vtt/episode_026_small.vtt

WEBVTT

00:00.000 --> 00:02.760
 The following is a conversation with Sean Carroll.

00:02.760 --> 00:04.920
 He's a theoretical physicist at Caltech,

00:04.920 --> 00:08.800
 specializing in quantum mechanics, gravity, and cosmology.

00:08.800 --> 00:11.640
 He's the author of several popular books,

00:11.640 --> 00:15.360
 one on the Arrow of Time called From Eternity to Hear,

00:15.360 --> 00:17.840
 one on the Higgs Boson called Particle

00:17.840 --> 00:19.160
 at the End of the Universe,

00:19.160 --> 00:22.560
 and one on Science and Philosophy called The Big Picture,

00:22.560 --> 00:26.360
 on the origins of life, meaning, and the universe itself.

00:26.360 --> 00:28.720
 He has an upcoming book on quantum mechanics

00:28.720 --> 00:32.760
 that you can preorder now called Something Deeply Hidden.

00:32.760 --> 00:36.040
 He writes one of my favorite blogs on his website,

00:36.040 --> 00:37.960
 preposterousuniverse.com.

00:37.960 --> 00:40.440
 I recommend clicking on the greatest hits link

00:40.440 --> 00:44.400
 that lists accessible, interesting posts on the Arrow of Time,

00:44.400 --> 00:47.600
 dark matter, dark energy, the Big Bang, general relativity,

00:47.600 --> 00:49.560
 string theory, quantum mechanics,

00:49.560 --> 00:53.160
 and the big meta questions about the philosophy of science,

00:53.160 --> 00:57.600
 God, ethics, politics, academia, and much, much more.

00:57.600 --> 01:00.280
 Finally, and perhaps most famously,

01:00.280 --> 01:03.640
 he's the host of a podcast called Mindscape

01:03.640 --> 01:06.920
 that you should subscribe to and support on Patreon.

01:06.920 --> 01:08.800
 Along with the Joe Rogan experience,

01:08.800 --> 01:10.480
 Sam Harris is Making Sense,

01:10.480 --> 01:13.080
 and Dan Carlin's Hardcore History.

01:13.080 --> 01:15.840
 Sean's Mindscape podcast is one of my favorite ways

01:15.840 --> 01:18.800
 to learn new ideas or explore different perspectives

01:18.800 --> 01:22.120
 and ideas that I thought I understood.

01:22.120 --> 01:25.880
 It was truly an honor to meet and spend a couple hours

01:25.880 --> 01:27.200
 with Sean.

01:27.200 --> 01:30.480
 It's a bit heartbreaking to say that for the first time ever,

01:30.480 --> 01:32.760
 the audio recorder for this podcast died

01:32.760 --> 01:34.880
 in the middle of our conversation.

01:34.880 --> 01:36.280
 There are technical reasons for this,

01:36.280 --> 01:38.360
 having to do with phantom power

01:38.360 --> 01:41.040
 that I now understand and will avoid.

01:41.040 --> 01:44.200
 It took me one hour to notice and fix the problem.

01:44.200 --> 01:48.280
 So, much like the universe's 68% dark energy,

01:48.280 --> 01:51.280
 roughly the same amount from this conversation was lost,

01:51.280 --> 01:54.160
 except in the memories of the two people involved

01:54.160 --> 01:56.280
 and in my notes.

01:56.280 --> 01:59.920
 I'm sure we'll talk again and continue this conversation

01:59.920 --> 02:02.440
 on this podcast or on Sean's.

02:02.440 --> 02:05.320
 And of course, I look forward to it.

02:05.320 --> 02:07.840
 This is the Artificial Intelligence podcast.

02:07.840 --> 02:09.960
 If you enjoy it, subscribe on YouTube,

02:09.960 --> 02:12.520
 iTunes, support on Patreon,

02:12.520 --> 02:16.680
 or simply connect with me on Twitter at Lex Freedman.

02:16.680 --> 02:21.360
 And now, here's my conversation with Sean Carroll.

02:21.360 --> 02:23.520
 What do you think is more interesting and impactful?

02:23.520 --> 02:25.840
 Understanding how the universe works

02:25.840 --> 02:26.880
 at a fundamental level

02:26.880 --> 02:29.200
 or understanding how the human mind works?

02:29.200 --> 02:32.440
 You know, of course this is a crazy meaningless

02:32.440 --> 02:33.960
 unanswerable question in some sense,

02:33.960 --> 02:35.160
 because they're both very interesting

02:35.160 --> 02:37.520
 and there's no absolute scale of interestingness

02:37.520 --> 02:39.160
 that we can rate them on.

02:39.160 --> 02:41.160
 There's a glib answer that says the human brain

02:41.160 --> 02:43.080
 is part of the universe, right?

02:43.080 --> 02:44.400
 And therefore, understanding the universe

02:44.400 --> 02:47.000
 is more fundamental than understanding the human brain.

02:47.000 --> 02:49.600
 But do you really believe that once we understand

02:49.600 --> 02:51.520
 the fundamental way the universe works

02:51.520 --> 02:52.680
 at the particle level,

02:52.680 --> 02:55.800
 the forces we would be able to understand how the mind works?

02:55.800 --> 02:56.640
 No, certainly not.

02:56.640 --> 02:58.760
 We cannot understand how ice cream works

02:58.760 --> 03:01.040
 just from understanding how particles work, right?

03:01.040 --> 03:02.760
 So I'm a big believer in emergence.

03:02.760 --> 03:05.320
 I'm a big believer that there are different ways

03:05.320 --> 03:06.640
 of talking about the world

03:07.880 --> 03:11.200
 beyond just the most fundamental microscopic one.

03:11.200 --> 03:13.880
 You know, when we talk about tables and chairs

03:13.880 --> 03:15.120
 and planets and people,

03:15.120 --> 03:16.400
 we're not talking the language

03:16.400 --> 03:18.360
 of particle physics and cosmology.

03:18.360 --> 03:20.880
 So, but understanding the universe,

03:20.880 --> 03:24.040
 you didn't say just at the most fundamental level, right?

03:24.040 --> 03:28.200
 So understanding the universe at all levels is part of that.

03:28.200 --> 03:29.960
 I do think, you know, to be a little bit more fair

03:29.960 --> 03:33.960
 to the question, there probably are general principles

03:33.960 --> 03:38.520
 of complexity, biology, information processing,

03:38.520 --> 03:41.840
 memory, knowledge, creativity

03:41.840 --> 03:45.600
 that go beyond just the human brain, right?

03:45.600 --> 03:47.800
 And maybe one could count understanding those

03:47.800 --> 03:49.120
 as part of understanding the universe.

03:49.120 --> 03:53.040
 The human brain, as far as we know, is the most complex thing

03:53.040 --> 03:54.320
 in the universe.

03:54.320 --> 03:57.440
 So there's, it's certainly absurd to think

03:57.440 --> 03:58.880
 that by understanding the fundamental laws

03:58.880 --> 04:00.400
 of particle physics,

04:00.400 --> 04:02.880
 you get any direct insight on how the brain works.

04:02.880 --> 04:04.360
 But then there's this step

04:04.360 --> 04:06.840
 from the fundamentals of particle physics

04:06.840 --> 04:08.680
 to information processing,

04:08.680 --> 04:10.840
 which a lot of physicists and philosophers

04:10.840 --> 04:12.520
 may be a little bit carelessly take

04:12.520 --> 04:14.680
 when they talk about artificial intelligence.

04:14.680 --> 04:18.080
 Do you think of the universe

04:18.080 --> 04:21.360
 as a kind of a computational device?

