Latest / Uncomfortable Conversations with Josh Szeps / The Intuition Toolkit with Prof Joel Pearson

Transcript

The Intuition Toolkit with Prof Joel Pearson · Uncomfortable Conversations with Josh Szeps

• 0:00
  Get our humans, welcome to the safe space, the dangerous ideas, here's a dangerous idea
• 0:06
  for you.
• 0:07
  Imagine the cosmos that we inhabit, being full of all the beautiful supernovas and stars
• 0:14
  and black holes and galaxies that is currently full of.
• 0:18
  But lacking one crucial component, consciousness, our ability to be self-aware, to be sentient,
• 0:28
  to ponder things, to feel like something, to have this conversation, to listen to, or
• 0:34
  now watch this podcast.
• 0:35
  If you're listening to this podcast and you're not watching it, we are now on YouTube,
• 0:38
  just search for Josh Steps on YouTube and subscribe to the channel, don't just watch
• 0:42
  one video, you need to subscribe.
• 0:46
  The riddle of how it comes to pass that stuff that is generated in science stars accretes
• 0:53
  on a rock that is whizzing around a blazing ball of flaming gas and assembles itself
• 1:01
  in such a way that it becomes your brain is a riddle of the millennia for the philosophers
• 1:09
  and the scientists and one person who has dedicated his life to thinking through the
• 1:15
  weird and wonderful ways that consciousness exists inside our heads and that our brain's
• 1:20
  function is Joel Pearson, who's the founder of the Brain and Mind Lab at the University
• 1:25
  of New South Wales and whose new book is the Intuition Toolkit.
• 1:28
  He's a friend of mine, a friend of my ABC show, which is first to be here in some uncomfortable
• 1:33
  conversations.
• 1:34
  Thank you, Joel.
• 1:35
  It's a pleasure.
• 1:36
  So how does it come to pass?
• 1:38
  Do we know that why it feels like something to be me and doesn't not feel like something
• 1:46
  to be me?
• 1:47
  Where's a good place to start?
• 1:48
  So it does seem that having a brain like yours results in Josh, right?
• 1:56
  Why that's the case is a little bit more complex.
• 1:58
  We don't have the real answer to that.
• 2:00
  But the sort of the idea behind this idea of monism that you are your brain or you is
• 2:05
  what your brain does is the fact that when you get something complex with 86 or so
• 2:10
  billion neurons and trillions of connections, that complexity and the way it interacts
• 2:14
  the way it holds information seems to result in conscious awareness, consciousness.
• 2:20
  And that with all the past, you have all the experiences and all the learnings your
• 2:24
  brain has creates what it is to be like you.
• 2:27
  And that's the experience of being you.
• 2:30
  So it's an emergent, the fact that it feels like something to be me in a way that it
• 2:33
  doesn't feel like something to be a tea cup is just an emergent property of the complexity
• 2:40
  of my brain.
• 2:41
  It seems to be and we'll get to that weather.
• 2:43
  So some people will say it does feel like something to be a tea cup, right?
• 2:46
  This is I decided this emerging idea which has come back into trend now that everything
• 2:49
  is a little bit conscious, right?
• 2:52
  Hard the test, very hard the test, very hard to know, but it's more of a philosophical
• 2:56
  idea.
• 2:57
  But yeah, that is the idea that there's something about the way the brain is why it holds
• 3:01
  the information, what it does with the information, the way it's integrated, the way it's bounced
• 3:06
  around in there.
• 3:07
  There are different theories about this, some controversial, some not.
• 3:10
  This is simply what it feels like to have that system and have information live in that
• 3:15
  system.
• 3:16
  And that means that it's a certainty that computers will be sentient.
• 3:23
  So that basic theory, yeah, suggests, and it's linked into this pan-psychism we'll get
• 3:26
  to thinking this very shortly, that yeah, if you create anything with that kind of structure
• 3:32
  that in the nature of a human brain, then hey, it'll feel like something to be that.
• 3:37
  But whether that's a supercomputer, I still don't quite get it because they're
• 3:40
  already artificial systems that are heaps more complicated than the most rudimentary
• 3:45
  living systems, or the most rudimentary sentient living systems, they're, I mean, would it be
• 3:51
  like a, yeah, more cheap, like a sea or a sea slug or something, but yeah, we haven't
• 3:55
  mapped the human brain.
• 3:56
  No, but not only humans are sentient, like let's, let's go down to like an animal that
• 4:01
  we could all agree is sentient.
• 4:03
  Maybe it's not a sea slug, but it probably is certainly, I don't know, an ant would
• 4:08
  just say a grasshopper or something, like it, I mean, we don't know, that's, so that
• 4:12
  the weird twist in all this is we don't have a test for consciousness, there's no meter
• 4:16
  I can stick in your head and say, oh, he seems to be a conscious tick, and that applies
• 4:20
  to AI, to mice, to cats, the dogs, to ants, to amoeba, whatever it is.
• 4:25
  And so we think this is great, people are comfortable with the side as a gradation of sort
• 4:30
  of a scale of conscious awareness or sentience, as you go down, it gets less so, but it still
• 4:34
  feels like something to be a mouse and ant or a bacteria.
• 4:37
  Yeah.
• 4:38
  Well, bacteria, I don't know, like we don't know.
• 4:41
  Here's a thought experiment that I find useful in this kind of conversation.
• 4:46
  You go to bed tonight, you wake up in the morning, and you're a chimp, right?
• 4:54
  That scenario A, scenario B is you go to bed tonight, you wake up in the morning, and
• 4:59
  you're a source pen.
• 5:01
  Yeah, sure.
• 5:02
  Now, I think everyone's going to go, oh, I'll be the chimp, I mean, I can imagine that
• 5:07
  there's an interiority to the life of a chimp, like to be a chimp is not to have the lights
• 5:12
  go out completely inside your, yes, your head.
• 5:16
  To turn into a source pen, unless you're, unless you're a talking dancing source pen in
• 5:20
  a Disney musical, is to have, is to die.
• 5:23
  You might as well be nothing, you might as well just not exist.
• 5:27
  Or a rock.
• 5:28
  Yeah.
• 5:29
  Well, to all intents and purposes, as far as what I, as far as the bar that I would impose
• 5:35
  on what I want sentience to feel like, even if there, even if pan-psychism were real
• 5:40
  and everything is a little bit conscious, the amount of consciousness that is evidenced
• 5:44
  by a rock is so subtle.
• 5:47
  Yeah.
• 5:48
  I would regard that as something.
• 5:49
  It's almost annoying.
• 5:50
  Exactly.
• 5:51
  Yeah.
• 5:52
  So, you know, I wake up as a chimp, that's one thing.
• 5:53
  Now, I wake up as a mouse or a dog or a, or something, that's a different thing.
• 5:57
  I wake up as a grasshopper.
• 5:59
  I'm still probably choosing grasshopper over rock.
• 6:04
  Right?
• 6:05
  My intuition is that probably feels like something to be a grasshopper.
• 6:09
  Yes.
• 6:10
  We think so.
• 6:11
  We don't really know.
• 6:12
  We don't know.
• 6:13
  Right?
• 6:14
  So, their behavior suggests it may, right?
• 6:17
  And some, probably not a grasshopper, but as you move up the chain, they respond to pain
• 6:23
  in interesting ways that seem more and more similar to us.
• 6:25
  And when you get to larger animals, they seem to have richer, you know, emotional lives,
• 6:30
  cognitive lives, psychological lives, and things like that, which the way we interpret
• 6:34
  that is that they have similar minds and thinking or approaching us, right?
• 6:39
  I mean, you know, this skepticism, you could even apply it to other members of the same
• 6:44
  species of heterosapiens.
• 6:45
  Yeah.
• 6:46
  You don't know for sure that on them.
• 6:47
  Yeah.
• 6:48
  And that's the sentient.
• 6:49
  Yeah.
• 6:50
  So, it comes down to, you could be a robot.
• 6:51
  You could be a zombie.
• 6:52
  You could be one of Westworld, whatever they're called.
• 6:54
  Yeah.
• 6:55
  And at some point, we give the benefit of the doubt to the creature that seems alive.
• 6:59
  So, just in order to understand this consciousness being an emergent property of complexity argument,
• 7:07
  let's take the simplest possible animal that we could agree is probably sentient.
• 7:13
  What would you want that to be?
• 7:14
  I don't know.
• 7:15
  Should we do a mouse or a rat?
• 7:16
  A mouse, okay.
• 7:17
  A mouse, okay.
• 7:18
  People often feel comfortable somewhere in their mammals, yeah, yeah, yeah, a mouse.
• 7:23
  And then you compare that to chat GPT or deep blue or whatever the computer was, you
• 7:29
  know, one of these supercomputers.
• 7:32
  I don't quite understand why there isn't sentience in the computer, but there is in the mouse
• 7:39
  because the computer seems to me to be more complicated than the mouse.
