Transcript
00:00:00 The following is a conversation with Sean Carroll.
00:00:02 He’s a theoretical physicist at Caltech
00:00:04 specializing in quantum mechanics, gravity, and cosmology.
00:00:08 He’s the author of several popular books,
00:00:11 one on the arrow of time called From Eternity to Here,
00:00:15 one on the Higgs boson called Particle
00:00:17 at the End of the Universe,
00:00:19 and one on science and philosophy called The Big Picture
00:00:22 on the Origins of Life, Meaning, and the Universe Itself.
00:00:26 He has an upcoming book on quantum mechanics
00:00:28 that you can preorder now called Something Deeply Hidden.
00:00:32 He writes one of my favorite blogs on his website,
00:00:36 preposterousuniverse.com.
00:00:37 I recommend clicking on the Greatest Hits link
00:00:40 that lists accessible, interesting posts
00:00:43 on the arrow of time, dark matter, dark energy,
00:00:45 the Big Bang, general relativity,
00:00:47 string theory, quantum mechanics,
00:00:49 and the big meta questions about the philosophy of science,
00:00:53 God, ethics, politics, academia, and much, much more.
00:00:57 Finally, and perhaps most famously,
00:01:00 he’s the host of a podcast called Mindscape
00:01:03 that you should subscribe to and support on Patreon.
00:01:06 Along with the Joe Rogan experience,
00:01:08 Sam Harris’s Making Sense,
00:01:10 and Dan Carlin’s Hardcore History,
00:01:13 Sean’s Mindscape podcast is one of my favorite ways
00:01:15 to learn new ideas or explore different perspectives
00:01:18 and ideas that I thought I understood.
00:01:22 It was truly an honor to meet
00:01:24 and spend a couple hours with Sean.
00:01:27 It’s a bit heartbreaking to say
00:01:28 that for the first time ever,
00:01:30 the audio recorder for this podcast
00:01:32 died in the middle of our conversation.
00:01:34 There’s technical reasons for this,
00:01:36 having to do with phantom power
00:01:38 that I now understand and will avoid.
00:01:41 It took me one hour to notice and fix the problem.
00:01:44 So, much like the universe is 68% dark energy,
00:01:48 roughly the same amount from this conversation was lost,
00:01:51 except in the memories of the two people involved
00:01:54 and in my notes.
00:01:56 I’m sure we’ll talk again and continue this conversation
00:01:59 on this podcast or on Sean’s.
00:02:02 And of course, I look forward to it.
00:02:05 This is the Artificial Intelligence podcast.
00:02:07 If you enjoy it, subscribe on YouTube, iTunes,
00:02:11 support it on Patreon,
00:02:12 or simply connect with me on Twitter at Lex Friedman.
00:02:16 And now, here’s my conversation with Sean Carroll.
00:02:21 What do you think is more interesting and impactful,
00:02:23 understanding how the universe works at a fundamental level
00:02:26 or understanding how the human mind works?
00:02:29 You know, of course this is a crazy,
00:02:31 meaningless, unanswerable question in some sense,
00:02:33 because they’re both very interesting
00:02:35 and there’s no absolute scale of interestingness
00:02:37 that we can rate them on.
00:02:39 There’s a glib answer that says the human brain
00:02:41 is part of the universe, right?
00:02:43 And therefore, understanding the universe
00:02:44 is more fundamental than understanding the human brain.
00:02:47 But do you really believe that once we understand
00:02:49 the fundamental way the universe works
00:02:51 at the particle level, the forces,
00:02:53 we would be able to understand how the mind works?
00:02:55 No, certainly not.
00:02:56 We cannot understand how ice cream works
00:02:58 just from understanding how particles work, right?
00:03:01 So I’m a big believer in emergence.
00:03:02 I’m a big believer that there are different ways
00:03:05 of talking about the world
00:03:07 beyond just the most fundamental microscopic one.
00:03:11 You know, when we talk about tables and chairs
00:03:13 and planets and people,
00:03:15 we’re not talking the language of particle physics
00:03:17 and cosmology.
00:03:18 So, but understanding the universe,
00:03:20 you didn’t say just at the most fundamental level, right?
00:03:24 So understanding the universe at all levels
00:03:26 is part of that.
00:03:28 I do think, you know, to be a little bit more fair
00:03:29 to the question, there probably are general principles
00:03:33 of complexity, biology, information processing,
00:03:38 memory, knowledge, creativity
00:03:41 that go beyond just the human brain, right?
00:03:45 And maybe one could count understanding those
00:03:47 as part of understanding the universe.
00:03:49 The human brain, as far as we know,
00:03:50 is the most complex thing in the universe.
00:03:54 So there’s, it’s certainly absurd to think
00:03:57 that by understanding the fundamental laws
00:03:58 of particle physics,
00:04:00 you get any direct insight on how the brain works.
00:04:02 But then there’s this step from the fundamentals
00:04:05 of particle physics to information processing,
00:04:08 which a lot of physicists and philosophers
00:04:10 may be a little bit carelessly take
00:04:12 when they talk about artificial intelligence.