04:21.360 --> 04:22.200
 No.

04:22.200 --> 04:24.200
 To be like the honest answer there is no.

04:24.200 --> 04:27.600
 There's a sense in which the universe processes information

04:27.600 --> 04:29.200
 clearly.

04:29.200 --> 04:32.720
 There's a sense in which the universe is like a computer,

04:32.720 --> 04:33.920
 clearly.

04:33.920 --> 04:36.560
 But in some sense, I think,

04:36.560 --> 04:38.560
 I tried to say this once on my blog

04:38.560 --> 04:39.400
 and no one agreed with me,

04:39.400 --> 04:42.400
 but the universe is more like a computation

04:42.400 --> 04:45.080
 than a computer because the universe happens once.

04:45.080 --> 04:46.960
 A computer is a general purpose machine, right?

04:46.960 --> 04:48.680
 You can ask it different questions,

04:48.680 --> 04:50.120
 even a pocket calculator, right?

04:50.120 --> 04:52.960
 And it's set up to answer certain kinds of questions.

04:52.960 --> 04:54.320
 The universe isn't that.

04:54.320 --> 04:57.360
 So information processing happens in the universe,

04:57.360 --> 04:59.120
 but it's not what the universe is.

04:59.120 --> 05:01.560
 And I know your MIT colleague, Seth Lloyd,

05:01.560 --> 05:03.840
 feels very differently about this, right?

05:03.840 --> 05:07.240
 Well, you're thinking of the universe as a closed system.

05:07.240 --> 05:08.080
 I am.

05:08.080 --> 05:11.760
 So what makes a computer more like a PC,

05:13.000 --> 05:14.560
 like a computing machine,

05:14.560 --> 05:17.720
 is that there's a human that comes up to it

05:17.720 --> 05:19.120
 and moves the mouse around,

05:19.120 --> 05:21.680
 so input gives it input.

05:21.680 --> 05:26.320
 And that's why you're saying it's just a computation,

05:26.320 --> 05:29.280
 a deterministic thing that's just unrolling.

05:29.280 --> 05:32.240
 But the immense complexity of it

05:32.240 --> 05:34.440
 is nevertheless like processing.

05:34.440 --> 05:39.440
 There's a state and it changes with rules.

05:40.160 --> 05:41.680
 And there's a sense for a lot of people

05:41.680 --> 05:45.400
 that if the brain operates, the human brain operates

05:45.400 --> 05:46.520
 within that world,

05:46.520 --> 05:49.400
 then it's simply just a small subset of that.

05:49.400 --> 05:52.560
 And so there's no reason we can't build

05:52.560 --> 05:55.600
 arbitrarily great intelligences.

05:55.600 --> 05:56.440
 Yeah.

05:56.440 --> 05:58.720
 Do you think of intelligence in this way?

05:58.720 --> 05:59.640
 Intelligence is tricky.

05:59.640 --> 06:01.720
 I don't have a definition of it offhand.

06:01.720 --> 06:04.640
 So I remember this panel discussion

06:04.640 --> 06:06.240
 that I saw on YouTube, I wasn't there,

06:06.240 --> 06:07.720
 but Seth Lloyd was on the panel.

06:07.720 --> 06:10.560
 And so was Martin Rees, the famous astrophysicist.

06:10.560 --> 06:13.800
 And Seth gave his shtick for why the universe is a computer

06:13.800 --> 06:14.840
 and explained this.

06:14.840 --> 06:19.360
 And Martin Rees said, so what is not a computer?

06:19.360 --> 06:22.000
 And Seth is like, oh, that's a good question.

06:22.000 --> 06:22.840
 I'm not sure.

06:22.840 --> 06:24.960
 Because if you have a sufficiently broad definition

06:24.960 --> 06:28.360
 of what a computer is, then everything is, right?

06:28.360 --> 06:32.160
 And similarly, or the analogy gains force

06:32.160 --> 06:34.560
 when it excludes some things.

06:34.560 --> 06:38.640
 Is the moon going around the earth performing a computation?

06:38.640 --> 06:41.320
 I can come up with definitions in which the answer is yes,

06:41.320 --> 06:43.840
 but it's not a very useful computation.

06:43.840 --> 06:46.120
 I think that it's absolutely helpful

06:46.120 --> 06:49.600
 to think about the universe in certain situations,

06:49.600 --> 06:53.080
 certain contexts, as an information processing device.

06:53.080 --> 06:54.840
 I'm even guilty of writing a paper

06:54.840 --> 06:56.960
 called Quantum Circuit Cosmology, where

06:56.960 --> 06:59.280
 we modeled the whole universe as a quantum circuit.

06:59.280 --> 07:00.320
 As a circuit.

07:00.320 --> 07:01.440
 As a circuit, yeah.

07:01.440 --> 07:02.840
 With qubits kind of thing.

07:02.840 --> 07:05.000
 With qubits, basically, right.

07:05.000 --> 07:07.400
 So in qubits, becoming more and more entangled.

07:07.400 --> 07:09.640
 So do we want to digress a little bit?

07:09.640 --> 07:11.000
 Because this is kind of fun.

07:11.000 --> 07:13.680
 So here's a mystery about the universe

07:13.680 --> 07:16.840
 that is so deep and profound that nobody talks about it.

07:16.840 --> 07:19.040
 Space expands, right?

07:19.040 --> 07:21.880
 And we talk about, in a certain region of space,

07:21.880 --> 07:23.560
 a certain number of degrees of freedom,

07:23.560 --> 07:25.480
 a certain number of ways that the quantum fields

07:25.480 --> 07:28.800
 and the particles in that region can arrange themselves.

07:28.800 --> 07:32.200
 That number of degrees of freedom in a region of space

07:32.200 --> 07:33.800
 is arguably finite.

07:33.800 --> 07:36.640
 We actually don't know how many there are,

07:36.640 --> 07:39.440
 but there's a very good argument that says it's a finite number.

07:39.440 --> 07:44.920
 So as the universe expands and space gets bigger,

07:44.920 --> 07:46.560
 are there more degrees of freedom?

07:46.560 --> 07:48.520
 If it's an infinite number, it doesn't really matter.

07:48.520 --> 07:50.000
 Infinity times 2 is still infinity.

07:50.000 --> 07:53.160
 But if it's a finite number, then there's more space,

07:53.160 --> 07:54.480
 so there's more degrees of freedom.

07:54.480 --> 07:55.760
 So where did they come from?

07:55.760 --> 07:58.000
 That would mean the universe is not a closed system.

07:58.000 --> 08:01.520
 There's more degrees of freedom popping into existence.

08:01.520 --> 08:05.320
 So what we suggested was that there are more degrees of freedom.

08:05.320 --> 08:07.960
 And it's not that they're not there to start,

08:07.960 --> 08:10.880
 but they're not entangled to start.

08:10.880 --> 08:12.800
 So the universe that you and I know of,

08:12.800 --> 08:15.440
 the three dimensions around us that we see,

08:15.440 --> 08:18.080
 we said those are the entangled degrees of freedom

08:18.080 --> 08:19.640
 making up space time.

08:19.640 --> 08:22.640
 As the universe expands, there are a whole bunch of qubits

08:22.640 --> 08:26.840
 in their zero state that become entangled

08:26.840 --> 08:28.720
 with the rest of space time through the action

08:28.720 --> 08:31.200
 of these quantum circuits.

08:31.200 --> 08:37.080
 So what does it mean that there's now more degrees of freedom

08:37.080 --> 08:39.280
 as they become more entangled?