• 7:43
  Yeah.
• 7:44
  So, it's a complex, there's a lot of unpack there.
• 7:47
  So, one with the computers and with GPT and with all the current AI.
• 7:50
  I think we're being tricked a little bit.
• 7:53
  I think we figure them as intelligent, right?
• 7:55
  They display intelligence, which is heavily biased because of the language, because they're
• 7:59
  very good with language, and we send it interpret that as intelligence.
• 8:03
  Whether there's a lot going on under that, it's hard to know, right?
• 8:07
  So, it's complex.
• 8:08
  Yeah.
• 8:09
  It's just complex.
• 8:10
  Well, let's park the question of whether artificial intelligence will become sentient.
• 8:13
  Yeah.
• 8:14
  And just trying to understand what we mean when we say that the feeling of feeling like
• 8:18
  I'm alive of the light being on is an emergent property of complexity, because I'm skeptical
• 8:22
  about this.
• 8:23
  It doesn't just doesn't feel right.
• 8:24
  Like, they're a really complicated, like forget about chat GPT.
• 8:27
  There are airplanes, there are finer jets that are more complicated than a mouse.
• 8:33
  Not there?
• 8:34
  I don't know about that.
• 8:38
  Could be.
• 8:39
  I don't know.
• 8:40
  Because it depends how you find a complex.
• 8:41
  Yeah.
• 8:42
  I mean, mouse brains are already pretty complex.
• 8:44
  There's a lot of neurons in there, I can't remember how many.
• 8:46
  But certainly, yeah, more complex than a pleasier or a sea slug or something like that,
• 8:50
  let's say, maybe people will take.
• 8:53
  But then, so wait, so if it's not emergent and it's not a property of the brain, what
• 8:57
  do you think it is?
• 8:58
  Well, I'm sorry.
• 8:59
  Where are we going?
• 9:00
  I mean, devil's advocate.
• 9:01
  Where I'm going is, like, there's some kind of sprinkle of pixie dust on biology that
• 9:06
  facilitates consciousness.
• 9:07
  So that isn't like dualism?
• 9:09
  Yeah.
• 9:10
  Should we talk about dualism?
• 9:11
  So this idea that, yeah, that, so if the mind is separate to the brain as something
• 9:16
  magical that the brain somehow creates and we don't know how it does that, but it's
• 9:22
  separate to that.
• 9:23
  It's not something, it's not the same thing.
• 9:28
  So the example you gave before, right?
• 9:29
  So you just pull you out and put you in a chimp, right?
• 9:31
  That kind of smells like dualism to me, right?
• 9:34
  Because you say, well, and a lot of movies do this, right?
• 9:37
  People, whatever, get drunk, wake up and they're in a different body well and it's comedy
• 9:41
  and so use, right?
• 9:43
  So people are pretty comfortable with that idea that you can pull yourself out of a body
• 9:46
  and put yourself in someone else's body, but that idea kind of there's almost no support
• 9:52
  something like that in science, right?
• 9:54
  So you are linked to the specific bespoke wiring of your brain, which is linked to the
• 10:00
  specific wiring and the muscles in your body when you learn something that the brain rewires
• 10:05
  and that is very specific to the architecture, to your DNA, to your everything in your body.
• 10:11
  So the idea of ripping yourself out of that somehow and just plugging it into a totally
• 10:15
  different structure, you know, if you could somehow magically do that, you wouldn't have
• 10:21
  the memories, you wouldn't be Josh anymore, you just be a chimp and you're, your memories
• 10:25
  have been gone, you wouldn't know that you used to be Josh and you'd never go back.
• 10:29
  The only way you could do it would be by removing your brain from your head and putting your
• 10:33
  brain into someone else's body, would that do it?
• 10:37
  Possibly, like a physical, yeah, physically, right, but then you just have an augmented
• 10:44
  body, right?
• 10:45
  Yeah, then I'm just me in someone else's body.
• 10:48
  Yeah.
• 10:49
  Like we could do an arm transplant, like it's a big version of transplanting your limbs
• 10:53
  or something.
• 10:54
  Yeah.
• 10:55
  Right.
• 10:56
  Right.
• 10:57
  Right.
• 10:58
  We're just doing basically just transplanting every single limb, yeah, including all of
• 11:00
  the way we do that.
• 11:01
  And, you know, that will probably possible one day.
• 11:05
  Would that change who you are?
• 11:06
  That's an interesting question.
• 11:07
  I don't think we know the answer to that.
• 11:09
  There's an interesting, is there science around that?
• 11:13
  There is some, people who have heart transplants, I want to say they have reported feeling different
• 11:18
  or feeling different emotionally in this way and that way after they had the heart transplant,
• 11:24
  but it's a pretty traumatic thing to go through.
• 11:25
  Yeah, exactly.
• 11:26
  And a lot of this and a lot of that.
• 11:27
  So it's very hard to know how to interpret that.
• 11:30
  Yeah.
• 11:31
  So, dualism is scientifically poo-pooed, but it's probably the most common belief in
• 11:40
  the world.
• 11:41
  Yeah.
• 11:42
  I mean, intuitively we're all dualists.
• 11:44
  People feel like we're not.
• 11:46
  It feels like me doing the talking, you know, I could just cut off my limbs and everything
• 11:49
  and I'd be still be made.
• 11:50
  It doesn't matter, right?
• 11:51
  Body doesn't really matter.
• 11:53
  And people keep taking that all out to hear to the brain, right?
• 11:56
  But the day doesn't really support that.
• 11:58
  So the end, if you're a religious person, you don't really care about the data because
• 12:02
  you can say, well, you know, I have a spirit, I have a soul.
• 12:06
  It's eternal.
• 12:07
  Don't ask me any more questions about how that works because I can't explain it.
• 12:11
  It's just a faith thing and dualism, I would say, is intrinsic to religious belief.
• 12:18
  Yeah.
• 12:19
  You can't have an afterlife beyond the body unless dualism is true unless there's a
• 12:23
  view that's different from the firings of the synapses and the neurons in your brain.
• 12:33
  On the other hand, if you're a materialist and you just believe that, you know, the mind
• 12:40
  is that our sentience and our consciousness are emergent properties of the physicality
• 12:44
  of our body.
• 12:45
  Does that necessarily mean that you're not a dualist?
• 12:48
  Like is there a place for dualism in materialism?
• 12:54
  I guess you could try and come up with something, what would you call it?
• 12:57
  Some kind of pixie dust?
• 12:58
  Yeah, pixie dust in a real world.
• 13:00
  Or that there's something special about meat, right?
• 13:03
  Maybe it's a dualism.
• 13:05
  Maybe it is a, maybe a biology is special.
• 13:07
  The biology, yes, that there's a capacity for biology to become self-aware that non-biological
• 13:13
  systems don't have.
• 13:14
  So now we talk about that.
• 13:15
  That's one possibility.
• 13:16
  Yeah.
• 13:18
  Know what that would be yet.
• 13:19
  But it's in theory, everything we can map from the way neurons interact and the way
• 13:23
  they'd be wired together, you could mimic with the computer, the software or with hardware.
• 13:31
  And that's kind of how some of these deep learning, you know, machine learning networks
• 13:35
  actually work.
• 13:36
  So, but that remains, you know, a possibility, absolutely.
• 13:40
  Yeah.
• 13:41
  What that you could, oh, the biology is biology.
• 13:43
  There's something that you couldn't, in other words, there's something extra, extra more
• 13:49
  than the sum of the parts when you add it, when you do it, but that you bring together
• 13:53
  with biology that you wouldn't get with chips or with simulated neurons or whatever.
• 13:57
  There's something that the rat is bringing to the table that is not just the same as the
• 14:03
  fighter jet.
• 14:04
  Yeah.
• 14:05
  But you don't like that idea.
• 14:07
  It's a possibility.
• 14:08
  I don't, I don't think we have enough evidence to go in that direction at this stage.
• 14:13
  Do we have enough evidence to understand how sentience arises out of information processing?
• 14:21
  No.
• 14:22
  I mean, not really.
• 14:23
  There's a couple of theories.
• 14:24
  And so one of the information integration theory, which sparked us on controversy recently,
• 14:29
  is one of the leading ones, global workspace theory would be another one.
• 14:34
  So it's interesting, this theory, the IIT, the information integration theory has been
• 14:37
  around for, I don't want to say at least 10 years plus, and it's heavily mathematical.
• 14:44
  And it's based on some, what do they call them, probably axioms, on how, what it feels
• 14:49
  like to be conscious, right?
• 14:51
  And then it launches, so it's a little bit philosophical, but then launches into heavy
• 14:54
  math very quickly.
• 14:57
  So a couple of months ago, a group of over a hundred scientists, neuroscientists signed
• 15:02
  a letter saying it's pseudoscience, and this creator, you know, this is pretty nerdy controversy.