00:04:14 Do you think of the universe
00:04:18 as a kind of a computational device?
00:04:21 No, to be like, the honest answer there is no.
00:04:24 There’s a sense in which the universe
00:04:26 processes information, clearly.
00:04:29 There’s a sense in which the universe
00:04:30 is like a computer, clearly.
00:04:33 But in some sense, I think,
00:04:36 I tried to say this once on my blog
00:04:38 and no one agreed with me,
00:04:39 but the universe is more like a computation
00:04:42 than a computer because the universe happens once.
00:04:45 A computer is a general purpose machine, right?
00:04:46 That you can ask it different questions,
00:04:48 even a pocket calculator, right?
00:04:50 And it’s set up to answer certain kinds of questions.
00:04:52 The universe isn’t that.
00:04:54 So information processing happens in the universe,
00:04:57 but it’s not what the universe is.
00:04:59 And I know your MIT colleague, Seth Lloyd,
00:05:01 feels very differently about this, right?
00:05:03 Well, you’re thinking of the universe as a closed system.
00:05:07 I am.
00:05:08 So what makes a computer more like a PC,
00:05:12 like a computing machine is that there’s a human
00:05:15 that every once comes up to it and moves the mouse around.
00:05:19 So input.
00:05:19 Gives it input.
00:05:20 Gives it input.
00:05:23 And that’s why you’re saying it’s just a computation,
00:05:26 a deterministic thing that’s just unrolling.
00:05:29 But the immense complexity of it
00:05:32 is nevertheless like processing.
00:05:34 There’s a state and then it changes with good rules.
00:05:40 And there’s a sense for a lot of people
00:05:41 that if the brain operates,
00:05:44 the human brain operates within that world,
00:05:46 then it’s simply just a small subset of that.
00:05:49 And so there’s no reason we can’t build
00:05:52 arbitrarily great intelligences.
00:05:55 Yeah.
00:05:56 Do you think of intelligence in this way?
00:05:58 Intelligence is tricky.
00:05:59 I don’t have a definition of it offhand.
00:06:01 So I remember this panel discussion that I saw on YouTube.
00:06:05 I wasn’t there, but Seth Lloyd was on the panel.
00:06:07 And so was Martin Rees, the famous astrophysicist.
00:06:10 And Seth gave his shtick for why the universe is a computer
00:06:13 and explained this.
00:06:14 And Martin Rees said, so what is not a computer?
00:06:19 And Seth was like, oh, that’s a good question.
00:06:22 I’m not sure.
00:06:22 Because if you have a sufficiently broad definition
00:06:24 of what a computer is, then everything is, right?
00:06:28 And the simile or the analogy gains force
00:06:32 when it excludes some things.
00:06:34 You know, is the moon going around the earth
00:06:36 performing a computation?
00:06:38 I can come up with definitions in which the answer is yes,
00:06:41 but it’s not a very useful computation.
00:06:43 I think that it’s absolutely helpful
00:06:46 to think about the universe in certain situations,
00:06:49 certain contexts, as an information processing device.
00:06:53 I’m even guilty of writing a paper
00:06:54 called Quantum Circuit Cosmology,
00:06:56 where we modeled the whole universe as a quantum circuit.
00:06:59 As a circuit.
00:07:00 As a circuit, yeah.
00:07:01 With qubits kind of thing?
00:07:02 With qubits basically, right, yeah.
00:07:05 So, and qubits becoming more and more entangled.
00:07:07 So do we wanna digress a little bit?
00:07:09 Let’s do it.
00:07:10 It’s kind of fun.
00:07:11 So here’s a mystery about the universe
00:07:13 that is so deep and profound that nobody talks about it.
00:07:16 Space expands, right?
00:07:19 And we talk about, in a certain region of space,
00:07:21 a certain number of degrees of freedom,
00:07:23 a certain number of ways that the quantum fields
00:07:25 and the particles in that region can arrange themselves.
00:07:28 That number of degrees of freedom in a region of space
00:07:32 is arguably finite.
00:07:33 We actually don’t know how many there are,
00:07:36 but there’s a very good argument
00:07:37 that says it’s a finite number.
00:07:39 So as the universe expands and space gets bigger,
00:07:44 are there more degrees of freedom?
00:07:46 If it’s an infinite number, it doesn’t really matter.
00:07:48 Infinity times two is still infinity.
00:07:49 But if it’s a finite number, then there’s more space,
00:07:53 so there’s more degrees of freedom.
00:07:54 So where did they come from?
00:07:55 That would mean the universe is not a closed system.
00:07:58 There’s more degrees of freedom popping into existence.
00:08:01 So what we suggested was
00:08:03 that there are more degrees of freedom,
00:08:05 and it’s not that they’re not there to start,
00:08:07 but they’re not entangled to start.
00:08:10 So the universe that you and I know of,
00:08:12 the three dimensions around us that we see,
00:08:15 we said those are the entangled degrees of freedom
00:08:18 making up space time.
00:08:19 And as the universe expands,
00:08:20 there are a whole bunch of qubits in their zero state
00:08:25 that become entangled with the rest of space time
00:08:28 through the action of these quantum circuits.