08:39.280 --> 08:40.280
 Yeah.

08:40.280 --> 08:41.640
 As the universe expands.

08:41.640 --> 08:41.960
 That's right.

08:41.960 --> 08:43.280
 So there's more and more degrees of freedom

08:43.280 --> 08:47.320
 that are entangled, that are playing the role of part

08:47.320 --> 08:49.600
 of the entangled space time structure.

08:49.600 --> 08:53.320
 So the underlying philosophy is that space time itself

08:53.320 --> 08:55.600
 arises from the entanglement of some fundamental quantum

08:55.600 --> 08:57.680
 degrees of freedom.

08:57.680 --> 08:58.280
 Wow.

08:58.280 --> 08:59.780
 OK.

08:59.780 --> 09:05.200
 At which point is most of the entanglement happening?

09:05.200 --> 09:07.400
 Are we talking about close to the Big Bang?

09:07.400 --> 09:11.840
 Are we talking about throughout the time of the life of the

09:11.840 --> 09:12.340
 universe?

09:12.340 --> 09:12.840
 Yeah.

09:12.840 --> 09:15.080
 So the idea is that at the Big Bang,

09:15.080 --> 09:17.760
 almost all the degrees of freedom that the universe could

09:17.760 --> 09:22.400
 have were there, but they were unentangled with anything else.

09:22.400 --> 09:23.840
 And that's a reflection of the fact

09:23.840 --> 09:25.560
 that the Big Bang had a low entropy.

09:25.560 --> 09:28.080
 It was a very simple, very small place.

09:28.080 --> 09:31.360
 And as space expands, more and more degrees of freedom

09:31.360 --> 09:34.240
 become entangled with the rest of the world.

09:34.240 --> 09:35.960
 Well, I have to ask John Carroll,

09:35.960 --> 09:38.160
 what do you think of the thought experiment from Nick

09:38.160 --> 09:41.560
 Bostrom that we're living in a simulation?

09:41.560 --> 09:44.880
 So I think let me contextualize that a little bit more.

09:44.880 --> 09:48.320
 I think people don't actually take this thought experiment.

09:48.320 --> 09:50.360
 I think it's quite interesting.

09:50.360 --> 09:52.880
 It's not very useful, but it's quite interesting.

09:52.880 --> 09:55.440
 From the perspective of AI, a lot of the learning

09:55.440 --> 09:59.280
 that can be done usually happens in simulation,

09:59.280 --> 10:01.440
 artificial examples.

10:01.440 --> 10:03.040
 And so it's a constructive question

10:03.040 --> 10:09.360
 to ask how difficult is our real world to simulate,

10:09.360 --> 10:12.400
 which is kind of a dual part of, if we're

10:12.400 --> 10:16.400
 living in a simulation and somebody built that simulation,

10:16.400 --> 10:18.840
 if you were to try to do it yourself, how hard would it be?

10:18.840 --> 10:21.080
 So obviously, we could be living in a simulation.

10:21.080 --> 10:22.960
 If you just want the physical possibility,

10:22.960 --> 10:25.360
 then I completely agree that it's physically possible.

10:25.360 --> 10:27.360
 I don't think that we actually are.

10:27.360 --> 10:31.880
 So take this one piece of data into consideration.

10:31.880 --> 10:35.080
 We live in a big universe.

10:35.080 --> 10:38.480
 There's two trillion galaxies in our observable universe

10:38.480 --> 10:41.640
 with 200 billion stars in each galaxy, et cetera.

10:41.640 --> 10:44.920
 It would seem to be a waste of resources

10:44.920 --> 10:47.600
 to have a universe that big going on just to do a simulation.

10:47.600 --> 10:50.120
 So in other words, I want to be a good Bayesian.

10:50.120 --> 10:54.920
 I want to ask, under this hypothesis, what do I expect to see?

10:54.920 --> 10:56.080
 So the first thing I would say is I

10:56.080 --> 11:00.280
 wouldn't expect to see a universe that was that big.

11:00.280 --> 11:02.560
 The second thing is I wouldn't expect the resolution

11:02.560 --> 11:05.000
 of the universe to be as good as it is.

11:05.000 --> 11:08.960
 So it's always possible that if our superhuman simulators only

11:08.960 --> 11:10.840
 have finite resources that they don't render

11:10.840 --> 11:14.360
 the entire universe, that the part that is out there,

11:14.360 --> 11:17.040
 the two trillion galaxies, isn't actually

11:17.040 --> 11:19.600
 being simulated fully.

11:19.600 --> 11:22.720
 But then the obvious extrapolation of that

11:22.720 --> 11:25.640
 is that only I am being simulated fully.

11:25.640 --> 11:29.240
 The rest of you are just nonplayer characters.

11:29.240 --> 11:30.520
 I'm the only thing that is real.

11:30.520 --> 11:32.720
 The rest of you are just chatbots.

11:32.720 --> 11:34.320
 Beyond this wall, I see the wall,

11:34.320 --> 11:37.360
 but there is literally nothing on the other side of the wall.

11:37.360 --> 11:39.000
 That is sort of the Bayesian prediction.

11:39.000 --> 11:40.400
 That's what it would be like to do

11:40.400 --> 11:42.240
 an efficient simulation of me.

11:42.240 --> 11:45.760
 So none of that seems quite realistic.

11:45.760 --> 11:50.880
 I don't see, I hear the argument that it's just possible

11:50.880 --> 11:53.280
 and easy to simulate lots of things.

11:53.280 --> 11:57.280
 I don't see any evidence from what we know about our universe

11:57.280 --> 11:59.280
 that we look like a simulated universe.

11:59.280 --> 12:01.120
 Now, maybe you can say, well, we don't know what it would

12:01.120 --> 12:03.000
 look like, but that's just abandoning

12:03.000 --> 12:04.520
 your Bayesian responsibilities.

12:04.520 --> 12:07.680
 Like your job is to say, under this theory,

12:07.680 --> 12:09.480
 here's what you would expect to see.

12:09.480 --> 12:11.680
 Yeah, so certainly if you think about a simulation

12:11.680 --> 12:16.680
 as a thing that's like a video game where only a small subset

12:16.680 --> 12:22.880
 is being readied, but say all the laws of physics,

12:22.880 --> 12:26.560
 the entire closed system of the quote unquote universe,

12:26.560 --> 12:27.800
 it had a creator.

12:27.800 --> 12:29.640
 Yeah, it's always possible.

12:29.640 --> 12:32.280
 So that's not useful to think about

12:32.280 --> 12:34.040
 when you're thinking about physics.

12:34.040 --> 12:38.080
 The way Nick Bostrom phrases it, if it's possible

12:38.080 --> 12:40.520
 to simulate a universe, eventually we'll do it.

12:40.520 --> 12:42.720
 Right.

12:42.720 --> 12:45.560
 You can use that, by the way, for a lot of things.

12:45.560 --> 12:49.840
 But I guess the question is, how hard is it

12:49.840 --> 12:52.320
 to create a universe?

12:52.320 --> 12:53.800
 I wrote a little blog post about this,

12:53.800 --> 12:55.440
 and maybe I'm missing something.

12:55.440 --> 12:57.680
 But there's an argument that says not only

12:57.680 --> 13:00.480
 that it might be possible to simulate a universe,

13:00.480 --> 13:05.400
 but probably, if you imagine that you actually

13:05.400 --> 13:07.320
 attribute consciousness and agency

13:07.320 --> 13:09.920
 to the little things that we're simulating,

13:09.920 --> 13:12.400
 to our little artificial beings, there's probably

13:12.400 --> 13:15.000
 a lot more of them than there are ordinary organic beings

13:15.000 --> 13:17.400
 in the universe, or there will be in the future.