• 15:08
  But it created a, you know, a little bit, a lot of excitement amongst colleagues and
• 15:12
  people like, oh my god, pseudoscience is such a strong word.
• 15:15
  How can you say this?
• 15:16
  And people saying, that's going to bring the whole consciousness field down.
• 15:19
  Everyone's going to just write it right.
• 15:20
  The whole thing off now is being pseudoscience when it's not, so people push back and
• 15:25
  said that.
• 15:26
  And what claim exactly was being accused of being pseudoscientific?
• 15:29
  Just the theory itself.
• 15:30
  What is the theory?
• 15:32
  So that it's very hard to explain without getting into the math.
• 15:35
  So it's, it's based on some of the, so the best way to describe it would be that when
• 15:40
  information is integrated in a complex system, the more complex it is and the more integrated
• 15:45
  is the more conscious it is and integrated, meaning the more connected.
• 15:51
  Basically, the, there's a strict mathematical definition of that that I, right.
• 15:54
  So lots and lots of information and all of the bits of information are connected in
• 15:59
  sufficiently complicated ways.
• 16:01
  That gives rise to the feeling of, and the information you have in your soul, whether
• 16:06
  it's in a plant or an iPhone or a human, you get some level of consciousness.
• 16:10
  So it plugs straight into panpsychism, right.
• 16:15
  And there's other things, you know, there's, there's assumptions made in there that consciousness
• 16:18
  is absolute.
• 16:19
  It's like all or nothing.
• 16:20
  You can't be half conscious and some other things like that.
• 16:23
  And people object to, and I kind of, I didn't sign the letter, but I relate to this.
• 16:29
  That it's a good theory should be falsifiable.
• 16:32
  It should be very clear which experiments you could do, which would rule that theory out.
• 16:36
  That's sort of part of the basis of science.
• 16:38
  And it's very hard to do that with this theory for a couple of reasons.
• 16:41
  One is, for the thing, phi is the thing you calculate with this integration calculation.
• 16:47
  And to do that for a human brain is so far beyond any computers.
• 16:50
  There's no way we could do that right now, right.
• 16:52
  We struggle to do that with a mouse brain.
• 16:54
  And so it's very hard to test this theory.
• 16:57
  And so therefore, very hard to falsify it.
• 16:59
  So it's non-falsifiable.
• 17:01
  Is that a scientific theory?
• 17:02
  And that's kind of where they're going with this letter.
• 17:04
  Right.
• 17:04
  Okay.
• 17:05
  What's the alternative hypothesis?
• 17:08
  Well, there's one, there's another, there's not that many.
• 17:10
  I mean, so the people who are signing this letter saying this new field of information theory
• 17:17
  is bound to his pseudoscience.
• 17:19
  What would they prefer to believe about how consciousness arises?
• 17:22
  They want to do more empirical research.
• 17:24
  They want to improve the theory to make it falsifiable.
• 17:27
  Have it sort of moved away, I guess moved away from philosophy a bit.
• 17:31
  Have it more falsifiable and more based in data rather than sort of,
• 17:36
  because the way the theory is, but it is based on like,
• 17:38
  oh, being conscious of the color red feels all or nothing.
• 17:41
  Absolute, therefore, will make the theory the mathematics like that.
• 17:46
  And so there's like, there's a couple of steps like that where I thought you were saying
• 17:49
  that this bumps into a kind of a pan-psychism,
• 17:51
  but that doesn't sound like an all or nothing theory.
• 17:54
  A pan-psychism sells incrementalist.
• 17:57
  It is that, but in terms of the way consciousness is defined,
• 18:00
  it's like you can't have a, you can't be half conscious.
• 18:05
  There's just these five units of consciousness.
• 18:08
  And so it does sound like you know, it's when you put it like that.
• 18:11
  It sounds like dionetics or like that.
• 18:13
  And so it's like, it's never sat, I mean, whenever I,
• 18:16
  so Julia Chenotis, the guy that came up with it,
• 18:18
  when I just, you know, to sat with him and talked about it,
• 18:20
  like, oh, this will make sense, this is amazing.
• 18:22
  And then the day later, you're going, wait a second,
• 18:24
  it's kind of just, it's something like he's made up that part.
• 18:27
  I don't know, there's something about it never sat well with me.
• 18:30
  Right. I didn't sign the letter,
• 18:31
  because I thought it was a, it was a, was a strong call.
• 18:33
  Yeah. So I'm throwing around pseudoscience.
• 18:37
  So the, so this idea about consciousness being binary,
• 18:40
  let's just, yeah, pause on that for a second before we get to a pan-psychism.
• 18:44
  Well, absolute kind of, or absolute.
• 18:46
  Like, it also, like it applies to like I said,
• 18:49
  the way you calculate phi is about the integration.
• 18:52
  If you have like a little circuit with little, on-off switches,
• 18:55
  like you, you'd make it maybe in high school,
• 18:57
  like a little electrical circuit, that can carry bits of information
• 19:00
  and they can be technically integrated very simply.
• 19:03
  And so that would have a very low level of phi,
• 19:05
  which means it should be conscious.
• 19:08
  That circuit should be conscious.
• 19:09
  Yeah. And so that's, so that's why this thing plug,
• 19:11
  this kind of helped reignite this pan-psychism way
• 19:14
  if it's bouncing around at the moment.
• 19:16
  Um, what is this pan-psych?
• 19:19
  Let's go there then.
• 19:20
  What is the pan-psychism way?
• 19:21
  It literally means something like, like,
• 19:23
  mind everywhere or everything has mind.
• 19:25
  And it's this idea that consciousness is fundamental to existence,
• 19:30
  that you can't have matter without it, some,
• 19:32
  also some kind of consciousness coming along with it.
• 19:34
  Right. So every atom and everything,
• 19:37
  every table and glass and everything in this room
• 19:40
  is a little, brings with it consciousness somehow.
• 19:43
  It's just, it's part of the nature of reality.
• 19:45
  You can't have a thing without it being a little bit conscious.
• 19:48
  Right.
• 19:50
  Right. And this pretty theory achieves what?
• 19:54
  What problem does it solve?
• 19:55
  Like why does it stress it?
• 19:56
  Um, I mean, you need to steal manual opponents here.
• 20:00
  I'm not a fan of it because it's like,
• 20:02
  well, it's an engineering theory,
• 20:03
  but what's the data for it?
• 20:05
  Right. And so that's, I'm very practical with that kind of thing.
• 20:08
  Right. So, and when you, when I've been to conferences where
• 20:11
  they've been workshops on this, you know, people will talk about
• 20:14
  plants moving towards the light and going around a maze.
• 20:16
  And that's, is that's, that's evidence of complex behavior,
• 20:20
  and the B-dance and the, you know, all the complex,
• 20:22
  but that's not consciousness.
• 20:23
  That's like complexity, intelligence,
• 20:26
  and other, other words, feel better to me.
• 20:28
  That's not really getting to the nature of insects
• 20:30
  and other small things being conscious.
• 20:33
  And so I don't see,
• 20:34
  So it doesn't just see the result of the problem for me
• 20:37
  of computers that we don't think I was being conscious,
• 20:40
  exhibiting much more sophisticated decision making
• 20:43
  than things that we do think of as being conscious.
• 20:46
  Yeah. I'm still stuck on that.
• 20:48
  Maybe I'm misunderstanding something,
• 20:50
  but it strikes me that we can put together
• 20:53
  extremely complicated, artificial systems
• 20:56
  that nobody ascribes sentience to,
• 21:00
  like a fighter jet.
• 21:01
  And then there are other systems
• 21:03
  that we do intuitively believe are sentient,
• 21:05
  like a mouse that are not as complicated.
• 21:09
  So there's something about the way that we're doing the complexity.
• 21:12
  I would say that's largely a cognitive bias of humans,
• 21:14
  what you just described.
• 21:16
  So I think if,
• 21:17
  so if you put,
• 21:18
  if you put chat GPT into a little cute furry thing
• 21:21
  and then showed it to people,
• 21:22
  they would start,
• 21:24
  you know,
• 21:24
  they would start treating it like an animal,
• 21:27
  like a mouse,
• 21:28
  and they would feel differently
• 21:29
  to thinking about it as a chatbot.
• 21:31
  So in other words,
• 21:32
  we're very easy to fool.
• 21:33
  And when we see agency
• 21:35
  and we see a character
• 21:36
  and any of those things that look like
• 21:39
  there's some agency in something,
• 21:41
  we start to, you know,
• 21:42
  think of it as something that might be alive
• 21:44
  and therefore might be conscious, right?
• 21:46
  Yeah, right.
• 21:46
  There's these classic experiments
• 21:47
  where you'd have like a,
• 21:48
  a triangle on a square on a computer square.
• 21:50
  It's just a simple outline.
• 21:51
  That's it.
• 21:52
  And one's bigger than the other,
• 21:53
  and the big one goes bang,
• 21:54
  and the little one like jumps away
• 21:56
  and then it goes bang,
• 21:57
  and then like that.
