back to indexSean Carroll: The Nature of the Universe, Life, and Intelligence | Lex Fridman Podcast #26
link |
The following is a conversation with Sean Carroll.
link |
He's a theoretical physicist at Caltech,
link |
specializing in quantum mechanics, gravity, and cosmology.
link |
He's the author of several popular books,
link |
one on the Arrow of Time called From Eternity to Hear,
link |
one on the Higgs Boson called Particle
link |
at the End of the Universe,
link |
and one on Science and Philosophy called The Big Picture,
link |
on the origins of life, meaning, and the universe itself.
link |
He has an upcoming book on quantum mechanics
link |
that you can preorder now called Something Deeply Hidden.
link |
He writes one of my favorite blogs on his website,
link |
preposterousuniverse.com.
link |
I recommend clicking on the greatest hits link
link |
that lists accessible, interesting posts on the Arrow of Time,
link |
dark matter, dark energy, the Big Bang, general relativity,
link |
string theory, quantum mechanics,
link |
and the big meta questions about the philosophy of science,
link |
God, ethics, politics, academia, and much, much more.
link |
Finally, and perhaps most famously,
link |
he's the host of a podcast called Mindscape
link |
that you should subscribe to and support on Patreon.
link |
Along with the Joe Rogan experience,
link |
Sam Harris is Making Sense,
link |
and Dan Carlin's Hardcore History.
link |
Sean's Mindscape podcast is one of my favorite ways
link |
to learn new ideas or explore different perspectives
link |
and ideas that I thought I understood.
link |
It was truly an honor to meet and spend a couple hours
link |
It's a bit heartbreaking to say that for the first time ever,
link |
the audio recorder for this podcast died
link |
in the middle of our conversation.
link |
There are technical reasons for this,
link |
having to do with phantom power
link |
that I now understand and will avoid.
link |
It took me one hour to notice and fix the problem.
link |
So, much like the universe's 68% dark energy,
link |
roughly the same amount from this conversation was lost,
link |
except in the memories of the two people involved
link |
I'm sure we'll talk again and continue this conversation
link |
on this podcast or on Sean's.
link |
And of course, I look forward to it.
link |
This is the Artificial Intelligence podcast.
link |
If you enjoy it, subscribe on YouTube,
link |
iTunes, support on Patreon,
link |
or simply connect with me on Twitter at Lex Freedman.
link |
And now, here's my conversation with Sean Carroll.
link |
What do you think is more interesting and impactful?
link |
Understanding how the universe works
link |
at a fundamental level
link |
or understanding how the human mind works?
link |
You know, of course this is a crazy meaningless
link |
unanswerable question in some sense,
link |
because they're both very interesting
link |
and there's no absolute scale of interestingness
link |
that we can rate them on.
link |
There's a glib answer that says the human brain
link |
is part of the universe, right?
link |
And therefore, understanding the universe
link |
is more fundamental than understanding the human brain.
link |
But do you really believe that once we understand
link |
the fundamental way the universe works
link |
at the particle level,
link |
the forces we would be able to understand how the mind works?
link |
No, certainly not.
link |
We cannot understand how ice cream works
link |
just from understanding how particles work, right?
link |
So I'm a big believer in emergence.
link |
I'm a big believer that there are different ways
link |
of talking about the world
link |
beyond just the most fundamental microscopic one.
link |
You know, when we talk about tables and chairs
link |
and planets and people,
link |
we're not talking the language
link |
of particle physics and cosmology.
link |
So, but understanding the universe,
link |
you didn't say just at the most fundamental level, right?
link |
So understanding the universe at all levels is part of that.
link |
I do think, you know, to be a little bit more fair
link |
to the question, there probably are general principles
link |
of complexity, biology, information processing,
link |
memory, knowledge, creativity
link |
that go beyond just the human brain, right?
link |
And maybe one could count understanding those
link |
as part of understanding the universe.
link |
The human brain, as far as we know, is the most complex thing
link |
So there's, it's certainly absurd to think
link |
that by understanding the fundamental laws
link |
of particle physics,
link |
you get any direct insight on how the brain works.