13:17.400 --> 13:19.600
 So there's an argument that not only is being a simulation

13:19.600 --> 13:23.520
 possible, it's probable, because in the space

13:23.520 --> 13:25.480
 of all living consciousnesses, most of them

13:25.480 --> 13:26.600
 are being simulated.

13:26.600 --> 13:28.840
 Most of them are not at the top level.

13:28.840 --> 13:30.520
 I think that argument must be wrong,

13:30.520 --> 13:34.080
 because it follows from that argument that if we're simulated,

13:34.080 --> 13:36.880
 but we can also simulate other things.

13:36.880 --> 13:38.800
 Well, but if we can simulate other things,

13:38.800 --> 13:41.800
 they can simulate other things.

13:41.800 --> 13:44.280
 If we give them enough power and resolution,

13:44.280 --> 13:46.000
 and ultimately, we'll reach a bottom,

13:46.000 --> 13:47.800
 because the laws of physics in our universe

13:47.800 --> 13:51.120
 have a bottom, we're made of atoms and so forth.

13:51.120 --> 13:55.080
 So there will be the cheapest possible simulations.

13:55.080 --> 13:57.680
 And if you believe the original argument,

13:57.680 --> 13:59.920
 you should conclude that we should be in the cheapest

13:59.920 --> 14:02.560
 possible simulation, because that's where most people are.

14:02.560 --> 14:03.640
 But we don't look like that.

14:03.640 --> 14:06.880
 It doesn't look at all like we're at the edge of resolution,

14:06.880 --> 14:09.960
 that we're 16 bit things.

14:09.960 --> 14:12.840
 It seems much easier to make much lower level things

14:12.840 --> 14:14.160
 than we are.

14:14.160 --> 14:18.200
 So, and also, I question the whole approach

14:18.200 --> 14:19.840
 to the anthropic principle that says

14:19.840 --> 14:22.320
 we are typical observers in the universe.

14:22.320 --> 14:23.640
 I think that that's not actually,

14:23.640 --> 14:27.320
 I think that there's a lot of selection that we can do

14:27.320 --> 14:30.120
 that were typical within things we already know,

14:30.120 --> 14:32.240
 but not typical within all the universe.

14:32.240 --> 14:35.760
 So do you think there is intelligent life,

14:35.760 --> 14:37.800
 however you would like to define intelligent life

14:37.800 --> 14:39.920
 out there in the universe?

14:39.920 --> 14:44.640
 My guess is that there is not intelligent life

14:44.640 --> 14:46.840
 in the observable universe other than us.

14:48.320 --> 14:52.480
 Simply on the basis of the fact that the likely number

14:52.480 --> 14:56.320
 of other intelligent species in the observable universe,

14:56.320 --> 15:00.280
 there's two likely numbers, zero or billions.

15:01.480 --> 15:02.560
 And if there had been billions,

15:02.560 --> 15:04.000
 you would have noticed already.

15:05.040 --> 15:07.320
 For there to be literally like a small number,

15:07.320 --> 15:12.320
 like Star Trek, there's a dozen intelligent civilizations

15:12.440 --> 15:15.040
 in our galaxy, but not a billion.

15:16.240 --> 15:18.480
 That's weird, that's sort of bizarre to me.

15:18.480 --> 15:21.000
 It's easy for me to imagine that there are zero others

15:21.000 --> 15:22.600
 because there's just a big bottleneck

15:22.600 --> 15:24.960
 to making multicellular life

15:24.960 --> 15:27.040
 or technological life or whatever.

15:27.040 --> 15:28.560
 It's very hard for me to imagine

15:28.560 --> 15:30.160
 that there's a whole bunch out there

15:30.160 --> 15:32.280
 that have somehow remained hidden from us.

15:32.280 --> 15:34.880
 The question I'd like to ask is,

15:34.880 --> 15:37.240
 what would intelligent life look like?

15:37.240 --> 15:41.120
 What I mean by that question and where it's going is,

15:41.120 --> 15:45.120
 what if intelligent life is just fundamentally,

15:45.120 --> 15:49.120
 in some very big ways, different than the one

15:49.120 --> 15:51.480
 that has on Earth.

15:51.480 --> 15:53.880
 That there's all kinds of intelligent life

15:53.880 --> 15:57.560
 that operates at different scales of both size and temporal.

15:57.560 --> 15:59.280
 That's a great possibility

15:59.280 --> 16:00.800
 because I think we should be humble

16:00.800 --> 16:02.640
 about what intelligence is, what life is.

16:02.640 --> 16:04.040
 We don't even agree on what life is,

16:04.040 --> 16:06.040
 much less what intelligent life is, right?

16:06.040 --> 16:08.200
 So that's an argument for humility,

16:08.200 --> 16:10.080
 saying there could be intelligent life

16:10.080 --> 16:12.800
 of a very different character, right?

16:12.800 --> 16:17.240
 You could imagine that dolphins are intelligent

16:17.240 --> 16:19.760
 but never invent space travel

16:19.760 --> 16:20.760
 because they live in the ocean

16:20.760 --> 16:22.760
 and they don't have thumbs, right?

16:22.760 --> 16:25.840
 So they never invent technology, they never invent smelting.

16:26.840 --> 16:31.200
 Maybe the universe is full of intelligent species

16:31.200 --> 16:33.200
 that just don't make technology, right?

16:33.200 --> 16:35.440
 That's compatible with the data, I think.

16:35.440 --> 16:38.560
 And I think maybe what you're pointing at

16:38.560 --> 16:42.560
 is even more out there versions of intelligence,

16:42.560 --> 16:46.240
 you know, intelligence in intermolecular clouds

16:46.240 --> 16:48.160
 or on the surface of a neutron star

16:48.160 --> 16:50.360
 or in between the galaxies in giant things

16:50.360 --> 16:52.840
 where the equivalent of a heartbeat is 100 million years.

16:54.840 --> 16:56.760
 On the one hand, yes,

16:56.760 --> 16:58.560
 we should be very open minded about those things.

16:58.560 --> 17:03.560
 On the other hand, we all of us share the same laws of physics.

17:03.560 --> 17:07.040
 There might be something about the laws of physics

17:07.040 --> 17:08.560
 even though we don't currently know exactly

17:08.560 --> 17:12.560
 what that thing would be that makes meters

17:12.560 --> 17:16.560
 and years the right length and time scales

17:16.560 --> 17:19.560
 for intelligent life, maybe not.

17:19.560 --> 17:22.560
 But we're made of atoms, atoms have a certain size,

17:22.560 --> 17:25.560
 we orbit stars, our stars have a certain lifetime.

17:25.560 --> 17:28.560
 It's not impossible to me that there's a sweet spot

17:28.560 --> 17:30.560
 for intelligent life that we find ourselves in.

17:30.560 --> 17:33.560
 So I'm open minded either way, I'm open minded either being humble

17:33.560 --> 17:35.560
 and there's all sorts of different kinds of life

17:35.560 --> 17:37.560
 or no, there's a reason we just don't know it yet

17:37.560 --> 17:40.560
 why life like ours is the kind of life that's out there.

17:40.560 --> 17:43.560
 Yeah, I'm of two minds too, but I often wonder

17:43.560 --> 17:48.560
 if our brains is just designed to, quite obviously,

17:48.560 --> 17:53.560
 to operate and see the world on these time scales.

17:53.560 --> 17:57.560
 And we're almost blind and the tools we've created

17:57.560 --> 18:01.560
 for detecting things are blind to the kind of observation

18:01.560 --> 18:04.560
 needed to see intelligent life at other scales.

