machines can do it far better than almost all people.

I mean, they're the super recognizers.

But even they can't match the best machines.

Now, ten years ago, what I just said was the exact reverse, right?

Marc Andreessen: That's right, yeah.

So faces, handwriting, and then voice, being able to

understand voice just as a user.

If you use Google Docs, it has a built-in voice transcription.

They have sort of the best industry leading kind of voice transcription.

If you use a voice transcription in Google Docs, it's breathtakingly good.

You just speak into it and it just types what you're saying.

Andrew Huberman: Well, that's good, because in my phone, every once in

a while, I'll say I need to go pick up a f ew things and it'll say,

I need to pick up a few thongs.

And so Apple needs to get on board.

Whatever the voice recognition is that Google's using--

Marc Andreessen: --Maybe it knows you better than you think.

Andrew Huberman: [LAUGHS] That was not the topic I was avoiding discussing.

Marc Andreessen: No.

So that's on the list, right?

That's on your...

Actually, there's a reason, actually, why Google's so good and Apple is

not right now at that kind of thing.

And it actually goes to actually an ideological thing, of all things.

Apple does not permit pooling of data for any purpose, including

training AI, whereas Google does.

And Apple's just like, stake their brand on privacy.

And among that is sort of a pledge that they don't pool your data.

And so all of Apple's AI is like, AI that has to happen locally on your phone.

Whereas Google's AI can happen in the cloud.

Right? It can happen across pool data.

Now, by the way, some people think that that's bad because

they think pooling data is bad.

But that's an example of the shift that's happening in the industry right now,

which is you have this separation between the people who are embracing the new

way of training AIs and the people who basically, for whatever reason, are not.

Andrew Huberman: Excuse me, you say that some people think it's

bad because of privacy issues or they think it's bad because of the

reduced functionality of that AI.

Marc Andreessen: Oh, no. So you're definitely going to get...

there's three reasons AIs have started to work.

One of them is just simply larger data sets, larger amounts of data.

Specifically, the reason why objects and images are now, the reason

machines are now better than humans at recognizing objects, images or

recognizing faces is because modern facial recognition AIs are trained across

all photos on the Internet of people.

Billions and billions and billions of photos, right?

Unlimited number of photos of people on the Internet.

Attempts to train facial recognition systems.

Ten or 20 years ago, they'd be trained on thousands or tens of thousands of photos.

Andrew Huberman: So the input data is simply much m ore vast

. Marc Andreessen: Much larger.

This is the reason to get to the conclusion on this.

This is the reason why ChatGPT works so well.

One of the reasons ChatGPT works so well is it's trained

on the entire Internet of text.

And the entire Internet of text was not something that was available for

you to train an AI on until it came to actually exist itself, which is

new in the last, basically decade.

Andrew Huberman: So in the case of face recognition, I could see how

having a much larger input data set would be beneficial if the goal is

to recognize Marc Andreessen's face, because you are looking for signal to

noise against everything else, right?

But in the case of ChatGPT, when you're pooling all text on the internet and you

ask ChatGPT to, say, construct a paragraph about Marc Andreessen's prediction of

the future of human beings over the next ten years and the likely to be most

successful industries, give ChatGPT that.

If it's pooling across all text, how does it know what is

authentically Marc Andreessen's text?

Because in the case of face recognition, you've got a standard to work from a

verified image versus everything else.

In the case of text, you have to make sure that what you're starting with is

verified text from your mouth, which makes sense if it's coming from video.

But then if that video is deep faked, all of a sudden, what's true?

Your valid Marc Andreessen is in question.

And then everything ChatGPT is producing, that is then of question.

Marc Andreessen: So I would say there's a before and after thing here.

There's like a before ChatGPT and after GPT question, because the existence

of GPT itself changes the answer.

So before ChatGPT.

So the version you're using today is trained on data up till September 2021.

They're cut off with the training set.

Up till September 2021, almost all text on the Internet was written by a human being.

And then most of that was written by people under their own names.