link |
But then there's this step
link |
from the fundamentals of particle physics
link |
to information processing,
link |
which a lot of physicists and philosophers
link |
may be a little bit carelessly take
link |
when they talk about artificial intelligence.
link |
Do you think of the universe
link |
as a kind of a computational device?
link |
To be like the honest answer there is no.
link |
There's a sense in which the universe processes information
link |
There's a sense in which the universe is like a computer,
link |
But in some sense, I think,
link |
I tried to say this once on my blog
link |
and no one agreed with me,
link |
but the universe is more like a computation
link |
than a computer because the universe happens once.
link |
A computer is a general purpose machine, right?
link |
You can ask it different questions,
link |
even a pocket calculator, right?
link |
And it's set up to answer certain kinds of questions.
link |
The universe isn't that.
link |
So information processing happens in the universe,
link |
but it's not what the universe is.
link |
And I know your MIT colleague, Seth Lloyd,
link |
feels very differently about this, right?
link |
Well, you're thinking of the universe as a closed system.
link |
So what makes a computer more like a PC,
link |
like a computing machine,
link |
is that there's a human that comes up to it
link |
and moves the mouse around,
link |
so input gives it input.
link |
And that's why you're saying it's just a computation,
link |
a deterministic thing that's just unrolling.
link |
But the immense complexity of it
link |
is nevertheless like processing.
link |
There's a state and it changes with rules.
link |
And there's a sense for a lot of people
link |
that if the brain operates, the human brain operates
link |
within that world,
link |
then it's simply just a small subset of that.
link |
And so there's no reason we can't build
link |
arbitrarily great intelligences.
link |
Do you think of intelligence in this way?
link |
Intelligence is tricky.
link |
I don't have a definition of it offhand.
link |
So I remember this panel discussion
link |
that I saw on YouTube, I wasn't there,
link |
but Seth Lloyd was on the panel.
link |
And so was Martin Rees, the famous astrophysicist.
link |
And Seth gave his shtick for why the universe is a computer
link |
and explained this.
link |
And Martin Rees said, so what is not a computer?
link |
And Seth is like, oh, that's a good question.
link |
Because if you have a sufficiently broad definition
link |
of what a computer is, then everything is, right?
link |
And similarly, or the analogy gains force
link |
when it excludes some things.
link |
Is the moon going around the earth performing a computation?
link |
I can come up with definitions in which the answer is yes,
link |
but it's not a very useful computation.
link |
I think that it's absolutely helpful
link |
to think about the universe in certain situations,
link |
certain contexts, as an information processing device.
link |
I'm even guilty of writing a paper
link |
called Quantum Circuit Cosmology, where
link |
we modeled the whole universe as a quantum circuit.
link |
As a circuit, yeah.
link |
With qubits kind of thing.
link |
With qubits, basically, right.
link |
So in qubits, becoming more and more entangled.
link |
So do we want to digress a little bit?
link |
Because this is kind of fun.
link |
So here's a mystery about the universe
link |
that is so deep and profound that nobody talks about it.
link |
Space expands, right?
link |
And we talk about, in a certain region of space,
link |
a certain number of degrees of freedom,
link |
a certain number of ways that the quantum fields
link |
and the particles in that region can arrange themselves.
link |
That number of degrees of freedom in a region of space
link |
is arguably finite.
link |
We actually don't know how many there are,
link |
but there's a very good argument that says it's a finite number.
link |
So as the universe expands and space gets bigger,
link |
are there more degrees of freedom?
link |
If it's an infinite number, it doesn't really matter.
link |
Infinity times 2 is still infinity.
link |
But if it's a finite number, then there's more space,
link |
so there's more degrees of freedom.
link |
So where did they come from?
link |
That would mean the universe is not a closed system.
link |
There's more degrees of freedom popping into existence.
link |
So what we suggested was that there are more degrees of freedom.
link |
And it's not that they're not there to start,
link |
but they're not entangled to start.