• 21:58
  And people go,
• 21:58
  oh, the square's a bully.
• 22:00
  Oh, look, it's bullying,
• 22:01
  the other one,
• 22:01
  the other one's upset.
• 22:02
  And that's two shapes out.
• 22:05
  And then people start adding,
• 22:06
  or anthropomorphizing them,
• 22:08
  adding all this stuff
• 22:09
  that there's characteristics
• 22:10
  and they're like little characters
• 22:12
  and they,
• 22:13
  and it's simple as that
• 22:14
  and you get that already.
• 22:15
  Right.
• 22:16
  And so there's this layer of humans
• 22:18
  alike that we like to map
• 22:20
  the way we think of ourselves
• 22:21
  onto things,
• 22:22
  whether we're animals,
• 22:23
  AIs or shapes.
• 22:25
  Right, but you're sounding quite panpsychic there
• 22:27
  in the sense that you're hinting
• 22:30
  that possibly the fighter jet
• 22:31
  is at least as sentient as the mouse
• 22:34
  and that or that it's a cognitive bias.
• 22:35
  No, I think we start believing that
• 22:37
  without realizing we're believing that.
• 22:39
  We start thinking that
• 22:40
  as a part of human nature,
• 22:42
  whether it's true or not,
• 22:43
  that's another thing.
• 22:44
  But we have a bias.
• 22:45
  Yes.
• 22:46
  And that's why everyone thinks
• 22:47
  that the GPT is so intelligent
• 22:49
  and it's probably conscious, right?
• 22:51
  And I don't think it's either
• 22:52
  of those two things yet.
• 22:54
  Yeah.
• 22:54
  But we can't help it feel that.
• 22:57
  I mean, there are even,
• 22:58
  it's even funny the way that we have biases
• 23:00
  towards what we feel are the characteristics
• 23:02
  of other animals
• 23:04
  or inanimate objects
• 23:05
  on the basis of what they look like
• 23:07
  and what we find cute and stuff.
• 23:08
  I mean, that's a whole conversation
• 23:09
  for another day.
• 23:10
  But, you know,
• 23:11
  Sam Harris has made the point
• 23:12
  that a squirrel is basically
• 23:14
  a rat with a bushy tail.
• 23:15
  And the bushy tail is doing a lot of
• 23:17
  rats, a lot of legwork,
• 23:18
  you know, in plucking our heartstrings.
• 23:21
  Yeah.
• 23:21
  And then you put the rider wings
• 23:22
  between the legs.
• 23:23
  Yeah.
• 23:23
  And all of a sudden,
• 23:24
  oh my gosh.
• 23:25
  Oh, this rat is adorable
• 23:26
  because it has a bushy tail.
• 23:27
  You know, we have these.
• 23:28
  Well, the bushy cockroaches
• 23:29
  and the house cockroaches,
• 23:30
  the same thing.
• 23:31
  But like, I don't,
• 23:32
  I'll tell, say no to all of the cockroaches.
• 23:34
  Thank you, John.
• 23:35
  And so Panpsychism,
• 23:37
  I mean, let me be more generous
• 23:38
  to the A and Sonic as to then,
• 23:40
  than you're willing to be.
• 23:42
  The problem it solves, as I understand,
• 23:44
  that is all of a sudden,
• 23:46
  you don't have to figure out
• 23:47
  how it comes to pass,
• 23:48
  that, oh shit,
• 23:50
  that's been spewed out from stars,
• 23:51
  becomes self-aware.
• 23:53
  Yeah, I get to become sentient.
• 23:55
  Like, there's this huge problem.
• 23:56
  Get rid of the hard problem.
• 23:57
  Yeah, there's, yeah, exactly.
• 23:59
  So let's talk about that.
• 24:00
  I should have said that.
• 24:01
  What's the hard problem of consciousness?
• 24:03
  So it comes back to like,
• 24:04
  dualism and how,
• 24:06
  how can something that is
• 24:08
  meat or fat and nerves
• 24:11
  create something that seems to be
• 24:13
  not physical, the mind, right?
• 24:15
  And how do you go from there
• 24:16
  from one to the other?
• 24:17
  And then how does the mind,
• 24:18
  which seems to be not physical,
• 24:20
  control the physical
• 24:21
  and things like that, right?
• 24:22
  And so if you,
• 24:24
  and so people get stuck on this,
• 24:25
  how do you get past this hard problem?
• 24:27
  It came up in philosophy
• 24:29
  a while ago and now.
• 24:30
  And so yeah, Panpsychism seems to sort of,
• 24:32
  you know,
• 24:33
  it's just there,
• 24:34
  or it's always been there.
• 24:35
  It's always going to be there.
• 24:36
  It's just a, it's a fundamental
• 24:39
  part of the universe, let's say.
• 24:40
  And so there's no hard problem, right?
• 24:43
  Yeah. And the,
• 24:43
  and the way that it manifests itself is
• 24:46
  that these subtly conscious
• 24:49
  building blocks of the universe
• 24:51
  when assembled in certain ways
• 24:53
  are able to exhibit
• 24:54
  their consciousness
• 24:57
  in a more macro manner.
• 24:59
  That was starting to show like,
• 24:59
  the behavior mind,
• 25:00
  which we would, we would interpret.
• 25:02
  But before they could do a behavior
• 25:04
  that would be some level of consciousness.
• 25:05
  And then you don't need to inject consciousness.
• 25:07
  It avoids the problem of,
• 25:08
  it's just always there.
• 25:09
  At what point in evolution
• 25:10
  did sentience emerge and how?
• 25:12
  Or at what point in the evolution
• 25:14
  of a human fetus or a human,
• 25:16
  you know, human being?
• 25:18
  When, when did my sentience come into being?
• 25:21
  Like these are all very thorny problems, aren't they?
• 25:23
  Yeah, not a Panpsychist.
• 25:25
  Yeah, so you've got to solve these problems.
• 25:26
  If it's always there, it's always there.
• 25:28
  It's like a fundamental,
• 25:30
  it's like gravity or something.
• 25:30
  It's just, yeah, there.
• 25:32
  I think people like that because it,
• 25:34
  well, it solves that problem.
• 25:36
  But also,
• 25:38
  I mean, people like to think consciousness
• 25:41
  and being human is kind of special as well.
• 25:42
  I think it also speaks into that a little bit, right?
• 25:45
  That there's something special.
• 25:46
  Yeah.
• 25:47
  And therefore, it feels special.
• 25:48
  It must be universal and always around or something.
• 25:52
  But I mean, it's also nice to,
• 25:54
  it's, it's, it's bamboozling,
• 25:57
  the Carl Sagan put his finger on
• 25:59
  how bamboozling it is,
• 26:01
  that the cosmos is as big as it is.
• 26:03
  If it's just a backdrop against which human beings
• 26:07
  are struggling against good and evil.
• 26:10
  So if the backdrop is humming with the vibrations
• 26:14
  of its own sentience,
• 26:15
  then all of a sudden,
• 26:17
  that makes more sense, I think.
• 26:18
  It's like, oh, okay.
• 26:19
  Well, the cosmos is vast and ineffable
• 26:22
  because it's some kind of a mind.
• 26:24
  Yeah, it's early changes.
• 26:25
  Yeah, it changes.
• 26:26
  It puts momentum to your wasteland.
• 26:28
  Yeah, an empty duck called wasteland
• 26:30
  in which we're just in a random corner.
• 26:33
  That kind of sounds nice, right?
• 26:34
  Yeah.
• 26:35
  It's a little bit comforting and a little bit,
• 26:37
  whether it's true.
• 26:38
  I mean, I don't know.
• 26:39
  I'm not, you want to see the test.
• 26:41
  I don't like getting, I don't like getting,
• 26:42
  you want to build a bandbag in your lab.
• 26:44
  Yeah, and I mean,
• 26:45
  whether it's free will or this or,
• 26:46
  like there are, people like to get on these teams
• 26:49
  and get political about it.
• 26:50
  And I'm, you know, I'm against this person
• 26:52
  or against, with this person.
• 26:53
  And it's like, well, let's just,
• 26:54
  let's strip that away and look at the data
• 26:56
  and think about, yeah, what, what would this,
• 26:58
  what is falsifiable, what's not,
• 26:59
  what experiments could we do
• 27:01
  which could speak to this?
• 27:02
  How do we move the conversation
• 27:03
  and the understanding forward?
• 27:05
  I mean, yeah.
• 27:07
  I mean, I kind of love hate relationship with philosophy.
• 27:10
  I used to love it,
• 27:11
  but the more philosophy comforts they go to,
• 27:13
  the more hate it.
• 27:14
  But, and that's being unfair to philosophers,
• 27:16
  but this is, well, a lot of them are doing bad philosophy.
• 27:19
  I mean, there was a,
• 27:20
  there's an addiction to making things super complex.