Some of it wasn't, but a lot of it was.

And why do you know it's for me is because it was published in a magazine

under my name, or it's a podcast transcript and it's under my name.

And generally speaking, if you just did a search on what are things Marc

Andreessen has written and said, 90% plus of that would be correct,

and somebody might have written a fake parody article or something.

Like that.

But not that many people were spending that much time writing fake

articles about things that I said.

Andrew Huberman: Right now, so many people can pretend to be you.

Marc Andreessen: Exactly right.

And so, generally speaking, you can kind of get your arms around

the idea that there's a corpus of material associated with me.

Or by the way, same thing with you.

There's a corpus of YouTube transcripts and other, your academic papers

and talks you've given, and you can kind of get your hands around that.

And that's how these systems are trained.

They take all that data collectively, they put it in there.

And that's why this works as well as it does.

And that's why if you ask ChatGPT to speak or write like me or like you or like

somebody else, it will actually generally do a really good job because it has all

of our prior text in its training data.

That said, from here on out, this gets harder.

And of course, the reason this gets harder is because now we have AI that

can create text and we have AI that can create text at industrial scale.

Andrew Huberman: Is it watermarked as AI generated text?

Marc Andreessen: No.

Andrew Huberman: How hard would it be to do that?

Marc Andreessen: I think it's impossible.

I think it's impossible.

There are people who are trying to do that.

This is a hot topic in the classroom.

I was just talking to a friend who's got like a 14 year old kid in a class, and

there's like these recurring scandals.

Every kid in the class is using ChatGPT to write their essays or to help them write

their essays, and then the teacher is using one of, there's a tool that you can

use that purports to be able to tell you whether something was written by ChatGPT.

But it's like, only right like 60% of the time.

And so there was this case where the student wrote an essay where their

parent sat and watched them write the essay, and then they submitted it, and

this tool got the conclusion incorrect.

And then the student feels outraged because he got unfairly cheated.

But the teacher is like, well, you're all using the tool.

Then it turns out there's another tool that basically you feed in

text, and they call it a summarizer.

But what it really is is it's a cheating mechanism to basically

just shuffle the words around enough so that it sheds whatever

characteristics were associated with AI.

So, there's like an arms race going on in educational settings right

now around this exact question.

I don't think it's possible to do.

There are people working on the watermarking.

I don't think it's possible to do the watermarking.

And I think it's just kind of obvious why it's not possible to do that, which is

you can just read the output for yourself.

It's really good.

How are you actually going to tell the difference between that and

something that a real person wrote?

And then, by the way, you can also ask ChatGPT to write

in different styles, right?

So you can tell it, like, write in the style of a 15 year old.

You can tell it to write in the style of a non native English speaker.

Or if you're a non native English speaker, you can tell it to

write in the style of an English speaker, native English speaker.

And so the tool itself will help you evade.

I think there's a lot of people who are going to want to

distinguish, "real" versus fake.

I think those days are over.

Andrew Huberman: Genie's out of the bottle.

Marc Andreessen: Genie is completely out of the bottle.

And by the way, I actually think this is good.

This doesn't map to my worldview of how we use this technology

anyway, which we can come back to.

So there's that, and then there's the problem, therefore of the

so-called deep fake problem.

So then there's the problem of, like, deliberate basically, manipulation.

And that's like one of your many enemies, one of your increasingly

long list of enemies like mine, who basically is like, wow, I know

how I'm going to get him, right?

I'm going to use it to create something that looks like a Huberman

transcript and I'm going to have him say all these bad things.

Andrew Huberman: Or a video.

Marc Andreessen: Or a video, or a video.

Andrew Huberman: I mean, Joe Rogan and I were deep faked in a video.

I don't want to flag people to it, so I won't talk about what it was about, but

where it, for all the world looked like a conversation that we were having and

we never had that specific conversation.

Marc Andreessen: Yeah, that's right.

So that's going to happen for sure.

So what there's going to need to be is there need to be basically

registries where basically in your case, you will submit your legitimate

content into a registry under your unique cryptographic key, right.