00:08:31 So what does it mean that there’s now more
00:08:35 degrees of freedom as they become more entangled?
00:08:39 Yeah, so.
00:08:40 As the universe expands.
00:08:41 That’s right, so there’s more and more degrees of freedom
00:08:43 that are entangled, that are playing part,
00:08:46 playing the role of part
00:08:47 of the entangled space time structure.
00:08:49 So the basic, the underlying philosophy is
00:08:51 that space time itself arises from the entanglement
00:08:54 of some fundamental quantum degrees of freedom.
00:08:57 Wow, okay, so at which point
00:09:00 is most of the entanglement happening?
00:09:05 Are we talking about close to the Big Bang?
00:09:07 Are we talking about throughout the time of the life?
00:09:11 Throughout history, yeah.
00:09:12 So the idea is that at the Big Bang,
00:09:15 almost all the degrees of freedom
00:09:16 that the universe could have were there,
00:09:19 but they were unentangled with anything else.
00:09:22 And that’s a reflection of the fact
00:09:23 that the Big Bang had a low entropy.
00:09:25 It was a very simple, very small place.
00:09:28 And as space expands, more and more degrees of freedom
00:09:31 become entangled with the rest of the world.
00:09:34 Well, I have to ask John Carroll,
00:09:35 what do you think of the thought experiment
00:09:37 from Nick Bostrom that we’re living in a simulation?
00:09:41 So I think, let me contextualize that a little bit more.
00:09:44 I think people don’t actually take this thought experiments.
00:09:48 I think it’s quite interesting.
00:09:50 It’s not very useful, but it’s quite interesting.
00:09:52 From the perspective of AI,
00:09:54 a lot of the learning that can be done usually happens
00:09:58 in simulation from artificial examples.
00:10:00 And so it’s a constructive question to ask,
00:10:04 how difficult is our real world to simulate?
00:10:08 Right.
00:10:09 Which is kind of a dual part of,
00:10:12 if we’re living in a simulation
00:10:14 and somebody built that simulation,
00:10:16 if you were to try to do it yourself, how hard would it be?
00:10:18 So obviously we could be living in a simulation.
00:10:21 If you just want the physical possibility,
00:10:23 then I completely agree that it’s physically possible.
00:10:25 I don’t think that we actually are.
00:10:27 So take this one piece of data into consideration.
00:10:30 You know, we live in a big universe, okay?
00:10:35 There’s two trillion galaxies in our observable universe
00:10:38 with 200 billion stars in each galaxy, et cetera.
00:10:41 It would seem to be a waste of resources
00:10:44 to have a universe that big going on
00:10:46 just to do a simulation.
00:10:47 So in other words, I want to be a good Bayesian.
00:10:50 I want to ask under this hypothesis,
00:10:52 what do I expect to see?
00:10:54 So the first thing I would say is I wouldn’t expect
00:10:56 to see a universe that was that big, okay?
00:11:00 The second thing is I wouldn’t expect the resolution
00:11:02 of the universe to be as good as it is.
00:11:05 So it’s always possible that if our superhuman simulators
00:11:08 only have finite resources,
00:11:09 that they don’t render the entire universe, right?
00:11:12 That the part that is out there,
00:11:14 the two trillion galaxies,
00:11:16 isn’t actually being simulated fully, okay?
00:11:19 But then the obvious extrapolation of that
00:11:22 is that only I am being simulated fully.
00:11:25 Like the rest of you are just non player characters, right?
00:11:29 I’m the only thing that is real.
00:11:30 The rest of you are just chat bots.
00:11:32 Beyond this wall, I see the wall,
00:11:34 but there is literally nothing
00:11:36 on the other side of the wall.
00:11:37 That is sort of the Bayesian prediction.
00:11:39 That’s what it would be like
00:11:40 to do an efficient simulation of me.
00:11:42 So like none of that seems quite realistic.
00:11:45 I don’t see, I hear the argument that it’s just possible
00:11:50 and easy to simulate lots of things.
00:11:53 I don’t see any evidence from what we know
00:11:55 about our universe that we look like a simulated universe.
00:11:59 Now, maybe you can say,
00:12:00 well, we don’t know what it would look like,
00:12:01 but that’s just abandoning your Bayesian responsibilities.
00:12:04 Like your job is to say under this theory,
00:12:07 here’s what you would expect to see.
00:12:09 Yeah, so certainly if you think about simulation
00:12:11 as a thing that’s like a video game
00:12:14 where only a small subset is being rendered.
00:12:17 But say the entire, all the laws of physics,
00:12:22 the entire closed system of the quote unquote universe,
00:12:26 it had a creator.
00:12:27 Yeah, it’s always possible.
00:12:29 Right, so that’s not useful to think about
00:12:32 when you’re thinking about physics.
00:12:34 The way Nick Bostrom phrases it,
00:12:36 if it’s possible to simulate a universe,
00:12:39 eventually we’ll do it.
00:12:40 Right.
00:12:42 You can use that by the way for a lot of things.
00:12:44 Well, yeah.