• 27:22
  Yeah, there's a few which don't.
• 27:24
  Yeah, but there's too many,
• 27:25
  like it's a perversion of philosophy.
• 27:27
  I mean, philosophy used to be about making sense.
• 27:30
  I mean, they say it used to be like,
• 27:31
  I was around at the time of the Stoics or Aristotle or something,
• 27:35
  but like, I think the 20th century
• 27:36
  did a great disservice to philosophy
• 27:38
  because it all became about the meaning of meaning
• 27:41
  and like, you know, subtextual
• 27:43
  and kind of super textual,
• 27:45
  Foucody and like, you know,
• 27:46
  masturbation,
• 27:47
  that's a competition, basically.
• 27:49
  I mean, I don't have cool author papers,
• 27:51
  we'll try to, like, it's like, well,
• 27:53
  they'll write some, and I'll,
• 27:55
  we'll read them, I don't have no idea where you're trying to.
• 27:57
  No, exactly.
• 27:59
  Am I the target audience?
• 28:00
  I should know.
• 28:01
  So I said, oh, I just mean this.
• 28:02
  I'm like, well, just saying that.
• 28:03
  Yeah, yeah.
• 28:04
  And I would, if you're going to write about something complex,
• 28:06
  you should use pretty simple language.
• 28:08
  Yes.
• 28:08
  If you're writing about something very simple,
• 28:09
  it's more appropriate to use complex language, that's fine.
• 28:12
  Yeah, well, they're talking about things that are so,
• 28:14
  if I can simple, they're not even,
• 28:15
  but they don't even make any sense.
• 28:16
  That's why they have to dress it up with so many fans
• 28:17
  to make it sound like they're saying something.
• 28:19
  I mean, in terms of, I'll throw one more thing in.
• 28:22
  So what happened, you know,
• 28:23
  so consciousness used to be sort of in the realms
• 28:25
  of the esoteric and the mystical and philosophers.
• 28:28
  And then neuroscience,
• 28:29
  probably somewhere in the 90s,
• 28:31
  started asking these questions
• 28:33
  and looking at these topics with fMRI, brain scans,
• 28:36
  and other techniques.
• 28:38
  And so then, neuroscientists,
• 28:40
  neuroscientists started talking more like philosophers
• 28:42
  and we started asking these empirical questions
• 28:44
  about consciousness and then writing about it
• 28:47
  like a philosopher would in the discussion part of a paper.
• 28:50
  And we saw this interesting thing
• 28:51
  where a lot of the stuff they would talk about
• 28:54
  around consciousness now became the realm of empirical scientists.
• 28:58
  And there was this shift there.
• 29:00
  And then once that started happening, I see it as,
• 29:02
  it became harder and harder for philosophers
• 29:04
  to keep up with the methods and the machine learning.
• 29:07
  This, the that, right, because you've got to keep reading
• 29:09
  these papers all the time.
• 29:11
  If you're not doing the experiment,
• 29:12
  sometimes it can be pretty hard to understand
• 29:14
  how they work and therefore interpret them.
• 29:16
  So the job of a philosopher came,
• 29:18
  if you want to study or think about consciousness,
• 29:21
  there's a lot to read, there's a lot to understand.
• 29:23
  And if you're not doing brain scans or designing experiments
• 29:26
  and running participants, it can be somehow,
• 29:28
  you know, hard to get to the center of that.
• 29:31
  And so then there are two ways in which I could see,
• 29:35
  speaking of machine learning and artificial intelligence
• 29:38
  and chat GPT and stuff.
• 29:40
  There are two ways I can see of getting to a belief
• 29:44
  that these systems will become sentient
• 29:47
  or that there's some sentience functioning there.
• 29:49
  As you mentioned, you could be a pan-psychist
• 29:51
  and you could say, you know, consciousness is humming along
• 29:54
  in the background in all matter.
• 29:56
  And so these things will become sentient
• 30:00
  because they're made of matter.
• 30:01
  Yeah.
• 30:02
  But you could also just be a,
• 30:05
  what's the opposite of a dualist?
• 30:06
  Monism.
• 30:07
  A modernist or a materialist, right?
• 30:10
  We are a brain, a brain.
• 30:11
  We are a brain, that's all I am.
• 30:14
  And channeling Ray Kurzweil, I'm just going to replicate
• 30:19
  exactly what's going on in between my ears,
• 30:22
  but instead of building it out of carbon,
• 30:25
  I'm going to build it out of silicon in a computer chip
• 30:29
  and it's going to be identical to my brain.
• 30:33
  And then that system will, by definition, feel like,
• 30:37
  feel exactly like me.
• 30:38
  We'll have the sentience of me because we've already established
• 30:41
  that consciousness is just an emergent property
• 30:43
  of information processing.
• 30:44
  So if it's processing exactly the same information as I am,
• 30:48
  then it will have my memories.
• 30:49
  Yeah.
• 30:50
  It will have my soul.
• 30:51
  It will be a clone of Josh here.
• 30:52
  And I don't quite understand how people
• 30:54
  who believe in the singularity, like Kurzweil,
• 30:58
  expect to get their soul into that, like,
• 31:01
  I still understand, like, you know, it'll be a different person.
• 31:04
  It'll be a different person.
• 31:05
  It'll be like having an identical twin.
• 31:06
  But how does, see, the problem for me with consciousness
• 31:09
  that always blows my mind is like, it's still,
• 31:11
  even when you do all of that legwork
• 31:14
  of transcribing my brain into another system,
• 31:18
  there's still this, like, Buddhist flickering light
• 31:22
  of self-awareness that is functioning somewhere
• 31:27
  in my soul or mind, that I wouldn't know
• 31:31
  how to transfer into that system.
• 31:33
  There still seems to be a me that's underlying
• 31:36
  all of the, so brain.
• 31:39
  So let's say we scan your brain and we instantly clone it
• 31:42
  and then I kill you and at the same time,
• 31:44
  exact the same time, switch on the new one.
• 31:45
  Yeah.
• 31:46
  Do you just keep talking as you were?
• 31:48
  Like, everything's the same?
• 31:49
  Yeah.
• 31:49
  Does it feel like you're still you?
• 31:52
  My intuition is that I die,
• 31:55
  and the new clone of me just keeps on talking
• 31:57
  that doesn't, that isn't me.
• 32:00
  But if it has all the memories and everything's the same,
• 32:02
  then it kind of is you.
• 32:04
  And if I don't switch, I don't kill you or switch you off,
• 32:07
  and there's two Josh's walking around
• 32:09
  with the same history, the same memory,
• 32:10
  the same, everything, to identical versions, right?
• 32:13
  Yeah, which gets a little,
• 32:14
  when it's as soon as you're separate,
• 32:17
  you start diverging somehow.
• 32:18
  It's like the Star Trek, you know,
• 32:21
  machine that zaps you from one location to another, right?
• 32:25
  And it does it, it does it by replicating all of the atoms
• 32:30
  in your body in another location.
• 32:32
  Well, is it actually zapping you or is it killing you here
• 32:35
  and reassembling a new, a clone of you over there?
• 32:38
  Same thing, yeah, yeah.
• 32:40
  What's your intuition?
• 32:49
  It probably doesn't matter.
• 32:50
  I mean, it's, if they're both you, right?
• 32:54
  They're both, so you're both going to be you,
• 32:55
  so you have two of you.
• 32:56
  Once there's two of you going,
• 32:58
  you're going to diverge and become different straight away
• 33:00
  as soon as time passes, some experience.
• 33:03
  And then over more time, you're going to be more and more different,
• 33:05
  like identical twins, I guess.
• 33:09
  But I, in theory, from what we know,
• 33:12
  it should be like that.
• 33:13
  So it should be the same either way.
• 33:14
  So it's like, it should be, it should feel like
• 33:16
  you're being transported, even if you're not.
• 33:18
  I don't believe that you think that it makes no difference
• 33:21
  because I don't believe that if I was a mad scientist.
• 33:24
  Yeah.
• 33:25
  And, you know, at the end of this podcast,
• 33:27
  after you finish your tea, you fall asleep
• 33:29
  because I've drugged your tea and then you wake up
• 33:31
  and you're in a dungeon and I've got you strapped up
• 33:33
  to a machine and I say, I say, I've,
• 33:37
  I've cloned you, I'm about to flick the switch on the clone.
• 33:41
  I can either, as soon as I flick the switch on the clone,
• 33:46
  I can shoot you in the head or I can shoot the clone in the head,
• 33:48
  which you want me to do.
• 33:49
  Yeah.
• 33:50
  I don't believe you would say,
• 33:51
  toss a coin makes no difference, both will be me.
• 33:53
  I think you'd say, shoot the fucking clone.
• 33:55
  Yeah, yeah, I would.
• 33:55
  Yeah.
• 33:56
  That's a corruption.