And then basically there will be a way to check against that registry to

see whether that was the real thing.

And I think this needs to be done for sure.

For public figures, it needs to be done for politicians,

it needs to be done for music.

Andrew Huberman: What about taking what's already out there and being able to

authenticate it or not in the same way that many times per week, I get asked,

is this your account about a direct message that somebody got on Instagram?

And I always tell them, look, I only have the one account,

this one verified account.

Although now, with the advent of pay to play, verification makes it

a little less potent as a security blanket for knowing if it's not

this account, then it's not me.

But in any case, these accounts pop up all the time pretending to be me.

And I'm relatively low on the scale.

Not low, but relatively low on the scale to say, like a Beyonce

or something like that, who has hundreds of millions of followers.

So is there a system in mind where people could go in and

verify text, click yes or no.

This is me.

This is not me.

And even there, there's the opportunity for people to fudge, to eliminate things

about themselves that they don't want out there, by saying, no, that's not me.

I didn't actually say that.

Or create that.

Marc Andreessen: Yeah, no, that's right.

Technologically, it's actually pretty straightforward.

So the way to implement this technologically is with a public key.

It's called public key cryptography, which is the basis for how cryptography

information is secured in the world today.

And so basically, the implementation form of this would be, you would pick whatever

is your most trusted channel, and let's say it's your YouTube channel as an

example, where just everybody just knows that it's you on your YouTube channel

because you've been doing it for ten years or whatever, and it's just obvious.

And you would just publish in the about me page on YouTube, you

would just publish your public cryptographic key that's unique to you.

Right.

And then anytime anybody wants to check to see whether any piece

of content is actually you, they go to a registry in the cloud

somewhere, and they basically submit.

They basically say, okay, is this him?

And then they can basically see whether somebody with your public

key, you had actually certified that this was something that you made.

Now, who runs that registry is an interesting question.

If that registry is run by the government, we will call that the Ministry of Truth.

I think that's probably a bad idea.

If that registry is run by a company, we would call that basically the

equivalent of, like, a credit bureau or something like that.

Maybe that's how it happens.

The problem with that is that company now becomes hacking target number one,

right, of every bad person on Earth.

Because if anybody breaks into that company, they can

fake all kinds of things.

Andrew Huberman: They own the truth.

Marc Andreessen: Right. They own the truth.

And by the way, insider threat, also, their employees can monkey with it.

So you have to really trust that company.

The third way to do it is with a blockchain.

And so this, with the crypto blockchain technology, you could have

a distributed system, basically, a distributed database in the cloud

that is run through a blockchain.

And then it implements this cryptography and this certification process.

Andrew Huberman: What about quantum Internet?

Is that another way to encrypt these things?

I know most of our listeners are probably not familiar with quantum

Internet, but put simply, it's a way to secure communications on the Internet.

Let's just leave it at that.

It's sophisticated, and we'll probably do a whole episode about this at some point.

But maybe you have a succinct way of describing quantum Internet,

but that would be better.

And if so, please offer it up.

But is quantum Internet going to be one way to secure these

kinds of data and resources?

Marc Andreessen: Maybe in the future, years in the future?

We don't yet have working quantum computers in practice, so it's

not currently something you could do, but maybe in a decade or two?

Andrew Huberman: Tell me.

I'm going to take a stab at defining quantum Internet in one sentence.

It's a way in which if anyone were to try and peer in on a conversation on the

Internet, it essentially would be futile because of the way that quantum Internet

changes the way that the communication is happening so fast and so many times in any

one conversation, essentially changing the translation or the language so fast that

there's just no way to keep up with it.

Is that more or less accurate?

Marc Andreessen: Yeah, conceivably not yet, but someday.

Andrew Huberman: So, going back to AI, most people who

hear about AI are afraid of AI.

Marc Andreessen: Well?

Andrew Huberman: I think most people who aren't informed--

Marc Andreessen: --This goes back to our elites versus masses thing.