link |
So the universe that you and I know of,
link |
the three dimensions around us that we see,
link |
we said those are the entangled degrees of freedom
link |
making up space time.
link |
As the universe expands, there are a whole bunch of qubits
link |
in their zero state that become entangled
link |
with the rest of space time through the action
link |
of these quantum circuits.
link |
So what does it mean that there's now more degrees of freedom
link |
as they become more entangled?
link |
As the universe expands.
link |
So there's more and more degrees of freedom
link |
that are entangled, that are playing the role of part
link |
of the entangled space time structure.
link |
So the underlying philosophy is that space time itself
link |
arises from the entanglement of some fundamental quantum
link |
degrees of freedom.
link |
At which point is most of the entanglement happening?
link |
Are we talking about close to the Big Bang?
link |
Are we talking about throughout the time of the life of the
link |
So the idea is that at the Big Bang,
link |
almost all the degrees of freedom that the universe could
link |
have were there, but they were unentangled with anything else.
link |
And that's a reflection of the fact
link |
that the Big Bang had a low entropy.
link |
It was a very simple, very small place.
link |
And as space expands, more and more degrees of freedom
link |
become entangled with the rest of the world.
link |
Well, I have to ask John Carroll,
link |
what do you think of the thought experiment from Nick
link |
Bostrom that we're living in a simulation?
link |
So I think let me contextualize that a little bit more.
link |
I think people don't actually take this thought experiment.
link |
I think it's quite interesting.
link |
It's not very useful, but it's quite interesting.
link |
From the perspective of AI, a lot of the learning
link |
that can be done usually happens in simulation,
link |
artificial examples.
link |
And so it's a constructive question
link |
to ask how difficult is our real world to simulate,
link |
which is kind of a dual part of, if we're
link |
living in a simulation and somebody built that simulation,
link |
if you were to try to do it yourself, how hard would it be?
link |
So obviously, we could be living in a simulation.
link |
If you just want the physical possibility,
link |
then I completely agree that it's physically possible.
link |
I don't think that we actually are.
link |
So take this one piece of data into consideration.
link |
We live in a big universe.
link |
There's two trillion galaxies in our observable universe
link |
with 200 billion stars in each galaxy, et cetera.
link |
It would seem to be a waste of resources
link |
to have a universe that big going on just to do a simulation.
link |
So in other words, I want to be a good Bayesian.
link |
I want to ask, under this hypothesis, what do I expect to see?
link |
So the first thing I would say is I
link |
wouldn't expect to see a universe that was that big.
link |
The second thing is I wouldn't expect the resolution
link |
of the universe to be as good as it is.
link |
So it's always possible that if our superhuman simulators only
link |
have finite resources that they don't render
link |
the entire universe, that the part that is out there,
link |
the two trillion galaxies, isn't actually
link |
being simulated fully.
link |
But then the obvious extrapolation of that
link |
is that only I am being simulated fully.
link |
The rest of you are just nonplayer characters.
link |
I'm the only thing that is real.
link |
The rest of you are just chatbots.
link |
Beyond this wall, I see the wall,
link |
but there is literally nothing on the other side of the wall.
link |
That is sort of the Bayesian prediction.
link |
That's what it would be like to do
link |
an efficient simulation of me.
link |
So none of that seems quite realistic.
link |
I don't see, I hear the argument that it's just possible
link |
and easy to simulate lots of things.
link |
I don't see any evidence from what we know about our universe
link |
that we look like a simulated universe.
link |
Now, maybe you can say, well, we don't know what it would
link |
look like, but that's just abandoning
link |
your Bayesian responsibilities.
link |
Like your job is to say, under this theory,
link |
here's what you would expect to see.
link |
Yeah, so certainly if you think about a simulation
link |
as a thing that's like a video game where only a small subset
link |
is being readied, but say all the laws of physics,
link |
the entire closed system of the quote unquote universe,
link |
Yeah, it's always possible.
link |
So that's not useful to think about
link |
when you're thinking about physics.