18:04.560 --> 18:06.560
 Well, I'm totally open to that,

18:06.560 --> 18:08.560
 but so here's another argument I would make.

18:08.560 --> 18:10.560
 We have looked for intelligent life,

18:10.560 --> 18:13.560
 but we've looked at for it in the dumbest way we can

18:13.560 --> 18:15.560
 by turning radio telescopes to the sky.

18:15.560 --> 18:20.560
 And why in the world would a super advanced civilization

18:20.560 --> 18:23.560
 randomly beam out radio signals wastefully

18:23.560 --> 18:25.560
 in all directions into the universe?

18:25.560 --> 18:28.560
 It just doesn't make any sense, especially because

18:28.560 --> 18:30.560
 in order to think that you would actually contact

18:30.560 --> 18:33.560
 another civilization, you would have to do it forever.

18:33.560 --> 18:35.560
 You have to keep doing it for millions of years.

18:35.560 --> 18:37.560
 That sounds like a waste of resources.

18:37.560 --> 18:42.560
 If you thought that there were other solar systems

18:42.560 --> 18:45.560
 with planets around them where maybe intelligent life

18:45.560 --> 18:48.560
 didn't yet exist, but might someday,

18:48.560 --> 18:51.560
 you wouldn't try to talk to it with radio waves.

18:51.560 --> 18:53.560
 You would send a spacecraft out there

18:53.560 --> 18:55.560
 and you would park it around there.

18:55.560 --> 18:57.560
 And it would be like, from our point of view,

18:57.560 --> 19:00.560
 it would be like 2001 where there was a monolith.

19:00.560 --> 19:02.560
 There could be an artifact.

19:02.560 --> 19:04.560
 In fact, the other way works also, right?

19:04.560 --> 19:07.560
 There could be artifacts in our solar system

19:07.560 --> 19:11.560
 that have been put there by other technologically advanced

19:11.560 --> 19:14.560
 civilizations, and that's how we will eventually contact them.

19:14.560 --> 19:16.560
 We just haven't explored the solar system well enough yet

19:16.560 --> 19:18.560
 to find them.

19:18.560 --> 19:20.560
 The reason why we don't think about that is because

19:20.560 --> 19:21.560
 we're young and impatient, right?

19:21.560 --> 19:23.560
 It's like it would take more than my lifetime

19:23.560 --> 19:25.560
 to actually send something to another star system

19:25.560 --> 19:27.560
 and wait for it and then come back.

19:27.560 --> 19:30.560
 But if we start thinking on hundreds of thousands of years

19:30.560 --> 19:32.560
 or a million year time scales,

19:32.560 --> 19:34.560
 that's clearly the right thing to do.

19:34.560 --> 19:38.560
 Are you excited by the thing that Elon Musk is doing with SpaceX

19:38.560 --> 19:41.560
 in general, but the idea of space exploration,

19:41.560 --> 19:45.560
 even though you're species is young and impatient?

19:45.560 --> 19:50.560
 No, I do think that space travel is crucially important, long term.

19:50.560 --> 19:52.560
 Even to other star systems.

19:52.560 --> 19:57.560
 And I think that many people overestimate the difficulty

19:57.560 --> 20:00.560
 because they say, look, if you travel 1% the speed of light

20:00.560 --> 20:03.560
 to another star system, we'll be dead before we get there, right?

20:03.560 --> 20:05.560
 And I think that it's much easier.

20:05.560 --> 20:07.560
 And therefore, when they write their science fiction stories,

20:07.560 --> 20:09.560
 they imagine we'd go faster than the speed of light

20:09.560 --> 20:11.560
 because otherwise they're too impatient, right?

20:11.560 --> 20:13.560
 We're not going to go faster than the speed of light,

20:13.560 --> 20:15.560
 but we could easily imagine that the human lifespan

20:15.560 --> 20:17.560
 gets extended to thousands of years.

20:17.560 --> 20:19.560
 And once you do that, then the stars are much closer.

20:19.560 --> 20:20.560
 Effectively, right?

20:20.560 --> 20:22.560
 What's 100 year trip, right?

20:22.560 --> 20:26.560
 So I think that that's going to be the future, the far future,

20:26.560 --> 20:29.560
 not my lifetime once again, but baby steps.

20:29.560 --> 20:31.560
 Unless your lifetime gets extended.

20:31.560 --> 20:33.560
 Well, it's in a race against time, right?

20:33.560 --> 20:37.560
 A friend of mine who actually thinks about these things said,

20:37.560 --> 20:39.560
 you know, you and I are going to die,

20:39.560 --> 20:42.560
 but I don't know about our grandchildren.

20:42.560 --> 20:45.560
 I don't know, predicting the future is hard,

20:45.560 --> 20:47.560
 but that's the least plausible scenario.

20:47.560 --> 20:51.560
 And so, yeah, no, I think that as we discussed earlier,

20:51.560 --> 20:56.560
 there are threats to the earth, known and unknown, right?

20:56.560 --> 21:02.560
 Having spread humanity and biology elsewhere

21:02.560 --> 21:04.560
 is a really important longterm goal.

21:04.560 --> 21:08.560
 What kind of questions can science not currently answer,

21:08.560 --> 21:11.560
 but might soon?

21:11.560 --> 21:14.560
 When you think about the problems and the mysteries before us,

21:14.560 --> 21:17.560
 that may be within reach of science.

21:17.560 --> 21:19.560
 I think an obvious one is the origin of life.

21:19.560 --> 21:21.560
 We don't know how that happened.

21:21.560 --> 21:24.560
 There's a difficulty in knowing how it happened historically,

21:24.560 --> 21:26.560
 actually, you know, literally on earth,

21:26.560 --> 21:29.560
 but starting life from nonlife

21:29.560 --> 21:32.560
 is something I kind of think we're close to, right?

21:32.560 --> 21:33.560
 You really think so?

21:33.560 --> 21:35.560
 Like, how difficult is it to start life?

21:35.560 --> 21:36.560
 I do.

21:36.560 --> 21:40.560
 Well, I've talked to people, including on the podcast, about this.

21:40.560 --> 21:42.560
 You know, life requires three things.

21:42.560 --> 21:44.560
 Life as we know it.

21:44.560 --> 21:46.560
 There's a difference between life, who knows what it is,

21:46.560 --> 21:47.560
 and life as we know it,

21:47.560 --> 21:50.560
 which we can talk about with some intelligence.

21:50.560 --> 21:53.560
 Life as we know it requires compartmentalization.

21:53.560 --> 21:56.560
 You need a little membrane around your cell.

21:56.560 --> 21:58.560
 Metabolism, you need to take in food and eat it

21:58.560 --> 22:00.560
 and let that make you do things.

22:00.560 --> 22:02.560
 And then replication.

22:02.560 --> 22:04.560
 You need to have some information about who you are,

22:04.560 --> 22:07.560
 that you pass down to future generations.

22:07.560 --> 22:11.560
 In the lab, compartmentalization seems pretty easy,

22:11.560 --> 22:13.560
 not hard to make lipid bilayers

22:13.560 --> 22:16.560
 that come into little cellular walls pretty easily.

22:16.560 --> 22:19.560
 Metabolism and replication are hard,

22:19.560 --> 22:21.560
 but replication we're close to.

22:21.560 --> 22:25.560
 People have made RNA like molecules in the lab that...

22:25.560 --> 22:28.560
 I think the state of the art is

22:28.560 --> 22:31.560
 they're not able to make one molecule that reproduces itself,

22:31.560 --> 22:34.560
 but they're able to make two molecules that reproduce each other.