Andrew Huberman: Oh, interesting.

Well, I heard you say that, a his is from a really wonderful tweet thread that we

will link in the show note captions that you put out not long ago and that I've

read now several times, and that everyone really should take the time to read it.

Probably takes about 20 minutes to read it carefully and to think about

each piece, and I highly recommend it.

But you said, and I'm quoting here, "Let's address the fifth, the

one thing I actually agree with, which is AI will make it easier

for bad people to do bad things."

Marc Andreessen: First of all, there is a general freak out happening around AI.

I think it's primarily, it's one of these, again, it's an elite driven freak out.

I don't think the man in the street knows, cares, or feels one way or the other.

It's just not a relevant concept, and it probably just sounds like science fiction.

So I think there's an elite driven freak out that's happening right now.

I think that elite driven freak out has many aspects to it that I think

are incorrect, which is not surprising.

I would think that, given that.

I think the elites are incorrect about a lot of things, but I think

they're very wrong about a number of things they're saying about AI.

But that said, look, this is a very powerful new technology, right?

This is like a new general purpose thinking technology.

So what if machines could think?

And what if you could use machines that think, and what if you

could have them think for you?

There's obviously a lot of good that could come from that.

But also, people, look, criminals could use them to plan better crimes.

Terrorists could use them to plan better terror attacks and so forth.

And so these are going to be tools that bad people can use

to do bad things, for sure.

Andrew Huberman: I can think of some ways that AI could be

leveraged to do fantastic things.

Like in the realm of medicine, an AI pathologist perhaps, can scan 10,000

slides of histology and find the one micro tumor, cellular aberration, that

would turn into a full blown tumor, whereas the even mildly fatigued or

well rested human pathologists, as great as they come, might miss that.

And perhaps the best solution is for both of them to do it, and then

for the human to verify what the AI has found and vice versa, right?

Marc Andreessen: That's right.

Andrew Huberman: And that's just one example.

I mean, I can come up with thousands of examples where this would be wonderful.

Marc Andreessen: I'll give you another one, by the way, medicine.

So you're talking about an analytic result, which is good and important.

The other is like, the machines are going to be much better at bedside manner.

They're going to be much better at dealing with the patient.

And we already know there's already been a study.

There's already been a study on this.

So there was already a study done on this where there was a study team that

scraped thousands of medical questions off of an Internet forum, and then they

had real doctors answer the questions, and then they had basically GPT4 answer

the questions, and then they had another panel of doctors score the responses.

So there were no patients experimented on here.

This was a test contained within the medical world.

The judges, the panel of doctors who are the judges, scored the

answers in both factual accuracy and on bedside manner, on empathy.

And the GPT4 was equal or better on most of the factual questions

analytically, already, and it's not even a specifically trained medical AI, but

it was overwhelmingly better on empathy.

Andrew Huberman: Amazing,

Marc Andreessen: Right?

Do you treat patients directly in your work?

You don't?

Andrew Huberman: No, I don't.

We run clinical trials.

Marc Andreessen: Right.

Andrew Huberman: But I don't do any direct clinical work.

Marc Andreessen: I've no direct experience with this.

But from the surgeons, if you talk to surgeons or you talk to people who

train surgeons, what they'll tell you is surgeons need to have an emotional

remove from their patients in order to do a good job with the surgery.

The side effect of that, and by the way, look, it's a hell of a job to have to

go in and tell somebody that they're going to die or that they have so you're

never going to recover, they're never going to walk again or whatever it is.

And so there's sort of something inherent in that job where they need

to keep an emotional reserve from the patient to be able to do the job.

And it's expected of them as professionals.

The machine has no such limitation.

The machine can be as sympathetic as you want it to be for as

long as you want it to be.

It can be infinitely sympathetic.

It's happy to talk to you at four in the morning.

It's happy to sympathize with you.

And by the way, it's not just sympathizing with you in the way

that, oh, it's just making up words to lie to you to make you feel good.