link |
The way Nick Bostrom phrases it, if it's possible
link |
to simulate a universe, eventually we'll do it.
link |
You can use that, by the way, for a lot of things.
link |
But I guess the question is, how hard is it
link |
to create a universe?
link |
I wrote a little blog post about this,
link |
and maybe I'm missing something.
link |
But there's an argument that says not only
link |
that it might be possible to simulate a universe,
link |
but probably, if you imagine that you actually
link |
attribute consciousness and agency
link |
to the little things that we're simulating,
link |
to our little artificial beings, there's probably
link |
a lot more of them than there are ordinary organic beings
link |
in the universe, or there will be in the future.
link |
So there's an argument that not only is being a simulation
link |
possible, it's probable, because in the space
link |
of all living consciousnesses, most of them
link |
are being simulated.
link |
Most of them are not at the top level.
link |
I think that argument must be wrong,
link |
because it follows from that argument that if we're simulated,
link |
but we can also simulate other things.
link |
Well, but if we can simulate other things,
link |
they can simulate other things.
link |
If we give them enough power and resolution,
link |
and ultimately, we'll reach a bottom,
link |
because the laws of physics in our universe
link |
have a bottom, we're made of atoms and so forth.
link |
So there will be the cheapest possible simulations.
link |
And if you believe the original argument,
link |
you should conclude that we should be in the cheapest
link |
possible simulation, because that's where most people are.
link |
But we don't look like that.
link |
It doesn't look at all like we're at the edge of resolution,
link |
that we're 16 bit things.
link |
It seems much easier to make much lower level things
link |
So, and also, I question the whole approach
link |
to the anthropic principle that says
link |
we are typical observers in the universe.
link |
I think that that's not actually,
link |
I think that there's a lot of selection that we can do
link |
that were typical within things we already know,
link |
but not typical within all the universe.
link |
So do you think there is intelligent life,
link |
however you would like to define intelligent life
link |
out there in the universe?
link |
My guess is that there is not intelligent life
link |
in the observable universe other than us.
link |
Simply on the basis of the fact that the likely number
link |
of other intelligent species in the observable universe,
link |
there's two likely numbers, zero or billions.
link |
And if there had been billions,
link |
you would have noticed already.
link |
For there to be literally like a small number,
link |
like Star Trek, there's a dozen intelligent civilizations
link |
in our galaxy, but not a billion.
link |
That's weird, that's sort of bizarre to me.
link |
It's easy for me to imagine that there are zero others
link |
because there's just a big bottleneck
link |
to making multicellular life
link |
or technological life or whatever.
link |
It's very hard for me to imagine
link |
that there's a whole bunch out there
link |
that have somehow remained hidden from us.
link |
The question I'd like to ask is,
link |
what would intelligent life look like?
link |
What I mean by that question and where it's going is,
link |
what if intelligent life is just fundamentally,
link |
in some very big ways, different than the one
link |
that has on Earth.
link |
That there's all kinds of intelligent life
link |
that operates at different scales of both size and temporal.
link |
That's a great possibility
link |
because I think we should be humble
link |
about what intelligence is, what life is.
link |
We don't even agree on what life is,
link |
much less what intelligent life is, right?
link |
So that's an argument for humility,
link |
saying there could be intelligent life
link |
of a very different character, right?
link |
You could imagine that dolphins are intelligent
link |
but never invent space travel
link |
because they live in the ocean
link |
and they don't have thumbs, right?
link |
So they never invent technology, they never invent smelting.
link |
Maybe the universe is full of intelligent species
link |
that just don't make technology, right?
link |
That's compatible with the data, I think.
link |
And I think maybe what you're pointing at
link |
is even more out there versions of intelligence,
link |
you know, intelligence in intermolecular clouds
link |
or on the surface of a neutron star
link |
or in between the galaxies in giant things
link |
where the equivalent of a heartbeat is 100 million years.
link |
On the one hand, yes,
link |
we should be very open minded about those things.
link |
On the other hand, we all of us share the same laws of physics.