22:34.560 --> 22:37.560
 So that's okay. That's pretty close.

22:37.560 --> 22:40.560
 Metabolism is harder, believe it or not,

22:40.560 --> 22:42.560
 even though it's sort of the most obvious thing,

22:42.560 --> 22:44.560
 but you want some sort of controlled metabolism

22:44.560 --> 22:48.560
 and the actual cellular machinery in our bodies is quite complicated.

22:48.560 --> 22:51.560
 It's hard to see it just popping into existence all by itself.

22:51.560 --> 22:53.560
 It probably took a while.

22:53.560 --> 22:55.560
 But we're making progress.

22:55.560 --> 22:58.560
 In fact, I don't think we're spending nearly enough money on it.

22:58.560 --> 23:01.560
 If I were the NSF, I would flood this area with money

23:01.560 --> 23:04.560
 because it would change our view of the world

23:04.560 --> 23:06.560
 if we could actually make life in the lab

23:06.560 --> 23:09.560
 and understand how it was made originally here on Earth.

23:09.560 --> 23:11.560
 I'm sure it would have some ripple effects

23:11.560 --> 23:13.560
 that help cure diseases and so on.

23:13.560 --> 23:15.560
 That's right.

23:15.560 --> 23:18.560
 Synthetic biology is a wonderful big frontier where we're making cells.

23:18.560 --> 23:21.560
 Right now, the best way to do that

23:21.560 --> 23:23.560
 is to borrow heavily from existing biology.

23:23.560 --> 23:26.560
 Craig Ventner several years ago created an artificial cell,

23:26.560 --> 23:28.560
 but all he did was...

23:28.560 --> 23:30.560
 not all he did, it was a tremendous accomplishment,

23:30.560 --> 23:33.560
 but all he did was take out the DNA from a cell

23:33.560 --> 23:36.560
 and put in entirely new DNA and let it boot up and go.

23:36.560 --> 23:43.560
 What about the leap to creating intelligent life on Earth?

23:43.560 --> 23:45.560
 However, again, we define intelligence, of course,

23:45.560 --> 23:49.560
 but let's just even say homo sapiens,

23:49.560 --> 23:54.560
 the modern intelligence in our human brain.

23:54.560 --> 23:58.560
 Do you have a sense of what's involved in that leap

23:58.560 --> 24:00.560
 and how big of a leap that is?

24:00.560 --> 24:02.560
 So AI would count in this?

24:02.560 --> 24:04.560
 Or do you really want life?

24:04.560 --> 24:06.560
 AI would count in some sense.

24:06.560 --> 24:08.560
 AI would count, I think.

24:08.560 --> 24:10.560
 Of course, AI would count.

24:10.560 --> 24:12.560
 Well, let's say artificial consciousness.

24:12.560 --> 24:14.560
 I do not think we are on the threshold

24:14.560 --> 24:16.560
 of creating artificial consciousness.

24:16.560 --> 24:18.560
 I think it's possible.

24:18.560 --> 24:20.560
 I'm not, again, very educated about how close we are,

24:20.560 --> 24:22.560
 but my impression is not that we're really close

24:22.560 --> 24:24.560
 because we understand how little we understand

24:24.560 --> 24:26.560
 of consciousness and what it is.

24:26.560 --> 24:28.560
 So if we don't have any idea what it is,

24:28.560 --> 24:30.560
 it's hard to imagine we're on the threshold

24:30.560 --> 24:32.560
 of making it ourselves.

24:32.560 --> 24:34.560
 But it's doable, it's possible.

24:34.560 --> 24:36.560
 I don't see any obstacles in principle,

24:36.560 --> 24:38.560
 so yeah, I would hold out some interest

24:38.560 --> 24:40.560
 in that happening eventually.

24:40.560 --> 24:42.560
 I think in general, consciousness,

24:42.560 --> 24:44.560
 I think it would be just surprised

24:44.560 --> 24:46.560
 how easy consciousness is

24:46.560 --> 24:48.560
 once we create intelligence.

24:48.560 --> 24:50.560
 I think consciousness is a thing

24:50.560 --> 24:54.560
 that's just something we all fake.

24:54.560 --> 24:56.560
 Well, good.

24:56.560 --> 24:58.560
 No, actually, I like this idea that, in fact,

24:58.560 --> 25:00.560
 consciousness is way less mysterious than we think

25:00.560 --> 25:02.560
 because we're all at every time,

25:02.560 --> 25:04.560
 at every moment, less conscious than we think we are.

25:04.560 --> 25:06.560
 We can fool things.

25:06.560 --> 25:08.560
 And I think that plus the idea that you

25:08.560 --> 25:10.560
 not only have artificial intelligence systems,

25:10.560 --> 25:12.560
 but you put them in a body,

25:12.560 --> 25:14.560
 give them a robot body,

25:14.560 --> 25:18.560
 that will help the faking a lot.

25:18.560 --> 25:20.560
 Yeah, I think creating consciousness

25:20.560 --> 25:22.560
 in artificial consciousness

25:22.560 --> 25:24.560
 is as simple

25:24.560 --> 25:26.560
 as asking a Roomba

25:26.560 --> 25:28.560
 to say, I'm conscious

25:28.560 --> 25:32.560
 and refusing to be talked out of it.

25:32.560 --> 25:34.560
 It could be.

25:34.560 --> 25:36.560
 I mean, I'm almost being silly,

25:36.560 --> 25:38.560
 but that's what we do.

25:38.560 --> 25:40.560
 That's what we do with each other.

25:40.560 --> 25:44.560
 The consciousness is also a social construct,

25:44.560 --> 25:46.560
 and a lot of our ideas of intelligence

25:46.560 --> 25:48.560
 is a social construct,

25:48.560 --> 25:50.560
 and so reaching that bar involves

25:50.560 --> 25:52.560
 something that's beyond,

25:52.560 --> 25:54.560
 that doesn't necessarily involve

25:54.560 --> 25:56.560
 the fundamental understanding

25:56.560 --> 25:58.560
 of how you go from

25:58.560 --> 26:00.560
 electrons to neurons

26:00.560 --> 26:02.560
 to cognition.

26:02.560 --> 26:04.560
 No, actually, I think that is an extremely good point,

26:04.560 --> 26:06.560
 and in fact,

26:06.560 --> 26:08.560
 what it suggests is,

26:08.560 --> 26:10.560
 so yeah, you referred to Kate Darling,

26:10.560 --> 26:12.560
 who I had on the podcast,

26:12.560 --> 26:14.560
 and who does these experiments with

26:14.560 --> 26:16.560
 very simple robots,

26:16.560 --> 26:18.560
 but they look like animals,

26:18.560 --> 26:20.560
 and they can look like they're experiencing pain,

26:20.560 --> 26:22.560
 and we human beings react

26:22.560 --> 26:24.560
 very negatively to these little robots

26:24.560 --> 26:26.560
 looking like they're experiencing pain,

26:26.560 --> 26:28.560
 and what you want to say is,

26:28.560 --> 26:30.560
 yeah, but they're just robots.

26:30.560 --> 26:32.560
 It's not really pain.

26:32.560 --> 26:34.560
 It's just some electrons going around,

26:34.560 --> 26:36.560
 but then you realize you and I

26:36.560 --> 26:38.560
 are just electrons going around,

26:38.560 --> 26:40.560
 and that's what pain is also.