It can also sympathize with you in terms of helping you through all

the things that you can actually do to improve your situation.

And so, boy, can you keep a patient actually on track with

a physical therapy program.

Can you keep a patient on track with a nutritional program?

Can you keep a patient off of drugs or alcohol?

And if they have a machine medical companion that's with them all the

time that they're talking to all the time, that's infinitely patient,

infinitely wise, infinitely loving, and it's just going to be there all the

time and it's going to be encouraging and it's going to be, you know, you

did such a great job yesterday, I know you can do this again today.

Cognitive behavioral therapy is an obvious fit here.

These things are going to be great at CBT and that's already starting.

You can already use ChatGPT as a CBT therapist if you want.

It's actually quite good at it.

There's, there's a universe here that's, it goes to what you said,

there's a universe here that's opening up, which is what I believe is it's

partnership between man and machine.

It's a symbiotic relationship, not an adversarial relationship.

And so the doctor is going to pair with the AI to do all the things

that you described, but the patient is also going to pair with the AI.

And I think this partnership that's going to emerge is going to lead,

among other things, to actually much better health outcomes.

Andrew Huberman: I've relied for so much of my life on excellent mentors from a

very young age, and still now, in order to make the best decisions possible with

the information I had, and rarely were they available at four in the morning

sometimes, but not on a frequent basis.

And they fatigue like anybody else, and they have their own stuff like anybody

else, baggage, events in their life, etc.

What you're describing is a sort of AI coach or therapist of sorts, that

hopefully would learn to identify our best self and encourage us to be our best self.

And when I say best self, I don't mean that in any kind of pop psychology way.

I could imagine AI very easily knowing how well I slept the night before and

what types of good or bad decisions I tend to make at 02:00 in the afternoon when

I've only had 5 hours of sleep, or maybe just less REM sleep the night before.

It might encourage me to take a little more time to think about something.

Might give me a little tap on the wrist through a device that no one else

would detect to refrain from something.

Marc Andreessen: Never going to judge you.

It's never going to be resentful.

It's never going to be upset that you didn't listen to it.

It's never going to go on vacation.

It's going to be there for you.

I think this is the way people are going to live.

It's going to start with kids, and then over time it's going to be adults.

I think the way people are going to live is they're going to have

a friend, therapist, companion, mentor, coach, teacher, assistant.

Or, by the way, maybe multiple of those.

It may be that we're actually talking about six, like, different personas

interacting, which is a whole 'nother possibility, but they're going to have--

Andrew Huberman: --A committee!

Marc Andreessen: A committee, yeah, exactly.

Actually different personas.

And maybe, by the way, when there are difficult decisions to be made in your

life, maybe what you want to hear is the argument among the different personas.

And so you're just going to grow up, you're just going to have this in

your life and you're going to always be able to talk to it and always

be able to learn from it and always be able to help it make, it's going

to be a symbiotic relationship.

I think it's going to be a much better way to live.

I think people are going to get a lot out of it.

Andrew Huberman: What modalities will it include?

So I can imagine my phone has this engine in it, this AI

companion, and I'm listening in headphones as I walk into work.

And it's giving me some, not just encouragement, some warnings, some

thoughts that things that I might ask Marc Andreessen today that I

might not have thought of and so on.

I could also imagine it having a more human form.

I could imagine it being tactile, having some haptic, so tapping to

remind me so that it's not going to enter our conversation in a way

that interferes or distracts you.

But I would be aware.

Oh, right.

Things of that sort.

I mean, how many different modalities are we going to allow these AI

coaches to approach us with?

And is anyone actually thinking about the hardware piece right now?

Because I'm hearing a lot about the software piece.

What does the hardware piece look like?

Marc Andreessen: Yeah, so this is where Silicon Valley is going to kick in.

So the entrepreneurial community is going to try all of those, right?

By the way, the big companies and startups are going to try all those.

And so obviously there's big companies that are working, the big

companies that have talked about a variety of these, including heads

up displays, AR, VR kinds of things.