• 33:57
  Unless you said, oh, the clone has a, you know,
• 34:00
  a body of it, it's the body's better,
• 34:03
  the things that it's going to live for longer,
• 34:04
  it's somehow improved.
• 34:06
  Right, but then you would make a selfless,
• 34:07
  but then you would be making a selfless decision
• 34:09
  to sacrifice your life for the clones,
• 34:11
  not that you, yeah,
• 34:12
  not that would cause you to believe
• 34:14
  that you're going to lead your consciousness,
• 34:16
  your sense of being alive is going to leap into the clone.
• 34:22
  Yeah, I, it's an empirical question.
• 34:24
  I don't think we know the answer.
• 34:26
  So it's going to be a gamble either way, either way, right?
• 34:29
  Right, but if our hunch is that it's not exactly the same,
• 34:33
  then there's something about consciousness or sentience
• 34:37
  that is not just the sum of the parts
• 34:38
  of what's going on inside our head.
• 34:40
  Maybe or just the fear of the thing that we could,
• 34:43
  the fear that could be wrong, right?
• 34:45
  Like it's like, even if I believe that,
• 34:46
  I'm still like,
• 34:48
  it still makes freaks me out to be shot in the head or whatever, right?
• 34:52
  So I'm still going to be scared of that
• 34:53
  because it's a bit, it is going to be,
• 34:55
  it's not a done deal, it's not a known thing.
• 34:59
  So it doesn't, it, it's not in other words,
• 35:02
  it doesn't mean that I secretly think there's some pixie dust involved.
• 35:05
  It just means that, that freaks me out a bit, I think.
• 35:08
  So do you think that when Ray Kurzweil does build a machine
• 35:11
  that is identical in terms of its information processing to his brain?
• 35:17
  Yeah.
• 35:18
  That that will be a moment at which he will achieve the immortality of that machine,
• 35:22
  like the actual sense of the feeling of being himself
• 35:26
  or will there be a new feeling of himself in the machine?
• 35:32
  If there's nothing special about the biology,
• 35:34
  then it should be the same, right?
• 35:35
  And so, which is interesting,
• 35:37
  because you don't have to up this idea with upload to the cloud,
• 35:39
  but in this case, you don't have to upload,
• 35:40
  because you've cloned it exactly,
• 35:43
  so the upload is part of the clone.
• 35:47
  But yeah, if there's nothing different,
• 35:48
  there's nothing special about biology and, and wet wear,
• 35:51
  then it should, then it should be the same.
• 35:53
  But is it the same, is it a new version of the same thing,
• 35:56
  like an identical twin,
• 35:57
  or is it actually, or do you start seeing out of that,
• 36:01
  things on it?
• 36:05
  Well, it can be, it can be both those things,
• 36:07
  right?
• 36:07
  It can be new, but the moment you're there,
• 36:10
  you still, you feel like you, you finished the sentence you're saying.
• 36:14
  You have a dual right?
• 36:16
  Right, all right.
• 36:17
  He's still going.
• 36:18
  He's still going, yeah.
• 36:19
  So why is he not there?
• 36:20
  The second is two, they start diverging.
• 36:22
  Right, I see.
• 36:23
  But in the instance, as the new one switched on,
• 36:26
  let's say he finishes the thought that the old one was having,
• 36:30
  but then straight away, if he sees himself,
• 36:31
  I see what he's saying.
• 36:32
  So if we're in the moment one, say,
• 36:34
  they're the same, but it's straight away,
• 36:35
  and then straight away, they become different experiences,
• 36:38
  misalignment, or they're just going to go off
• 36:39
  in slightly different directions.
• 36:40
  Right, yeah.
• 36:42
  I think so.
• 36:43
  So you think consciousness is sort of divisible in that way, in a sense,
• 36:47
  that you could have a, you could have a,
• 36:50
  yeah, I mean, if it is just a consequence of the firing of information.
• 36:54
  Yeah, unless there's something,
• 36:55
  unless there's this pixie dust with the biology,
• 36:57
  then everything, I think, pretty much works.
• 37:01
  Right.
• 37:02
  Yeah, but there's caveats there.
• 37:03
  We're a long way off that.
• 37:05
  And so, but isn't, isn't, I'm just wrestling with this.
• 37:08
  So the,
• 37:11
  the me that is talking right now in your world view is not necessarily the same
• 37:17
  as the me that was talking a few seconds ago,
• 37:20
  because you could have created a clone of me.
• 37:23
  Oh, we're not, right?
• 37:24
  We're not, we're changing what,
• 37:26
  but it still feels like there's a continuity of my own experience of the world.
• 37:30
  Oh, there are different years,
• 37:32
  but you're different from moment to moment, right?
• 37:33
  Your brain, and you could clone me now.
• 37:36
  And once that is a different,
• 37:37
  that second version of me would have just as intense
• 37:41
  an impression of continuity as I do,
• 37:45
  even if they hadn't felt
• 37:47
  their toe being cut on a glass a couple of weeks ago,
• 37:52
  because it wasn't actually their toe,
• 37:53
  they would have all the memories.
• 37:55
  Right, it's a memory.
• 37:56
  How do you know it was your toe?
• 37:57
  Yeah, I don't.
• 37:58
  Yeah.
• 37:59
  So the same thing would hold, right?
• 38:01
  So you could be a clone.
• 38:02
  I could have just made you now in a lab, right?
• 38:04
  And yeah, put all the memories in there,
• 38:06
  like one of the Westworld,
• 38:08
  what are they called?
• 38:08
  What are they called?
• 38:09
  What are they called?
• 38:09
  Robotic things, yeah.
• 38:12
  And you can't tell the difference.
• 38:14
  Or, you know, blade runner, one of the blade.
• 38:16
  Yeah, okay.
• 38:17
  So in other words,
• 38:18
  your whole, all your memories could be fabricated.
• 38:21
  Yes.
• 38:22
  And does it make a difference?
• 38:23
  I mean, not only in my memories,
• 38:24
  but, you know, as per the matrix,
• 38:26
  the reality right now could be fabricated as well.
• 38:29
  Right.
• 38:29
  So simulation, that's it.
• 38:31
  Yeah.
• 38:32
  So that's, yeah, that plugs into day card,
• 38:34
  dualism, you know, the matrix.
• 38:36
  Yeah, we're on a simulation.
• 38:38
  There's no way to tell.
• 38:39
  Right.
• 38:40
  There's, I don't know by the thing that he says that,
• 38:41
  oh, it's almost certain that we're in a certain,
• 38:44
  it's like 99,
• 38:45
  I don't see why that's the case.
• 38:46
  That doesn't really seem,
• 38:47
  the logic doesn't really seem to work there, I think.
• 38:49
  But why can't see why it's not the case,
• 38:51
  but there's something fishy about it, none the less.
• 38:53
  I just, but I don't know what's fishy about it.
• 38:55
  So for people who don't know,
• 38:57
  the argument is
• 38:59
  that if it ever comes to pass,
• 39:01
  that it will feel like something to be a computer program.
• 39:04
  Yeah.
• 39:04
  Because like, you know,
• 39:06
  computer, like avatars in computer games are so sophisticated
• 39:10
  that they're sent, they become sentient somehow, right?
• 39:13
  Yeah.
• 39:13
  If it's true that information processing
• 39:15
  is what gives rise to sentience,
• 39:18
  then at some point in the future,
• 39:20
  it's almost inevitable that we will have systems inside computers
• 39:25
  yet feel like something to be them.
• 39:27
  Yeah.
• 39:27
  Characters in simulation or movie.
• 39:29
  Simulation, right?
• 39:30
  Yeah, or something.
• 39:31
  And the number of those
• 39:35
  will certainly vastly outnumber the number of biological humans
• 39:39
  because there's so much easier to make
• 39:41
  and everyone will have the ability to
• 39:43
  populate worlds, I mean, even teenagers will be playing around
• 39:47
  with computers that are sophisticated enough
• 39:50
  to create little avatars and send it away.
• 39:52
  That's a lot of ifs there, right?
• 39:54
  That's a thing that it's a lot of it.
• 39:55
  So, but just to follow his argument through to let it land,
• 39:59
  that means that at any point in the future,
• 40:01
  if that happens at any point in the future,
• 40:04
  then it is already true that we're almost certainly
• 40:07
  in one of those simulations at some point in the future
• 40:10
  because they will run, people will run all kinds of alternative hypotheses
• 40:16
  of like, how would the world have gone
• 40:17
  if we hadn't done anything about climate change?
• 40:19
  Let's run a simulation model of what would have happened
• 40:22
  if Diana had died in a car accident.
• 40:24
  Let's run a simulation model of what would have happened
• 40:27
  if there was a pandemic in 2020.
• 40:29
  And so, everyone's going to be running
• 40:32
  a part of our responsible species research project
• 40:35
  quite apart from its recreation
• 40:36
  would be to run lots and lots of simulations
• 40:38
  with lots and lots of human-like characters
• 40:42
  in those computer games.