link |
There might be something about the laws of physics
link |
even though we don't currently know exactly
link |
what that thing would be that makes meters
link |
and years the right length and time scales
link |
for intelligent life, maybe not.
link |
But we're made of atoms, atoms have a certain size,
link |
we orbit stars, our stars have a certain lifetime.
link |
It's not impossible to me that there's a sweet spot
link |
for intelligent life that we find ourselves in.
link |
So I'm open minded either way, I'm open minded either being humble
link |
and there's all sorts of different kinds of life
link |
or no, there's a reason we just don't know it yet
link |
why life like ours is the kind of life that's out there.
link |
Yeah, I'm of two minds too, but I often wonder
link |
if our brains is just designed to, quite obviously,
link |
to operate and see the world on these time scales.
link |
And we're almost blind and the tools we've created
link |
for detecting things are blind to the kind of observation
link |
needed to see intelligent life at other scales.
link |
Well, I'm totally open to that,
link |
but so here's another argument I would make.
link |
We have looked for intelligent life,
link |
but we've looked at for it in the dumbest way we can
link |
by turning radio telescopes to the sky.
link |
And why in the world would a super advanced civilization
link |
randomly beam out radio signals wastefully
link |
in all directions into the universe?
link |
It just doesn't make any sense, especially because
link |
in order to think that you would actually contact
link |
another civilization, you would have to do it forever.
link |
You have to keep doing it for millions of years.
link |
That sounds like a waste of resources.
link |
If you thought that there were other solar systems
link |
with planets around them where maybe intelligent life
link |
didn't yet exist, but might someday,
link |
you wouldn't try to talk to it with radio waves.
link |
You would send a spacecraft out there
link |
and you would park it around there.
link |
And it would be like, from our point of view,
link |
it would be like 2001 where there was a monolith.
link |
There could be an artifact.
link |
In fact, the other way works also, right?
link |
There could be artifacts in our solar system
link |
that have been put there by other technologically advanced
link |
civilizations, and that's how we will eventually contact them.
link |
We just haven't explored the solar system well enough yet
link |
The reason why we don't think about that is because
link |
we're young and impatient, right?
link |
It's like it would take more than my lifetime
link |
to actually send something to another star system
link |
and wait for it and then come back.
link |
But if we start thinking on hundreds of thousands of years
link |
or a million year time scales,
link |
that's clearly the right thing to do.
link |
Are you excited by the thing that Elon Musk is doing with SpaceX
link |
in general, but the idea of space exploration,
link |
even though you're species is young and impatient?
link |
No, I do think that space travel is crucially important, long term.
link |
Even to other star systems.
link |
And I think that many people overestimate the difficulty
link |
because they say, look, if you travel 1% the speed of light
link |
to another star system, we'll be dead before we get there, right?
link |
And I think that it's much easier.
link |
And therefore, when they write their science fiction stories,
link |
they imagine we'd go faster than the speed of light
link |
because otherwise they're too impatient, right?
link |
We're not going to go faster than the speed of light,
link |
but we could easily imagine that the human lifespan
link |
gets extended to thousands of years.
link |
And once you do that, then the stars are much closer.
link |
Effectively, right?
link |
What's 100 year trip, right?
link |
So I think that that's going to be the future, the far future,
link |
not my lifetime once again, but baby steps.
link |
Unless your lifetime gets extended.
link |
Well, it's in a race against time, right?
link |
A friend of mine who actually thinks about these things said,
link |
you know, you and I are going to die,
link |
but I don't know about our grandchildren.
link |
I don't know, predicting the future is hard,
link |
but that's the least plausible scenario.
link |
And so, yeah, no, I think that as we discussed earlier,
link |
there are threats to the earth, known and unknown, right?
link |
Having spread humanity and biology elsewhere
link |
is a really important longterm goal.
link |
What kind of questions can science not currently answer,
link |
When you think about the problems and the mysteries before us,
link |
that may be within reach of science.
link |
I think an obvious one is the origin of life.
link |
We don't know how that happened.