There's lots of people doing voice.

Thing is, voice is a real possibility.

It may just be an earpiece.

There's a new startup that just unveiled a new thing where they actually project.

So you'll have like a pendant you wear on like a necklace, and it actually

projects, literally, it'll project images on your hand or on the table

or on the wall in front of you.

So maybe that's how it shows up.

Yeah.

There are people working on so-called haptic or touch based kinds of things.

There are people working on actually picking up nerve

signals, like out of your arm.

There's some science for being able to do basically like subvocalization.

So maybe you could pick up that way by bone conduction.

These are all going to be tried.

So that's one question is the physical form of it, and then the other

question is the software version of it, which is like, okay, what's the

level of abstraction that you want to deal with these things in right now?

It's like a question answer paradigm, so called chatbot, like, ask a question, get

an answer, ask a question, get an answer.

Well, you want that to go for sure to more of a fluid conversation.

You want it to build up more knowledge of who you are, and

you don't want to have to explain yourself a second time and so forth.

And then you want to be able to tell it things like, well, remind me this,

that, or be sure and tell me when X.

But then maybe over time, more and more, you want it actually deciding

when it's going to talk to you, right?

And when it thinks it has something to say, it says it,

and otherwise it stays silent.

Andrew Huberman: Normally, at least in my head, unless I make a

concerted effort to do otherwise, I don't think in complete sentences.

So presumably these machines could learn my style of fragmented internal dialogue.

And maybe I have an earpiece, and I'm walking in and I start hearing

something, but it's some advice, etc, encouragement, discouragement.

But at some point, those sounds that I hear in an earphone are very

different than seeing something or hearing something in the room.

We know this based on the neuroscience of musical perception

and language perception.

Hearing something in your head is very different.

And I could imagine at some point that the AI will cross a precipice where if

it has inline wiring to actually control neural activity in specific brain areas,

and I don't mean very precisely, even just stimulating a little more prefrontal

cortical activity, for instance, through the earpiece, a little ultrasound wave

now can stimulate prefrontal cortex in a non invasive way that's being

used clinically and experimentally, that the AI could decide that I need

to be a little bit more context aware.

This is something that is very beneficial for those listening that are trying to

figure out how to navigate through life.

It's like, you know, the context you're in and know the catalog of behaviors

and words that are appropriate for that situation and not, you know, this would

go along with agreeableness, perhaps, but strategic agreeableness, right.

Context is important.

There's nothing diabolical about that.

Context is important, but I could imagine the AI recognizing we're

entering a particular environment.

I'm now actually going to ramp up activity in prefrontal cortex a little bit in a

certain way that allows you to be more situationally aware of yourself and

others, which is great, unless I can't necessarily short circuit that influence,

because at some point, the AI is actually then controlling my brain activity

and my decision making and my speech.

I think that's what people fear is that once we cross that precipice that we

are giving up control to the artificial versions of our human intelligence.

Marc Andreessen: And look, I think we have to decide, we collectively,

and we as individuals, I think, have to decide exactly how to do that.

And this is the big thing that I believe about AI.

That's just a much more, I would say, practical view of the world than

a lot of the panic that you hear.

It's just like, these are machines.

They're able to do things that increasingly are like the things that

people can do in some circumstances.

But these are machines.

We built a machine, means we decide how to use the machines.

When we want the machines turned on, they're turned on, we want them

turned off, they're turned off.

I think that's absolutely the kind of thing that the individual person

should always be in charge of.

Andrew Huberman: Everyone was.

And I have to imagine some people are still afraid of CRISPR, of gene editing.

But gene editing stands to revolutionize our treatment of all sorts of disease,

you know, inserting and deleting particular genes in adulthood.

Not having to recombine in the womb.

A new organism is an immensely powerful tool.

And yet the Chinese scientist who did CRISPR on humans, this has been

done, actually did his postdoc at Stanford with Steve Quake, then

went to China, did CRISPR on babies.

Mutated something.

I believe it was one of the HIV receptors.