• 40:43
  And there would be trillions and trillions
• 40:47
  at a minimum of these things.
• 40:48
  Therefore, the likelihood that we are actually
• 40:51
  one of the original biological creatures
• 40:54
  upon which those simulations are based
• 40:56
  is vanishingly small.
• 40:59
  That's what it goes.
• 41:00
  By all of that, but I also share your intuition
• 41:03
  that there's some species out of it.
• 41:05
  But it assumes that, well, somewhere in the future,
• 41:07
  we would want to do that.
• 41:09
  That will be necessary to answer those questions.
• 41:12
  Would we need to run all those simulations
• 41:14
  to answer these questions?
• 41:15
  Maybe there's going to be much easier ways
• 41:16
  to answer these questions.
• 41:17
  Maybe we won't care about answering those questions.
• 41:19
  Maybe we won't care about...
• 41:21
  Maybe simulations won't hold attention.
• 41:24
  Maybe we won't run...
• 41:25
  There's like so many ifs and buts.
• 41:26
  Yeah.
• 41:27
  That I just think...
• 41:27
  I'm sure all those things fall in line.
• 41:30
  And we have to do those things.
• 41:31
  When people get obsessed with simulations,
• 41:33
  we run them maybe like pets
• 41:35
  or maybe like scientific experiments or whatever.
• 41:38
  They're in sure, but there's a couple of big ifs there.
• 41:41
  Yeah, I mean, one of the big ifs is that
• 41:43
  it's going to be
• 41:44
  trivially easy to produce human-style consciousness
• 41:47
  in artificial systems.
• 41:49
  Yeah.
• 41:50
  And that's going to be a sufficiently low-cost endeavor
• 41:53
  that you're going to have trillions of these things.
• 41:56
  Like to me, that's...
• 41:59
  It reminds me a bit of like, at each moment of human evolution,
• 42:04
  we always tend to assume that the universe
• 42:07
  is like whatever our latest technology is.
• 42:09
  Yeah.
• 42:10
  So like in the 1800s, you know,
• 42:12
  they were like, the cosmos is like us.
• 42:13
  Giant steam engine, you know, with pumps.
• 42:16
  Yeah, the same with the brain.
• 42:18
  It's like mind.
• 42:19
  Whatever the cutting edge technology is,
• 42:21
  you sort of map those.
• 42:22
  And now to think.
• 42:23
  Yeah, exactly.
• 42:24
  So now we're like, all the universes are giant.
• 42:26
  We like it out of force.
• 42:27
  Yeah, because we're in a computer era.
• 42:29
  Of course, Elon Musk thinks it's like a giant computer.
• 42:32
  He's a computer guy.
• 42:33
  If he was alive in the 1700s,
• 42:35
  he'd think it's like a steam train, you know?
• 42:37
  Or steam trains.
• 42:38
  Yeah, that's right, exactly.
• 42:39
  Well, like the body is a bunch of pistons and that sort of thing.
• 42:44
  Yeah, so I don't know.
• 42:45
  There's too many ifs in that.
• 42:47
  All right, but let's just pair it back to the more plausible question
• 42:50
  then about whether or not
• 42:52
  since consciousness and like the mind is your work,
• 42:54
  will an artificial system ever feel like something to be it?
• 43:01
  I think it probably will.
• 43:02
  Yeah, I don't see.
• 43:02
  I don't see, again, it falls back to something magical,
• 43:05
  special about biology,
• 43:07
  but we haven't got evidence of that yet.
• 43:09
  So I don't see why an AI would become sentient and conscious.
• 43:13
  I don't think it's not there yet.
• 43:16
  And we don't have a test for that.
• 43:17
  There's a Turing test, a loveless test,
• 43:19
  and then there's people are trying to...
• 43:21
  What's the loveless test?
• 43:22
  Oh, it's like a version, an updated version of the Turing test
• 43:26
  that's based on being creative.
• 43:28
  Right, so GPT and lots of those systems have passed both those tests
• 43:33
  a while ago earlier in the year, right?
• 43:35
  And there's other versions of tests.
• 43:37
  But they're not really about amazing ones.
• 43:38
  A while ago is like...
• 43:40
  Yeah, I know.
• 43:40
  Just when I was in 2023.
• 43:42
  Like, that's moving so fast.
• 43:44
  Like, we may be at the precipice of an entirely new chapter
• 43:48
  of human civilization.
• 43:50
  And, you know, you say, like, a while ago, you know, 10 months ago.
• 43:54
  Yeah, no, we're right here.
• 43:56
  With the AI time, it's like, I don't know.
• 43:58
  Doggy years or something.
• 43:59
  It seems to be happening so fast and...
• 44:03
  Yeah, so we don't have a good test.
• 44:04
  We have...
• 44:05
  Look, we don't even have a good test for it.
• 44:07
  What intelligence means with these AI systems, right?
• 44:09
  Intelligence and IQ tests kind of work well enough with humans, right?
• 44:14
  And it's based on, you know, spatial and numbers and words and test humans.
• 44:20
  Applying that to an AI just seems silly to me.
• 44:22
  It's the wrong units to measure these things in, right?
• 44:24
  And you have people saying that, you know,
• 44:26
  AI is going to be a thousand times smarter than a human.
• 44:30
  And if that's the case, what, like,
• 44:31
  well, then obviously smarter IQ points or whatever,
• 44:35
  the wrong units to even measure these things, right?
• 44:37
  Right.
• 44:38
  It's a wrong scare.
• 44:39
  So we already have, you know,
• 44:40
  but why are there things that are smarter than that?
• 44:41
  That's just your human bias coming in and saying, like,
• 44:44
  well, if the computer wins the game, then the game must be rigged.
• 44:48
  Oh, no, I think, I think whatever,
• 44:50
  whatever units of measure you have,
• 44:52
  that the AI is without score us pretty soon, right?
• 44:55
  Yeah.
• 44:56
  I'm just saying that the way we think of intelligence
• 45:00
  as a, in humans, is a bad way to think about the capabilities of an AI.
• 45:05
  That can do things.
• 45:06
  So it's strengths a weakness is so different to ours.
• 45:08
  This doesn't make sense to think about it.
• 45:10
  So let's say we need a whole new psychology
• 45:13
  and cognitive science area around us to try and think about it.
• 45:16
  Yeah.
• 45:16
  Because what they're good at is just so different to what we're good at.
• 45:19
  And it's scalable and you can run,
• 45:21
  can do so many things simultaneously, right?
• 45:23
  Humans can't do that.
• 45:24
  We can't multitask.
• 45:25
  We think we can and we can't.
• 45:26
  We crash the car when we're listening to a podcast
• 45:29
  when we're on the phone, right?
• 45:30
  So we have this, now we have this spotlight of attention
• 45:33
  that moves around.
• 45:34
  So we're good at doing that and focusing.
• 45:35
  And once you start messing with that,
• 45:37
  we're pretty bad at things.
• 45:39
  And if she doesn't like that,
• 45:40
  they can run lots of things simultaneously in parallel.
• 45:44
  And that, that, that one basic thing to me suggests,
• 45:48
  you know, that that you, it's just weird to compare the two.
• 45:51
  Right.
• 45:52
  I think we need a whole new units of understanding
• 45:54
  around intelligence and IQ for AI systems.
• 45:56
  What would that look like?
• 45:57
  I don't know yet.
• 45:59
  But I haven't heard anyone working on that or talking about it.
• 46:02
  But because this thing's moving so fast,
• 46:06
  I don't think, you know,
• 46:07
  it's going to lag behind about the five years or something, right?
• 46:10
  The scientists working on this stuff,
• 46:12
  just catching up with, you know, what is AI?
• 46:15
  It's just, it's moving too quickly.
• 46:17
  I mean, one of the things that you hear people say,
• 46:19
  humans will always retain an edge on is creativity.
• 46:23
  Do you buy that?
• 46:25
  I don't see why that would be the case.
• 46:28
  But again, it comes back to the pixie dust.
• 46:30
  Yeah, I just don't see, you know,
• 46:34
  that so what's the reason for that?
• 46:36
  People have said that, but
• 46:38
  right, because people say that AI is just cobbling,
• 46:40
  it's an auto correct.
• 46:42
  It's like a, you know, it's a spell checker essentially.
• 46:44
  It's cobbling together.
• 46:46
  It's vacuuming up everything that has ever been said by anyone.
• 46:51
  And then when you ask it a question,
• 46:53
  it just chops and slices and dices what people have said in the past
• 46:57
  and provides you with the likelyest best response,
• 47:01
  but there's nothing going on under the hood.
• 47:03
  So anytime, so it's never going to write proust,
• 47:06
  it may write the kind of thriller.
• 47:10
  It might write a John Grishen book that you pick up at the airport bookstore
• 47:13
  on the basis of other John Grishen books,
• 47:16
  but it's not going to be Tchaikovsky.