00:12:45 But I guess the question is,
00:12:48 how hard is it to create a universe?
00:12:52 I wrote a little blog post about this
00:12:53 and maybe I’m missing something,
00:12:55 but there’s an argument that says not only
00:12:57 that it might be possible to simulate a universe,
00:13:00 but probably if you imagine that you actually attribute
00:13:05 consciousness and agency to the little things
00:13:08 that we’re simulating, to our little artificial beings,
00:13:12 there’s probably a lot more of them
00:13:13 than there are ordinary organic beings in the universe
00:13:15 or there will be in the future, right?
00:13:17 So there’s an argument that not only
00:13:18 is being a simulation possible,
00:13:20 it’s probable because in the space
00:13:23 of all living consciousnesses,
00:13:24 most of them are being simulated, right?
00:13:26 Most of them are not at the top level.
00:13:28 I think that argument must be wrong
00:13:30 because it follows from that argument that,
00:13:33 if we’re simulated, but we can also simulate other things,
00:13:36 well, but if we can simulate other things,
00:13:38 they can simulate other things, right?
00:13:41 If we give them enough power and resolution
00:13:44 and ultimately we’ll reach a bottom
00:13:45 because the laws of physics in our universe have a bottom,
00:13:49 we’re made of atoms and so forth,
00:13:51 so there will be the cheapest possible simulations.
00:13:55 And if you believe the original argument,
00:13:57 you should conclude that we should be
00:13:59 in the cheapest possible simulation
00:14:00 because that’s where most people are.
00:14:02 But we don’t look like that.
00:14:03 It doesn’t look at all like we’re at the edge of resolution,
00:14:06 that we’re 16 bit things.
00:14:09 It seems much easier to make much lower level things
00:14:13 than we are.
00:14:14 And also, I questioned the whole approach
00:14:18 to the anthropic principle
00:14:19 that says we are typical observers in the universe.
00:14:22 I think that that’s not actually,
00:14:23 I think that there’s a lot of selection that we can do
00:14:27 that we’re typical within things we already know,
00:14:30 but not typical within all of the universe.
00:14:32 So do you think there’s intelligent life,
00:14:35 however you would like to define intelligent life,
00:14:37 out there in the universe?
00:14:39 My guess is that there is not intelligent life
00:14:44 in the observable universe other than us, simply
00:14:48 on the basis of the fact that the likely number
00:14:52 of other intelligent species in the observable universe,
00:14:56 there’s two likely numbers, zero or billions.
00:15:01 And if there had been billions,
00:15:02 you would have noticed already.
00:15:05 For there to be literally like a small number,
00:15:07 like, you know, Star Trek,
00:15:09 there’s a dozen intelligent civilizations in our galaxy,
00:15:13 but not a billion, that’s weird.
00:15:17 That’s sort of bizarre to me.
00:15:18 It’s easy for me to imagine that there are zero others
00:15:21 because there’s just a big bottleneck
00:15:22 to making multicellular life
00:15:24 or technological life or whatever.
00:15:27 It’s very hard for me to imagine
00:15:28 that there’s a whole bunch out there
00:15:30 that have somehow remained hidden from us.
00:15:32 The question I’d like to ask
00:15:34 is what would intelligent life look like?
00:15:38 What I mean by that question and where it’s going
00:15:40 is what if intelligent life is just in some very big ways
00:15:47 different than the one that has on Earth?
00:15:51 That there’s all kinds of intelligent life
00:15:53 that operates at different scales
00:15:55 of both size and temporal.
00:15:57 Right, that’s a great possibility
00:15:59 because I think we should be humble
00:16:00 about what intelligence is, what life is.
00:16:02 We don’t even agree on what life is,
00:16:04 much less what intelligent life is, right?
00:16:07 So that’s an argument for humility,
00:16:08 saying there could be intelligent life
00:16:10 of a very different character, right?
00:16:13 Like you could imagine the dolphins are intelligent
00:16:18 but never invent space travel
00:16:20 because they live in the ocean
00:16:21 and they don’t have thumbs, right?
00:16:24 So they never invent technology, they never invent smelting.
00:16:27 Maybe the universe is full of intelligent species
00:16:32 that just don’t make technology, right?
00:16:34 That’s compatible with the data, I think.
00:16:36 And I think maybe what you’re pointing at
00:16:39 is even more out there versions of intelligence,
00:16:44 intelligence in intermolecular clouds
00:16:47 or on the surface of a neutron star
00:16:49 or in between the galaxies in giant things
00:16:51 where the equivalent of a heartbeat is 100 million years.
00:16:56 On the one hand, yes,
00:16:58 we should be very open minded about those things.
00:16:59 On the other hand, all of us share the same laws of physics.
00:17:04 There might be something about the laws of physics,
00:17:08 even though we don’t currently know
00:17:09 exactly what that thing would be,
00:17:10 that makes meters and years
00:17:16 the right length and timescales for intelligent life.
00:17:19 Maybe not, but we’re made of atoms,
00:17:22 atoms have a certain size,
00:17:23 we orbit stars or stars have a certain lifetime.