I'm told it was with the intention of augmenting human memory.

It had very little to do, in fact, with limiting susceptibility to

HIV per se, to do with the way that receptor is involved in human memory.

The world demonized that person.

We actually don't know what happened to them.

Whether or not they have a laboratory now or they're sitting in jail, it's unclear.

But in China and elsewhere, people are doing CRISPR on humans.

We know this.

It's not legal in the US and other countries, but it's happening.

Do you think it's a mistake for us to fear these technologies so much that we back

away from them and end up 10, 20 years behind other countries that could use it

for both benevolent or malevolent reasons?

Marc Andreessen: Yeah, the details matter.

So it's technology by technology.

But I would say there's two things you always have to think about in

these questions, I think, in terms of counterfactuals and opportunity cost.

CRISPR is an interesting one.

CRISPR manipulates the human genome.

Nature manipulates the human, like, in all kinds of ways.

[LAUGHS] Andrew Huberman: Yeah.

[LAUGHS]

Marc Andreessen: When you pick a spouse and you--

Andrew Huberman: --Have a child with that spouse--

Marc Andreessen: --Oh, boy--

Andrew Huberman: --You're doing genetic recombination.

Marc Andreessen: Yes, you are.

Quite possibly, if you're Genghis Khan, you're determining the future

of humanity by those mutations.

This is the old question of, basically, this is all state of

nature, state of grace, basically.

Is nature good?

And then therefore, artificial things are bad, which is kind of shot.

A lot of people have ethical views like that.

I'm always of the view that nature is a bitch and wants us dead.

Nature is out to get us, man.

Nature wants to kill us, right?

Like, nature wants to evolve all kinds of horrible viruses.

Nature wants plagues.

Nature wants to do weather.

Nature wants to do all kinds of stuff.

I mean, look, nature religion was the original religion, right?

Like, that was the original thing people worshiped.

And the reason was because nature was the thing that was out to get you right before

you had scientific and technological methods to be able to deal with it.

So, the idea of not doing these things, to me is just saying, oh,

we're just going to turn over the future of everything to nature.

And I think that there's no reason to believe that that leads in a

particularly good direction or that that's not a value neutral decision.

And then the related thing that comes from that is always this question around

what's called the precautionary principle, which shows up in all these conversations

on things like CRISPR, which basically is this principle that basically says, the

inventors of a new technology should be required to prove that it will not have

negative effects before they roll it out.

This, of course, is a very new idea.

This is actually a new idea in the 1970s.

It's actually invented by the German Greens.

The 1970s.

Before that, people didn't think in those terms.

People just invented things and rolled them out.

And we got all of modern civilization by people inventing

things and rolling them out.

The German Greens came up with the precautionary principle

for one specific purpose.

I'll bet you can guess what it is.

It was to prevent...?

Andrew Huberman: Famine?

Marc Andreessen: Nuclear power.

It was to shut down attempts to do civilian nuclear power.

And if you fast forward 50 years later, you're like, wow, that was a big mistake.

So what they said at the time was, you have to prove that nuclear

reactors are not going to melt down and cause all kinds of problems.

And, of course, as an engineer, can you prove that will never happen?

You can't.

You can't rule out things that might happen in the future.

And so that philosophy was used to stop nuclear power by the way, not

just in Europe, but also in the US and around much of the rest of the world.

If you're somebody who's concerned about carbon emissions, of course, this

is the worst thing that happened in the last 50 years in terms of energy.

We actually have the silver bullet answer to unlimited energy with zero

carbon emissions, nuclear power.

We choose not to do it.

Not only do we choose not to do it, we're actually shutting down the

plants that we have now in California.

We just shut down the big plant.

Germany just shut down their plants.

Germany is in the middle of an energy war with Russia that, we are informed,

is existential for the future of Europe.

Andrew Huberman: But unless the risk of nuclear power plant meltdown has

increased, and I have to imagine it's gone the other way, what is

the rationale behind shutting down these plants and not expanding?