26:40.560 --> 26:42.560
 What I would have an easy time imagining

26:42.560 --> 26:44.560
 is that there is a spectrum

26:44.560 --> 26:46.560
 between these simple little robots

26:46.560 --> 26:48.560
 that Kate works with

26:48.560 --> 26:50.560
 and a human being,

26:50.560 --> 26:52.560
 where there are things that,

26:52.560 --> 26:54.560
 like a human touring test level thing

26:54.560 --> 26:56.560
 are not conscious,

26:56.560 --> 26:58.560
 but nevertheless walk and talk

26:58.560 --> 27:00.560
 like they're conscious,

27:00.560 --> 27:02.560
 and it could be that the future is,

27:02.560 --> 27:04.560
 I mean, Siri is close, right?

27:04.560 --> 27:06.560
 And so it might be the future

27:06.560 --> 27:08.560
 has a lot more agents like that,

27:08.560 --> 27:10.560
 and in fact, rather than someday going,

27:10.560 --> 27:12.560
 aha, we have consciousness,

27:12.560 --> 27:14.560
 we'll just creep up on it

27:14.560 --> 27:16.560
 with more and more accurate reflections

27:16.560 --> 27:18.560
 of what we expect.

27:18.560 --> 27:20.560
 And in the future, maybe the present,

27:20.560 --> 27:22.560
 and you're basically assuming

27:22.560 --> 27:24.560
 that I'm human.

27:24.560 --> 27:26.560
 I get a high probability.

27:26.560 --> 27:28.560
 At this time, because the,

27:28.560 --> 27:30.560
 but in the future,

27:30.560 --> 27:32.560
 there might be question marks around that, right?

27:32.560 --> 27:34.560
 Yeah, no, absolutely.

27:34.560 --> 27:36.560
 Certainly videos are almost to the point

27:36.560 --> 27:38.560
 where you shouldn't trust them already.

27:38.560 --> 27:40.560
 Photos you can't trust, right?

27:40.560 --> 27:42.560
 Videos is easier to trust,

27:42.560 --> 27:44.560
 but we're getting worse.

27:44.560 --> 27:46.560
 We're getting better at faking them, right?

27:46.560 --> 27:48.560
 Yeah, so physical, embodied people,

27:48.560 --> 27:50.560
 what's so hard about faking that?

27:50.560 --> 27:52.560
 This is very depressing,

27:52.560 --> 27:54.560
 this conversation we're having right now.

27:54.560 --> 27:56.560
 To me, it's exciting.

27:56.560 --> 27:58.560
 You're doing it, so it's exciting to you,

27:58.560 --> 28:00.560
 but it's a sobering thought.

28:00.560 --> 28:02.560
 We're very bad at imagining

28:02.560 --> 28:04.560
 what the next 50 years are going to be like

28:04.560 --> 28:06.560
 when we're in the middle of a phase transition

28:06.560 --> 28:08.560
 as we are right now.

28:08.560 --> 28:10.560
 Yeah, and in general,

28:10.560 --> 28:12.560
 I'm not blind to all the threats.

28:12.560 --> 28:14.560
 I am excited by the power of technology

28:14.560 --> 28:16.560
 to solve,

28:16.560 --> 28:18.560
 as they evolve.

28:18.560 --> 28:20.560
 I'm not as much as Steven Pinker

28:20.560 --> 28:22.560
 optimistic about the world,

28:22.560 --> 28:24.560
 but in everything I've seen,

28:24.560 --> 28:26.560
 all the brilliant people in the world

28:26.560 --> 28:28.560
 that I've met are good people.

28:28.560 --> 28:30.560
 So the army of the good

28:30.560 --> 28:32.560
 in terms of the development of technology is large.

28:32.560 --> 28:34.560
 Okay, you're way more

28:34.560 --> 28:36.560
 optimistic than I am.

28:36.560 --> 28:38.560
 I think that goodness and badness

28:38.560 --> 28:40.560
 are equally distributed among intelligent

28:40.560 --> 28:42.560
 and unintelligent people.

28:42.560 --> 28:44.560
 I don't see much of a correlation there.

28:44.560 --> 28:46.560
 Interesting.

28:46.560 --> 28:48.560
 Neither of us have proof.

28:48.560 --> 28:50.560
 Yeah, exactly. Again, opinions are free, right?

28:50.560 --> 28:52.560
 Nor definitions of good and evil.

28:52.560 --> 28:54.560
 Without definitions

28:54.560 --> 28:56.560
 or without data

28:56.560 --> 28:58.560
 opinions.

28:58.560 --> 29:00.560
 So what kind of questions can science not

29:00.560 --> 29:02.560
 currently answer

29:02.560 --> 29:04.560
 and may never be able to answer in your view?

29:04.560 --> 29:06.560
 Well, the obvious one is what is good and bad.

29:06.560 --> 29:08.560
 What is right and wrong?

29:08.560 --> 29:10.560
 I think that there are questions that science tells us

29:10.560 --> 29:12.560
 what happens, what the world is,

29:12.560 --> 29:14.560
 doesn't say what the world should do

29:14.560 --> 29:16.560
 or what we should do because we're part of the world.

29:16.560 --> 29:18.560
 But we are part of the world

29:18.560 --> 29:20.560
 and we have the ability to feel like

29:20.560 --> 29:22.560
 something's right, something's wrong.

29:22.560 --> 29:24.560
 And to make a very long story

29:24.560 --> 29:26.560
 very short, I think that the idea

29:26.560 --> 29:28.560
 of moral philosophy is

29:28.560 --> 29:30.560
 systematizing our intuitions of what is right

29:30.560 --> 29:32.560
 and what is wrong.

29:32.560 --> 29:34.560
 And science might be able to predict ahead of time

29:34.560 --> 29:36.560
 what we will do,

29:36.560 --> 29:38.560
 but it won't ever be able to judge

29:38.560 --> 29:40.560
 whether we should have done it or not.

29:40.560 --> 29:42.560
 You know, you're kind of unique in terms of scientists.

29:42.560 --> 29:44.560
 It doesn't

29:44.560 --> 29:46.560
 have to do with podcasts, but

29:46.560 --> 29:48.560
 even just reaching out, I think you refer to

29:48.560 --> 29:50.560
 as sort of doing interdisciplinary science.

29:50.560 --> 29:52.560
 So you reach out

29:52.560 --> 29:54.560
 and talk to people

29:54.560 --> 29:56.560
 that are outside of your discipline,

29:56.560 --> 29:58.560
 which I always

29:58.560 --> 30:00.560
 hope that's what science was for.

30:00.560 --> 30:02.560
 In fact, I was a little disillusioned

30:02.560 --> 30:04.560
 when I realized that academia

30:04.560 --> 30:06.560
 is very siloed.

30:06.560 --> 30:08.560
 Yeah.

30:08.560 --> 30:10.560
 The question is

30:10.560 --> 30:12.560
 how,

30:12.560 --> 30:14.560
 at your own level, how do you prepare for these conversations?

30:14.560 --> 30:16.560
 How do you think about these conversations?

30:16.560 --> 30:18.560
 How do you open your mind enough

30:18.560 --> 30:20.560
 to have these conversations?

30:20.560 --> 30:22.560
 And it may be a little bit broader.

30:22.560 --> 30:24.560
 How can you advise other scientists

30:24.560 --> 30:26.560
 to have these kinds of conversations?

30:26.560 --> 30:28.560
 Not at the podcast.

30:28.560 --> 30:30.560
 The fact that you're doing a podcast is awesome.

30:30.560 --> 30:32.560
 Other people get to hear them.

30:32.560 --> 30:34.560
 But it's also good to have it without mics in general.