00:17:27 It’s not impossible to me that there’s a sweet spot
00:17:30 for intelligent life that we find ourselves in.
00:17:32 So I’m open minded either way,
00:17:33 I’m open minded either being humble
00:17:35 and there’s all sorts of different kinds of life
00:17:37 or no, there’s a reason we just don’t know it yet
00:17:39 why life like ours is the kind of life that’s out there.
00:17:42 Yeah, I’m of two minds too,
00:17:43 but I often wonder if our brains is just designed
00:17:47 to quite obviously to operate and see the world
00:17:52 in these timescales and we’re almost blind
00:17:56 and the tools we’ve created for detecting things are blind
00:18:01 to the kind of observation needed
00:18:02 to see intelligent life at other scales.
00:18:04 Well, I’m totally open to that,
00:18:07 but so here’s another argument I would make,
00:18:09 we have looked for intelligent life,
00:18:11 but we’ve looked at for it in the dumbest way we can,
00:18:14 by turning radio telescopes to the sky.
00:18:16 And why in the world would a super advanced civilization
00:18:21 randomly beam out radio signals wastefully
00:18:24 in all directions into the universe?
00:18:25 That just doesn’t make any sense,
00:18:27 especially because in order to think
00:18:29 that you would actually contact another civilization,
00:18:32 you would have to do it forever,
00:18:33 you have to keep doing it for millions of years,
00:18:35 that sounds like a waste of resources.
00:18:38 If you thought that there were other solar systems
00:18:43 with planets around them,
00:18:44 where maybe intelligent life didn’t yet exist,
00:18:47 but might someday,
00:18:48 you wouldn’t try to talk to it with radio waves,
00:18:51 you would send a spacecraft out there
00:18:53 and you would park it around there
00:18:55 and it would be like, from our point of view,
00:18:57 it’d be like 2001, where there was a monolith.
00:19:00 Monolith.
00:19:01 There could be an artifact,
00:19:02 in fact, the other way works also, right?
00:19:04 There could be artifacts in our solar system
00:19:08 that have been put there
00:19:10 by other technologically advanced civilizations
00:19:12 and that’s how we will eventually contact them.
00:19:14 We just haven’t explored the solar system well enough yet
00:19:16 to find them.
00:19:18 The reason why we don’t think about that
00:19:20 is because we’re young and impatient, right?
00:19:21 Like, it would take more than my lifetime
00:19:24 to actually send something to another star system
00:19:26 and wait for it and then come back.
00:19:27 So, but if we start thinking on hundreds of thousands
00:19:30 of years or million year time scales,
00:19:32 that’s clearly the right thing to do.
00:19:34 Are you excited by the thing
00:19:36 that Elon Musk is doing with SpaceX in general?
00:19:39 Space, but the idea of space exploration,
00:19:41 even though your, or your species is young and impatient?
00:19:45 Yeah.
00:19:46 No, I do think that space travel is crucially important,
00:19:49 long term.
00:19:50 Even to other star systems.
00:19:52 And I think that many people overestimate the difficulty
00:19:57 because they say, look, if you travel 1% the speed of light
00:20:00 to another star system,
00:20:02 we’ll be dead before we get there, right?
00:20:04 And I think that it’s much easier.
00:20:06 And therefore, when they write their science fiction stories,
00:20:08 they imagine we’d go faster than the speed of light
00:20:09 because otherwise they’re too impatient, right?
00:20:11 We’re not gonna go faster than the speed of light,
00:20:13 but we could easily imagine that the human lifespan
00:20:16 gets extended to thousands of years.
00:20:18 And once you do that,
00:20:19 then the stars are much closer effectively, right?
00:20:21 And then what’s a hundred year trip, right?
00:20:23 So I think that that’s gonna be the future,
00:20:25 the far future, not my lifetime once again,
00:20:28 but baby steps.
00:20:30 Unless your lifetime gets extended.
00:20:32 Well, it’s in a race against time, right?
00:20:34 A friend of mine who actually thinks about these things
00:20:37 said, you know, you and I are gonna die,
00:20:40 but I don’t know about our grandchildren.
00:20:43 That’s, I don’t know, predicting the future is hard,
00:20:45 but that’s at least a plausible scenario.
00:20:47 And so, yeah, no, I think that as we discussed earlier,
00:20:51 there are threats to the earth, known and unknown, right?
00:20:56 Having spread humanity and biology elsewhere
00:21:02 is a really important longterm goal.
00:21:04 What kind of questions can science not currently answer,
00:21:08 but might soon?
00:21:11 When you think about the problems and the mysteries
00:21:13 before us that may be within reach of science.
00:21:17 I think an obvious one is the origin of life.
00:21:20 We don’t know how that happened.
00:21:22 There’s a difficulty in knowing how it happened historically
00:21:25 actually, you know, literally on earth,
00:21:27 but starting life from non life is something
00:21:30 I kind of think we’re close to, right?
00:21:32 We’re really.
00:21:33 You really think so?
00:21:34 Like how difficult is it to start life?
00:21:36 Well, I’ve talked to people,
00:21:39 including on the podcast about this.