link |
There's a difficulty in knowing how it happened historically,
link |
actually, you know, literally on earth,
link |
but starting life from nonlife
link |
is something I kind of think we're close to, right?
link |
You really think so?
link |
Like, how difficult is it to start life?
link |
Well, I've talked to people, including on the podcast, about this.
link |
You know, life requires three things.
link |
Life as we know it.
link |
There's a difference between life, who knows what it is,
link |
and life as we know it,
link |
which we can talk about with some intelligence.
link |
Life as we know it requires compartmentalization.
link |
You need a little membrane around your cell.
link |
Metabolism, you need to take in food and eat it
link |
and let that make you do things.
link |
And then replication.
link |
You need to have some information about who you are,
link |
that you pass down to future generations.
link |
In the lab, compartmentalization seems pretty easy,
link |
not hard to make lipid bilayers
link |
that come into little cellular walls pretty easily.
link |
Metabolism and replication are hard,
link |
but replication we're close to.
link |
People have made RNA like molecules in the lab that...
link |
I think the state of the art is
link |
they're not able to make one molecule that reproduces itself,
link |
but they're able to make two molecules that reproduce each other.
link |
So that's okay. That's pretty close.
link |
Metabolism is harder, believe it or not,
link |
even though it's sort of the most obvious thing,
link |
but you want some sort of controlled metabolism
link |
and the actual cellular machinery in our bodies is quite complicated.
link |
It's hard to see it just popping into existence all by itself.
link |
It probably took a while.
link |
But we're making progress.
link |
In fact, I don't think we're spending nearly enough money on it.
link |
If I were the NSF, I would flood this area with money
link |
because it would change our view of the world
link |
if we could actually make life in the lab
link |
and understand how it was made originally here on Earth.
link |
I'm sure it would have some ripple effects
link |
that help cure diseases and so on.
link |
Synthetic biology is a wonderful big frontier where we're making cells.
link |
Right now, the best way to do that
link |
is to borrow heavily from existing biology.
link |
Craig Ventner several years ago created an artificial cell,
link |
but all he did was...
link |
not all he did, it was a tremendous accomplishment,
link |
but all he did was take out the DNA from a cell
link |
and put in entirely new DNA and let it boot up and go.
link |
What about the leap to creating intelligent life on Earth?
link |
However, again, we define intelligence, of course,
link |
but let's just even say homo sapiens,
link |
the modern intelligence in our human brain.
link |
Do you have a sense of what's involved in that leap
link |
and how big of a leap that is?
link |
So AI would count in this?
link |
Or do you really want life?
link |
AI would count in some sense.
link |
AI would count, I think.
link |
Of course, AI would count.
link |
Well, let's say artificial consciousness.
link |
I do not think we are on the threshold
link |
of creating artificial consciousness.
link |
I think it's possible.
link |
I'm not, again, very educated about how close we are,
link |
but my impression is not that we're really close
link |
because we understand how little we understand
link |
of consciousness and what it is.
link |
So if we don't have any idea what it is,
link |
it's hard to imagine we're on the threshold
link |
of making it ourselves.
link |
But it's doable, it's possible.
link |
I don't see any obstacles in principle,
link |
so yeah, I would hold out some interest
link |
in that happening eventually.
link |
I think in general, consciousness,
link |
I think it would be just surprised
link |
how easy consciousness is
link |
once we create intelligence.
link |
I think consciousness is a thing
link |
that's just something we all fake.
link |
No, actually, I like this idea that, in fact,
link |
consciousness is way less mysterious than we think
link |
because we're all at every time,
link |
at every moment, less conscious than we think we are.
link |
We can fool things.
link |
And I think that plus the idea that you
link |
not only have artificial intelligence systems,
link |
but you put them in a body,
link |
give them a robot body,
link |
that will help the faking a lot.
link |
Yeah, I think creating consciousness
link |
in artificial consciousness
link |
as asking a Roomba
link |
to say, I'm conscious
link |
and refusing to be talked out of it.