30:34.560 --> 30:36.560
 It's a good question, but a tough one

30:36.560 --> 30:38.560
 to answer. I think about

30:38.560 --> 30:40.560
 a guy I know is a personal trainer

30:40.560 --> 30:42.560
 and he was asked on a podcast

30:42.560 --> 30:44.560
 how do we psych ourselves up

30:44.560 --> 30:46.560
 to do a workout? How do we make

30:46.560 --> 30:48.560
 that discipline to go and work out?

30:48.560 --> 30:50.560
 And he's like, why are you asking me?

30:50.560 --> 30:52.560
 I can't stop working out.

30:52.560 --> 30:54.560
 I don't need to psych myself up.

30:54.560 --> 30:56.560
 Likewise, you asked me

30:56.560 --> 30:58.560
 how do you get to have

30:58.560 --> 31:00.560
 interdisciplinary conversations and all sorts of different things

31:00.560 --> 31:02.560
 with all sorts of different people?

31:02.560 --> 31:04.560
 That's what makes me go.

31:04.560 --> 31:06.560
 I couldn't stop

31:06.560 --> 31:08.560
 doing that. I did that long before

31:08.560 --> 31:10.560
 any of them were recorded. In fact,

31:10.560 --> 31:12.560
 a lot of the motivation for starting recording it

31:12.560 --> 31:14.560
 was making sure I would read all these books

31:14.560 --> 31:16.560
 that I had purchased. All these books

31:16.560 --> 31:18.560
 I wanted to read. Not enough time to read them.

31:18.560 --> 31:20.560
 And now, if I have the motivation

31:20.560 --> 31:22.560
 because I'm going to interview Pat

31:22.560 --> 31:24.560
 Churchland, I'm going to finally read her

31:24.560 --> 31:26.560
 book.

31:26.560 --> 31:28.560
 And

31:28.560 --> 31:30.560
 it's absolutely true that academia is

31:30.560 --> 31:32.560
 extraordinarily siloed. We don't talk to people.

31:32.560 --> 31:34.560
 We rarely do.

31:34.560 --> 31:36.560
 And in fact, when we do, it's punished.

31:36.560 --> 31:38.560
 The people who do it successfully

31:38.560 --> 31:40.560
 generally first became

31:40.560 --> 31:42.560
 very successful within their little siloed discipline.

31:42.560 --> 31:44.560
 And only then

31:44.560 --> 31:46.560
 did they start expanding out.

31:46.560 --> 31:48.560
 If you're a young person, I have graduate students

31:48.560 --> 31:50.560
 and I try to be very, very

31:50.560 --> 31:52.560
 candid with them about this.

31:52.560 --> 31:54.560
 That it's

31:54.560 --> 31:56.560
 most graduate students do not become faculty members.

31:56.560 --> 31:58.560
 It's a tough road.

31:58.560 --> 32:00.560
 And so

32:00.560 --> 32:02.560
 you live the life you want to live

32:02.560 --> 32:04.560
 but do it with your eyes open

32:04.560 --> 32:06.560
 about what it does to your job chances.

32:06.560 --> 32:08.560
 And the more

32:08.560 --> 32:10.560
 broad you are and the less

32:10.560 --> 32:12.560
 time you spend hyper

32:12.560 --> 32:14.560
 specializing in your field, the lower

32:14.560 --> 32:16.560
 your job chances are. That's just an academic

32:16.560 --> 32:18.560
 reality. It's terrible. I don't like it.

32:18.560 --> 32:20.560
 But it's a reality.

32:20.560 --> 32:22.560
 And for some people

32:22.560 --> 32:24.560
 that's fine. Like there's plenty of people

32:24.560 --> 32:26.560
 who are wonderful scientists who have zero

32:26.560 --> 32:28.560
 interest in branching out and talking to

32:28.560 --> 32:30.560
 things to anyone outside their field.

32:30.560 --> 32:32.560
 But

32:32.560 --> 32:34.560
 it is disillusioning to me

32:34.560 --> 32:36.560
 some of the romantic notion

32:36.560 --> 32:38.560
 I had of the intellectual academic life

32:38.560 --> 32:40.560
 is belied by the reality

32:40.560 --> 32:42.560
 of it. The idea that we should

32:42.560 --> 32:44.560
 reach out beyond our discipline

32:44.560 --> 32:46.560
 and that is a positive good

32:46.560 --> 32:48.560
 is just so

32:48.560 --> 32:50.560
 rare in

32:50.560 --> 32:52.560
 universities that it may as well

32:52.560 --> 32:54.560
 not exist at all. But

32:54.560 --> 32:56.560
 that said, even though you're saying

32:56.560 --> 32:58.560
 you're doing it like the personal trainer

32:58.560 --> 33:00.560
 because you just can't help it, you're also

33:00.560 --> 33:02.560
 an inspiration to others.

33:02.560 --> 33:04.560
 Like I could speak for myself.

33:04.560 --> 33:06.560
 You know,

33:06.560 --> 33:08.560
 I also have a career I'm thinking about

33:08.560 --> 33:10.560
 right. And without

33:10.560 --> 33:12.560
 your podcast, I may have

33:12.560 --> 33:14.560
 not have been doing this at all.

33:14.560 --> 33:16.560
 Right. So it

33:16.560 --> 33:18.560
 makes me realize that these kinds

33:18.560 --> 33:20.560
 of conversations is kind of what science is about.

33:20.560 --> 33:22.560
 In many

33:22.560 --> 33:24.560
 ways. The reason we write papers

33:24.560 --> 33:26.560
 this exchange of ideas

33:26.560 --> 33:28.560
 is much harder to do

33:28.560 --> 33:30.560
 into the disciplinary papers, I would say.

33:30.560 --> 33:32.560
 Yeah. Right.

33:32.560 --> 33:34.560
 And conversations are easier.

33:34.560 --> 33:36.560
 So conversations is the beginning

33:36.560 --> 33:38.560
 and in the field of AI

33:38.560 --> 33:40.560
 that it's

33:40.560 --> 33:42.560
 obvious that we should think outside

33:42.560 --> 33:44.560
 of pure

33:44.560 --> 33:46.560
 computer vision competitions and in particular

33:46.560 --> 33:48.560
 data sets. We should think about the broader

33:48.560 --> 33:50.560
 impact of how this can be

33:50.560 --> 33:52.560
 you know, reaching

33:52.560 --> 33:54.560
 out to physics, to psychology

33:54.560 --> 33:56.560
 to neuroscience

33:56.560 --> 33:58.560
 and having these conversations.

33:58.560 --> 34:00.560
 So you're an inspiration

34:00.560 --> 34:02.560
 and so. Well, thank you very much.

34:02.560 --> 34:04.560
 Never know how the world

34:04.560 --> 34:06.560
 changes. I mean

34:06.560 --> 34:08.560
 the fact that this stuff is out there

34:08.560 --> 34:10.560
 and I've

34:10.560 --> 34:12.560
 a huge number of people come up to me

34:12.560 --> 34:14.560
 grad students really loving the

34:14.560 --> 34:16.560
 podcast inspired by it and

34:16.560 --> 34:18.560
 they will probably have that

34:18.560 --> 34:20.560
 there'll be ripple effects when they become faculty

34:20.560 --> 34:22.560
 and so on. So we can end

34:22.560 --> 34:24.560
 on a balance between pessimism

34:24.560 --> 34:26.560
 and optimism and Sean, thank you so much

34:26.560 --> 34:28.560
 for talking. It was awesome. No, Lex, thank you very

34:28.560 --> 34:52.560
 much for this conversation. It was great.