00:21:41 You know, life requires three things.
00:21:43 Life as we know it.
00:21:44 So there’s a difference with life,
00:21:45 which who knows what it is,
00:21:47 and life as we know it,
00:21:48 which we can talk about with some intelligence.
00:21:50 So life as we know it requires compartmentalization.
00:21:53 You need like a little membrane around your cell.
00:21:56 Metabolism, you need to take in food and eat it
00:21:58 and let that make you do things.
00:22:01 And then replication, okay?
00:22:02 So you need to have some information about who you are
00:22:04 that you pass down to future generations.
00:22:07 In the lab, compartmentalization seems pretty easy.
00:22:11 Not hard to make lipid bilayers
00:22:13 that come into little cellular walls pretty easily.
00:22:16 Metabolism and replication are hard,
00:22:20 but replication we’re close to.
00:22:21 People have made RNA like molecules in the lab
00:22:24 that I think the state of the art is,
00:22:28 they’re not able to make one molecule
00:22:30 that reproduces itself,
00:22:32 but they’re able to make two molecules
00:22:33 that reproduce each other.
00:22:35 So that’s okay.
00:22:36 That’s pretty close.
00:22:38 Metabolism is harder, believe it or not,
00:22:41 even though it’s sort of the most obvious thing,
00:22:42 but you want some sort of controlled metabolism
00:22:44 and the actual cellular machinery in our bodies
00:22:47 is quite complicated.
00:22:48 It’s hard to see it just popping into existence
00:22:50 all by itself.
00:22:51 It probably took a while,
00:22:53 but we’re making progress.
00:22:56 And in fact, I don’t think we’re spending
00:22:57 nearly enough money on it.
00:22:58 If I were the NSF, I would flood this area with money
00:23:01 because it would change our view of the world
00:23:05 if we could actually make life in the lab
00:23:06 and understand how it was made originally here on earth.
00:23:09 And I’m sure it’d have some ripple effects
00:23:11 that help cure disease and so on.
00:23:12 I mean, just that understanding.
00:23:14 So synthetic biology is a wonderful big frontier
00:23:16 where we’re making cells.
00:23:18 Right now, the best way to do that
00:23:21 is to borrow heavily from existing biology, right?
00:23:23 Well, Craig Venter several years ago
00:23:25 created an artificial cell, but all he did was,
00:23:28 not all he did, it was a tremendous accomplishment,
00:23:29 but all he did was take out the DNA from a cell
00:23:33 and put in entirely new DNA and let it boot up and go.
00:23:37 What about the leap to creating intelligent life on earth?
00:23:43 Yeah.
00:23:44 Again, we define intelligence, of course,
00:23:45 but let’s just even say Homo sapiens,
00:23:49 the modern intelligence in our human brain.
00:23:55 Do you have a sense of what’s involved in that leap
00:23:58 and how big of a leap that is?
00:24:00 So AI would count in this, or do you really want life?
00:24:03 Do you want really an organism in some sense?
00:24:06 AI would count, I think.
00:24:07 Okay.
00:24:08 Yeah, of course, of course AI would count.
00:24:11 Well, let’s say artificial consciousness, right?
00:24:13 So I do not think we are on the threshold
00:24:15 of creating artificial consciousness.
00:24:16 I think it’s possible.
00:24:18 I’m not, again, very educated about how close we are,
00:24:20 but my impression is not that we’re really close
00:24:22 because we understand how little we understand
00:24:24 of consciousness and what it is.
00:24:26 So if we don’t have any idea what it is,
00:24:28 it’s hard to imagine we’re on the threshold
00:24:29 of making it ourselves.
00:24:32 But it’s doable, it’s possible.
00:24:34 I don’t see any obstacles in principle.
00:24:35 So yeah, I would hold out some interest
00:24:38 in that happening eventually.
00:24:40 I think in general, consciousness,
00:24:42 I think we would be just surprised
00:24:44 how easy consciousness is once we create intelligence.
00:24:49 I think consciousness is a thing
00:24:50 that’s just something we all fake.
00:24:55 Well, good.
00:24:56 No, actually, I like this idea that in fact,
00:24:57 consciousness is way less mysterious than we think
00:25:00 because we’re all at every time, at every moment,
00:25:02 less conscious than we think we are, right?
00:25:04 We can fool things.
00:25:05 And I think that plus the idea
00:25:07 that you not only have artificial intelligent systems,
00:25:11 but you put them in a body, right,
00:25:12 give them a robot body,
00:25:15 that will help the faking a lot.
00:25:18 Yeah, I think creating consciousness
00:25:20 in artificial consciousness is as simple
00:25:25 as asking a Roomba to say, I’m conscious,
00:25:30 and refusing to be talked out of it.
00:25:32 Could be, it could be.
00:25:33 And I mean, I’m almost being silly,
00:25:36 but that’s what we do.
00:25:39 That’s what we do with each other.
00:25:40 This is the kind of,
00:25:42 that consciousness is also a social construct.
00:25:44 And a lot of our ideas of intelligence is a social construct.