link |
I mean, I'm almost being silly,
link |
but that's what we do.
link |
That's what we do with each other.
link |
The consciousness is also a social construct,
link |
and a lot of our ideas of intelligence
link |
is a social construct,
link |
and so reaching that bar involves
link |
something that's beyond,
link |
that doesn't necessarily involve
link |
the fundamental understanding
link |
of how you go from
link |
electrons to neurons
link |
No, actually, I think that is an extremely good point,
link |
what it suggests is,
link |
so yeah, you referred to Kate Darling,
link |
who I had on the podcast,
link |
and who does these experiments with
link |
very simple robots,
link |
but they look like animals,
link |
and they can look like they're experiencing pain,
link |
and we human beings react
link |
very negatively to these little robots
link |
looking like they're experiencing pain,
link |
and what you want to say is,
link |
yeah, but they're just robots.
link |
It's not really pain.
link |
It's just some electrons going around,
link |
but then you realize you and I
link |
are just electrons going around,
link |
and that's what pain is also.
link |
What I would have an easy time imagining
link |
is that there is a spectrum
link |
between these simple little robots
link |
that Kate works with
link |
and a human being,
link |
where there are things that,
link |
like a human touring test level thing
link |
are not conscious,
link |
but nevertheless walk and talk
link |
like they're conscious,
link |
and it could be that the future is,
link |
I mean, Siri is close, right?
link |
And so it might be the future
link |
has a lot more agents like that,
link |
and in fact, rather than someday going,
link |
aha, we have consciousness,
link |
we'll just creep up on it
link |
with more and more accurate reflections
link |
of what we expect.
link |
And in the future, maybe the present,
link |
and you're basically assuming
link |
I get a high probability.
link |
At this time, because the,
link |
but in the future,
link |
there might be question marks around that, right?
link |
Yeah, no, absolutely.
link |
Certainly videos are almost to the point
link |
where you shouldn't trust them already.
link |
Photos you can't trust, right?
link |
Videos is easier to trust,
link |
but we're getting worse.
link |
We're getting better at faking them, right?
link |
Yeah, so physical, embodied people,
link |
what's so hard about faking that?
link |
This is very depressing,
link |
this conversation we're having right now.
link |
To me, it's exciting.
link |
You're doing it, so it's exciting to you,
link |
but it's a sobering thought.
link |
We're very bad at imagining
link |
what the next 50 years are going to be like
link |
when we're in the middle of a phase transition
link |
as we are right now.
link |
Yeah, and in general,
link |
I'm not blind to all the threats.
link |
I am excited by the power of technology
link |
I'm not as much as Steven Pinker
link |
optimistic about the world,
link |
but in everything I've seen,
link |
all the brilliant people in the world
link |
that I've met are good people.
link |
So the army of the good
link |
in terms of the development of technology is large.
link |
Okay, you're way more
link |
optimistic than I am.
link |
I think that goodness and badness
link |
are equally distributed among intelligent
link |
and unintelligent people.
link |
I don't see much of a correlation there.
link |
Neither of us have proof.
link |
Yeah, exactly. Again, opinions are free, right?
link |
Nor definitions of good and evil.
link |
Without definitions
link |
So what kind of questions can science not
link |
and may never be able to answer in your view?
link |
Well, the obvious one is what is good and bad.
link |
What is right and wrong?
link |
I think that there are questions that science tells us
link |
what happens, what the world is,
link |
doesn't say what the world should do
link |
or what we should do because we're part of the world.
link |
But we are part of the world
link |
and we have the ability to feel like
link |
something's right, something's wrong.
link |
And to make a very long story
link |
very short, I think that the idea
link |
of moral philosophy is
link |
systematizing our intuitions of what is right
link |
and what is wrong.
link |
And science might be able to predict ahead of time
link |
but it won't ever be able to judge
link |
whether we should have done it or not.
link |
You know, you're kind of unique in terms of scientists.
link |
have to do with podcasts, but
link |
even just reaching out, I think you refer to
link |
as sort of doing interdisciplinary science.