Marc Andreessen: Because nuclear is bad.

Right. Nuclear is icky.

Nuclear has been tagged.

Andrew Huberman: It just sounds bad.

Nuclear.

Marc Andreessen: Yeah.

Andrew Huberman: Go nuclear.

Marc Andreessen: Well, so what happened?

Andrew Huberman: We didn't shut down postal offices and you hear go postal.

Marc Andreessen: So what happened was, so nuclear technology arrived

on planet Earth as a weapon, right?

So it arrived in the form of.

The first thing they did was in the middle of World War II.

The first thing they did was the atomic bomb they dropped on Japan.

And then there were all the debates that followed around

nuclear weapons and disarmament.

And there's a whole conversation to be had, by the way, about

that, because there's different views you could have on that.

And then it was in the.

Where they started to roll out civilian nuclear power.

And then there were accidents.

There was like, three Mile island melted down, and then Chernobyl melted

down in the Soviet Union, and then even recently, Fukushima melted down.

And so there have been meltdowns.

And so I think it was a combination of it's a weapon.

It is sort of icky scientists sometimes with the ick factor, right.

It glows green.

And by the way, it becomes like a mythical fictional thing.

And so you have all these movies of horrible supervillains powered by

nuclear energy and all this stuff.

Andrew Huberman: Well, the intro to the Simpsons, right.

Is the nuclear power plant and the three eyed fish and all the negative

implications of this nuclear power plant run by, at least in the Simpsons idiots.

And that is the dystopia, where people are unaware of just how bad it.

Marc Andreessen: Is and who owns the nuclear power plant.

Right. This evil capitalist.

Right.

So it's connected to capitalism.

Right.

Andrew Huberman: We're blaming Matt Gronig for the demise of a particular--

Marc Andreessen: --He certainly didn't help.

But it's literally, this amazing thing where if you're just like, thinking.

If you're just thinking rationally, scientifically, you're like, okay,

we want to get rid of carbon.

This is the obvious way to do it.

Okay, fun fact.

Richard Nixon did two things that really mattered on this.

So one is he defined in 1971 something called Project Independence, which

was to create 1000 new state of the art nuclear plants, civilian

nuclear plants, in the US by 1980.

And to get the US completely off of oil and cut the entire US energy grid

over to nuclear power, electricity, cut over to electric cars, the whole

thing, like, detach from carbon.

You'll notice that didn't happen.

Why did that not happen?

Because he also created the EPA and the Nuclear Regulatory Commission, which

then prevented that from happening.

Right.

And the Nuclear Regulatory Commission did not authorize a new nuclear

plant in the US for 40 years.

Andrew Huberman: Why would he hamstring himself like that?

Marc Andreessen: He got distracted by Watergate in Vietnam.

Andrew Huberman: I think Ellsberg just died recently, right?

The guy who released the Pentagon papers.

Marc Andreessen: Yeah. Andrew Huberman: So complicated.

Marc Andreessen: Yeah, exactly.

It's this thing.

He left office shortly thereafter.

He didn't have time to fully figure this out.

I don't know whether he would have figured it out or know.

Look, Ford could have figured it out.

Carter could have figured it out.

Reagan could have figured it out.

Any of these guys could have figured it out.

It's like the most obvious.

Knowing what we know today, it's the most obvious thing in the world.

The Russia thing is the amazing thing.

It's like Europe is literally funding Russia's invasion of Ukraine

by paying them for oil, right?

And they can't shut off the oil because they won't cut over to nuclear, right?

And then, of course, what happens?

Okay, so then here's the other kicker of what happens, right?

Which is they won't do nuclear, but they want to do renewables, right?

Sustainable energy.

And so what they do is they do solar and wind.

Solar and wind are not reliable because it sometimes gets dark out

and sometimes the wind doesn't blow.

And so then what happens is they fire up the coal plants, right?

And so the actual consequence of the precautionary principle for

the purpose it was invented is a massive spike in use of coal.

Andrew Huberman: That's taking us back over 100 years.