00:25:47 And so reaching that bar involves something that’s beyond,
00:25:52 that doesn’t necessarily involve
00:25:54 the fundamental understanding of how you go
00:25:57 from electrons to neurons to cognition.
00:26:02 No, actually, I think that is an extremely good point.
00:26:05 And in fact, what it suggests is,
00:26:08 so yeah, you referred to Kate Darling,
00:26:10 who I had on the podcast,
00:26:11 and who does these experiments with very simple robots,
00:26:16 but they look like animals,
00:26:18 and they can look like they’re experiencing pain,
00:26:20 and we human beings react very negatively
00:26:23 to these little robots
00:26:24 looking like they’re experiencing pain.
00:26:26 And what you wanna say is, yeah, but they’re just robots.
00:26:29 It’s not really pain, right?
00:26:31 It’s just some electrons going around.
00:26:33 But then you realize, you and I are just electrons
00:26:36 going around, and that’s what pain is also.
00:26:38 And so what I would have an easy time imagining
00:26:43 is that there is a spectrum
00:26:44 between these simple little robots that Kate works with
00:26:47 and a human being,
00:26:49 where there are things that sort of
00:26:50 by some strict definition,
00:26:52 Turing test level thing are not conscious,
00:26:55 but nevertheless walk and talk like they’re conscious.
00:26:58 And it could be that the future is,
00:27:00 I mean, Siri is close, right?
00:27:02 And so it might be the future
00:27:04 has a lot more agents like that.
00:27:07 And in fact, rather than someday going,
00:27:08 aha, we have consciousness,
00:27:10 we’ll just creep up on it with more and more
00:27:13 accurate reflections of what we expect.
00:27:15 And in the future, maybe the present,
00:27:18 for example, we haven’t met before,
00:27:20 and you’re basically assuming that I’m human as it’s a high
00:27:25 probability at this time because the yeah,
00:27:28 but in the future,
00:27:30 there might be question marks around that, right?
00:27:32 Yeah, no, absolutely.
00:27:33 Certainly videos are almost to the point
00:27:35 where you shouldn’t trust them already.
00:27:36 Photos you can’t trust, right?
00:27:39 Videos is easier to trust,
00:27:41 but we’re getting worse that,
00:27:44 we’re getting better at faking them, right?
00:27:46 Yeah, so physical embodied people,
00:27:48 what’s so hard about faking that?
00:27:51 So this is very depressing,
00:27:51 this conversation we’re having right now.
00:27:53 So I mean,
00:27:54 To me, it’s exciting.
00:27:55 To me, you’re doing it.
00:27:56 So it’s exciting to you,
00:27:57 but it’s a sobering thought.
00:27:59 We’re very bad, right?
00:28:00 At imagining what the next 50 years are gonna be like
00:28:02 when we’re in the middle of a phase transition
00:28:04 as we are right now.
00:28:05 Yeah, and I, in general,
00:28:06 I’m not blind to all the threats.
00:28:09 I am excited by the power of technology to solve,
00:28:14 to protect us against the threats as they evolve.
00:28:18 I’m not as much as Steven Pinker optimistic about the world,
00:28:22 but in everything I’ve seen,
00:28:23 all of the brilliant people in the world that I’ve met
00:28:27 are good people.
00:28:29 So the army of the good
00:28:30 in terms of the development of technology is large.
00:28:33 Okay, you’re way more optimistic than I am.
00:28:37 I think that goodness and badness
00:28:39 are equally distributed among intelligent
00:28:40 and unintelligent people.
00:28:42 I don’t see much of a correlation there.
00:28:44 Interesting.
00:28:46 Neither of us have proof.
00:28:47 Yeah, exactly.
00:28:48 Again, opinions are free, right?
00:28:50 Nor definitions of good and evil.
00:28:52 We come without definitions or without data opinions.
00:28:57 So what kind of questions can science not currently answer
00:29:01 and may never be able to answer in your view?
00:29:04 Well, the obvious one is what is good and bad?
00:29:06 What is right and wrong?
00:29:07 I think that there are questions that,
00:29:09 science tells us what happens,
00:29:11 what the world is and what it does.
00:29:13 It doesn’t say what the world should do
00:29:14 or what we should do,
00:29:15 because we’re part of the world.
00:29:17 But we are part of the world
00:29:19 and we have the ability to feel like something’s right,
00:29:21 something’s wrong.
00:29:22 And to make a very long story very short,
00:29:25 I think that the idea of moral philosophy
00:29:28 is systematizing our intuitions
00:29:30 of what is right and what is wrong.
00:29:31 And science might be able to predict ahead of time
00:29:34 what we will do,
00:29:36 but it won’t ever be able to judge
00:29:38 whether we should have done it or not.
00:29:39 So, you’re kind of unique in terms of scientists.
00:29:43 Listen, it doesn’t have to do with podcasts,
00:29:45 but even just reaching out,
00:29:47 I think you referred to as sort of
00:29:49 doing interdisciplinary science.
00:29:51 So you reach out and talk to people
00:29:54 that are outside of your discipline,
00:29:55 which I always hope that’s what science was for.