link |
and talk to people
link |
that are outside of your discipline,
link |
hope that's what science was for.
link |
In fact, I was a little disillusioned
link |
when I realized that academia
link |
at your own level, how do you prepare for these conversations?
link |
How do you think about these conversations?
link |
How do you open your mind enough
link |
to have these conversations?
link |
And it may be a little bit broader.
link |
How can you advise other scientists
link |
to have these kinds of conversations?
link |
Not at the podcast.
link |
The fact that you're doing a podcast is awesome.
link |
Other people get to hear them.
link |
But it's also good to have it without mics in general.
link |
It's a good question, but a tough one
link |
to answer. I think about
link |
a guy I know is a personal trainer
link |
and he was asked on a podcast
link |
how do we psych ourselves up
link |
to do a workout? How do we make
link |
that discipline to go and work out?
link |
And he's like, why are you asking me?
link |
I can't stop working out.
link |
I don't need to psych myself up.
link |
Likewise, you asked me
link |
how do you get to have
link |
interdisciplinary conversations and all sorts of different things
link |
with all sorts of different people?
link |
That's what makes me go.
link |
doing that. I did that long before
link |
any of them were recorded. In fact,
link |
a lot of the motivation for starting recording it
link |
was making sure I would read all these books
link |
that I had purchased. All these books
link |
I wanted to read. Not enough time to read them.
link |
And now, if I have the motivation
link |
because I'm going to interview Pat
link |
Churchland, I'm going to finally read her
link |
it's absolutely true that academia is
link |
extraordinarily siloed. We don't talk to people.
link |
And in fact, when we do, it's punished.
link |
The people who do it successfully
link |
generally first became
link |
very successful within their little siloed discipline.
link |
did they start expanding out.
link |
If you're a young person, I have graduate students
link |
and I try to be very, very
link |
candid with them about this.
link |
most graduate students do not become faculty members.
link |
It's a tough road.
link |
you live the life you want to live
link |
but do it with your eyes open
link |
about what it does to your job chances.
link |
broad you are and the less
link |
time you spend hyper
link |
specializing in your field, the lower
link |
your job chances are. That's just an academic
link |
reality. It's terrible. I don't like it.
link |
But it's a reality.
link |
And for some people
link |
that's fine. Like there's plenty of people
link |
who are wonderful scientists who have zero
link |
interest in branching out and talking to
link |
things to anyone outside their field.
link |
it is disillusioning to me
link |
some of the romantic notion
link |
I had of the intellectual academic life
link |
is belied by the reality
link |
of it. The idea that we should
link |
reach out beyond our discipline
link |
and that is a positive good
link |
universities that it may as well
link |
not exist at all. But
link |
that said, even though you're saying
link |
you're doing it like the personal trainer
link |
because you just can't help it, you're also
link |
an inspiration to others.
link |
Like I could speak for myself.
link |
I also have a career I'm thinking about
link |
right. And without
link |
your podcast, I may have
link |
not have been doing this at all.
link |
makes me realize that these kinds
link |
of conversations is kind of what science is about.
link |
ways. The reason we write papers
link |
this exchange of ideas
link |
is much harder to do
link |
into the disciplinary papers, I would say.
link |
And conversations are easier.
link |
So conversations is the beginning
link |
and in the field of AI
link |
obvious that we should think outside
link |
computer vision competitions and in particular
link |
data sets. We should think about the broader
link |
impact of how this can be
link |
you know, reaching
link |
out to physics, to psychology
link |
and having these conversations.
link |
So you're an inspiration
link |
and so. Well, thank you very much.
link |
Never know how the world
link |
the fact that this stuff is out there
link |
a huge number of people come up to me
link |
grad students really loving the
link |
podcast inspired by it and
link |
they will probably have that
link |
there'll be ripple effects when they become faculty
link |
and so on. So we can end
link |
on a balance between pessimism
link |
and optimism and Sean, thank you so much
link |
for talking. It was awesome. No, Lex, thank you very
link |
much for this conversation. It was great.