00:30:00 In fact, I was a little disillusioned
00:30:02 when I realized that academia is very siloed.
00:30:06 Yeah.
00:30:07 And so the question is,
00:30:10 how, at your own level,
00:30:13 how do you prepare for these conversations?
00:30:15 How do you think about these conversations?
00:30:16 How do you open your mind enough
00:30:18 to have these conversations?
00:30:20 And it may be a little bit broader,
00:30:21 how can you advise other scientists
00:30:24 to have these kinds of conversations?
00:30:26 Not at the podcast,
00:30:28 the fact that you’re doing a podcast is awesome,
00:30:29 other people get to hear them,
00:30:31 but it’s also good to have it without mics in general.
00:30:34 It’s a good question, but a tough one to answer.
00:30:37 I think about a guy I know who’s a personal trainer,
00:30:40 and he was asked on a podcast,
00:30:43 how do we psych ourselves up to do a workout?
00:30:45 How do we make that discipline to go and work out?
00:30:48 And he’s like, why are you asking me?
00:30:50 I can’t stop working out.
00:30:52 I don’t need to psych myself up.
00:30:54 So, and likewise, he asked me,
00:30:57 how do you get to have interdisciplinary conversations
00:30:59 on all sorts of different things,
00:31:00 all sorts of different people?
00:31:01 I’m like, that’s what makes me go, right?
00:31:04 Like that’s, I couldn’t stop doing that.
00:31:07 I did that long before any of them were recorded.
00:31:09 In fact, a lot of the motivation for starting recording it
00:31:12 was making sure I would read all these books
00:31:14 that I had purchased, right?
00:31:15 Like all these books I wanted to read,
00:31:17 not enough time to read them.
00:31:18 And now if I have the motivation,
00:31:20 cause I’m gonna interview Pat Churchland,
00:31:23 I’m gonna finally read her book.
00:31:25 You know, and it’s absolutely true
00:31:29 that academia is extraordinarily siloed, right?
00:31:31 We don’t talk to people.
00:31:32 We rarely do.
00:31:34 And in fact, when we do, it’s punished.
00:31:36 You know, like the people who do it successfully
00:31:38 generally first became very successful
00:31:41 within their little siloed discipline.
00:31:43 And only then did they start expanding out.
00:31:46 If you’re a young person, you know,
00:31:47 I have graduate students.
00:31:48 I try to be very, very candid with them about this,
00:31:52 that it’s, you know, most graduate students
00:31:55 are to not become faculty members, right?
00:31:57 It’s a tough road.
00:31:59 And so live the life you wanna live,
00:32:03 but do it with your eyes open
00:32:04 about what it does to your job chances.
00:32:06 And the more broad you are
00:32:09 and the less time you spend hyper specializing
00:32:12 in your field, the lower your job chances are.
00:32:15 That’s just an academic reality.
00:32:17 It’s terrible, I don’t like it, but it’s a reality.
00:32:20 And for some people, that’s fine.
00:32:22 Like there’s plenty of people who are wonderful scientists
00:32:24 who have zero interest in branching out
00:32:27 and talking to things, to anyone outside their field.
00:32:30 But it is disillusioning to me.
00:32:33 Some of the, you know, romantic notion I had
00:32:36 of the intellectual academic life
00:32:38 is belied by the reality of it.
00:32:39 The idea that we should reach out beyond our discipline
00:32:43 and that is a positive good is just so rare
00:32:48 in universities that it may as well not exist at all.
00:32:53 But that said, even though you’re saying you’re doing it
00:32:57 like the personal trainer, because you just can’t help it,
00:33:00 you’re also an inspiration to others.
00:33:02 Like I could speak for myself.
00:33:05 You know, I also have a career I’m thinking about, right?
00:33:09 And without your podcast,
00:33:12 I may have not have been doing this at all, right?
00:33:15 So it makes me realize that these kinds of conversations
00:33:19 is kind of what science is about in many ways.
00:33:23 The reason we write papers, this exchange of ideas,
00:33:27 is it’s much harder to do interdisciplinary papers,
00:33:30 I would say.
00:33:31 And conversations are easier.
00:33:35 So conversations is the beginning.
00:33:36 And in the field of AI, it’s obvious
00:33:41 that we should think outside of pure computer vision
00:33:45 competitions on a particular data sets.
00:33:47 We should think about the broader impact
00:33:49 of how this can be, you know, reaching out to physics,
00:33:53 to psychology, to neuroscience and having these
00:33:57 conversations so that you’re an inspiration.
00:34:00 And so never know how the world changes.
00:34:05 I mean, the fact that this stuff is out there
00:34:08 and I’ve a huge number of people come up to me,
00:34:12 grad students, really loving the podcast, inspired by it.
00:34:16 And they will probably have that,
00:34:18 they’ll be ripple effects when they become faculty
00:34:20 and so on and so on.
00:34:21 We can end on a balance between pessimism and optimism.
00:34:25 And Sean, thank you so much for talking to me, it was awesome.
00:34:27 No, Lex, thank you very much for this conversation.
00:34:29 It was great.