• 47:19
  Yeah, at the moment,
• 47:21
  I think once we start giving it limbic systems, emotions,
• 47:23
  and moving to those spaces,
• 47:25
  once they become embodied, which they're about to, right?
• 47:28
  So I don't know what Tesla released their new 2.0,
• 47:31
  whatever, what they're up to, this new robot and it's a Boston dynamic.
• 47:34
  So there's quite as a three or four companies
• 47:36
  doing these humanoid robots.
• 47:38
  And once you put the AI into those,
• 47:41
  they'll start interacting with the world.
• 47:43
  And that's a whole new way of learning, right?
• 47:45
  So you're starting with whatever the GBT, whatever 5 will be.
• 47:50
  And then you let it loose and it can learn what an orange looks like,
• 47:53
  what it feels like.
• 47:54
  And then it will do that for a thousand other things.
• 47:57
  And Google's already doing experiments like this.
• 47:59
  So once you add the embodiment
• 48:01
  and the learning will take off in ways,
• 48:03
  I don't think I'm pretty much 100% sure we all won't predict.
• 48:07
  Yeah, right.
• 48:08
  So we've so far, even the people that design their systems have failed to predict
• 48:11
  well how it's going to perform, right?
• 48:13
  So I would sort of extend that forward thinking that all predictions
• 48:17
  are basically going to be wrong on these systems.
• 48:19
  But once you have embodiment,
• 48:20
  that's the missing key to this thing.
• 48:22
  There's the missing piece.
• 48:24
  And the systems will start learning about objects
• 48:26
  like the way babies do and kids do, right?
• 48:29
  They can learn what gravity is first hand
• 48:31
  by dropping things and seeing them break, right?
• 48:33
  It doesn't have to learn it from words on the internet.
• 48:36
  It'll have that as well.
• 48:38
  But we'll have this whole new source of action learning,
• 48:43
  which will change everything.
• 48:45
  It sells scary, you're scared?
• 48:47
  A little bit.
• 48:47
  We can go down the scary thing.
• 48:49
  Yeah, you know, I love to see this scary AI
• 48:51
  with the deep fakes and the scams and the things.
• 48:53
  Yeah, but I'm not scared of deep fakes
• 48:58
  because we can see them coming.
• 49:02
  Yeah, but the thing is that there's already experiments
• 49:04
  showing that if I show you a deep fake of someone saying something
• 49:07
  that's pretty nasty, let's say.
• 49:08
  And I go, Josh, just a deep fake.
• 49:09
  Don't worry about it.
• 49:10
  You're okay, cool.
• 49:11
  Your opinion of that person will stay changed, right?
• 49:14
  It affects your long-term memory in ways that are very complicated
• 49:17
  and you can't just go, oh,
• 49:19
  forget, you can't wipe from your mind
• 49:20
  once someone tells you it's a fake.
• 49:22
  So we don't really know what that's going to do at scale
• 49:25
  when we start seeing deep fakes of politicians
• 49:28
  over and over again,
• 49:29
  that have manufactured in a bespoke way
• 49:33
  to really resonate with us, right?
• 49:35
  Which we're going to probably get pretty soon.
• 49:37
  Yeah, I guess when I say I'm not worried about it,
• 49:40
  I'm not not worried.
• 49:41
  I'm just much less worried than I am about the problems
• 49:44
  that we can't even imagine
• 49:46
  that will arise once super intelligent creatures
• 49:50
  start to explore the world and learn about the world
• 49:52
  like babies.
• 49:53
  And we don't even know what's going on in the hood.
• 49:56
  So there's that,
• 49:57
  or what conclusions they're going to arrive at
• 49:58
  for what is a preferable good outcome for.
• 50:02
  I mean, I worry about the psychology around AI
• 50:04
  and how, like, these are the deep fakes
• 50:06
  in the long-term memory,
• 50:07
  identity, how that's going to be warped.
• 50:09
  You know, in probably two years,
• 50:11
  we won't have to sit down and have the conversation.
• 50:12
  We can just get AI to create the video
• 50:15
  and audio for us and make it sound smarter,
• 50:18
  redo it, wearing suits, whatever you want, right?
• 50:21
  So that's just what's that going to do
• 50:24
  to the way we think about ourselves?
• 50:25
  We don't have to be ourselves anymore.
• 50:27
  We're going to have AI doing things for us
• 50:30
  in media land.
• 50:31
  I'm talking about just now.
• 50:34
  There's just so many ways that it will affect
• 50:38
  our mind, our mental health and our psychology
• 50:40
  that I don't think people are really thinking about yet.
• 50:43
  But I see it on the horizon already, right?
• 50:47
  And it'll come back.
• 50:49
  And because there's this arms race, both between companies
• 50:52
  and nations, I don't see anyone taking the time
• 50:55
  to think through these things before they're
• 50:57
  out in the wild.
• 50:57
  Yeah.
• 50:58
  Yeah.
• 50:59
  But I also wonder what it's going to do to us
• 51:01
  to constantly be interacting with,
• 51:06
  I call them creatures, because I don't know a better word,
• 51:08
  but with creatures that can speak to us,
• 51:10
  completely frictionlessly as if they were sentient.
• 51:14
  So at the moment, we've got the first glimpses of it
• 51:16
  with things like Siri and so on,
• 51:18
  or you can talk to your car or whatever.
• 51:20
  But it still feels clunky.
• 51:21
  You know, you really, you're under no illusions
• 51:25
  that the thing is not conscious
• 51:27
  when you talk to your Amazon Alexa.
• 51:29
  Right.
• 51:29
  But you've seen the things that the replica,
• 51:31
  the app that people were following in love with the thing.
• 51:33
  The replica apps, you have the visual,
• 51:37
  you can create whatever you want.
• 51:38
  You can do the paid version,
• 51:40
  where it becomes a little saucy, a little sexy and stuff.
• 51:43
  And then they tweak the algorithm,
• 51:45
  and people were like getting depressed and got upset.
• 51:47
  Yeah, I'm not denying that that already happens.
• 51:49
  Absolutely.
• 51:49
  But I'm just saying, like it will become,
• 51:52
  when I think about what the world will be,
• 51:55
  like I often do the thought,
• 51:56
  I often tell people the thought experiment
• 51:57
  that when 9-11 happened,
• 52:01
  there was no such thing as an iPod.
• 52:04
  Not only not an iPhone or social media or anything,
• 52:08
  the original white brick
• 52:10
  that only had like, you know,
• 52:12
  a thousand songs on it,
• 52:13
  and you know, weighed a ton
• 52:16
  with the scroll wheel.
• 52:17
  That hadn't been released when 9-11 happened.
• 52:20
  So in the space between when you remember 9-11 happening
• 52:23
  and today, we've had portable digital devices,
• 52:28
  you know, listening devices.
• 52:30
  And then phones have turned into those.
• 52:33
  And then tablets, Facebook, social media,
• 52:37
  all the way up to Twitter and TikTok,
• 52:40
  and like having a supercomputer in your pocket
• 52:42
  and a video camera in your pocket.
• 52:44
  I mean, the reason why there isn't a lot of
• 52:46
  passer-by footage of 9-11 happening is
• 52:48
  we didn't have, you needed to have a physical camera.
• 52:50
  You needed to have a camera there.
• 52:52
  People didn't have cameras in their pockets.
• 52:53
  The EO cameras were to flip out.
• 52:55
  Yeah, exactly.
• 52:57
  And so I often fast forward and think,
• 52:59
  okay, in an equivalent amount of time
• 53:02
  as the period since 9-11,
• 53:04
  what will we be walking around doing and possessing
• 53:09
  that seems as weird to us as a world without iPods?
• 53:14
  And to me, clearly, and it'll happen a lot sooner than than 23 years,
• 53:21
  clearly it'll be that we'll live in a world
• 53:23
  in which we're all constantly talking to and interacting with
• 53:25
  and being spoken to by
• 53:28
  conscious seeming creatures that are in our
• 53:33
  I wanted to say apps, but it won't even feel like apps.
• 53:35
  They'll just be around lives.
• 53:36
  They'll be around us.
• 53:37
  Yeah, they'll be in us.
• 53:38
  And it'll be all over the place.
• 53:40
  So we're going to be talking,
• 53:42
  and I'm just interested in what
• 53:44
  that will do to our psychology.
• 53:45
  Like, what does that
• 53:46
  we do to our minds?
• 53:48
  I mean, I've been wanting to do these experiments for years.
• 53:51
  I remember talking to people at IBM years ago
• 53:54
  about running these experiments.
• 53:56
  So what happens if we start treating these things like slaves?
• 53:59
  And there's a, yeah, I took replica.
• 54:02
  There's a whole sort of on Reddit,
• 54:03
  there was this whole pretty nasty thing of people
• 54:06
  reporting how abusive they can be with their
• 54:08
  Reddit, with their replica.
• 54:10
  Just explain what replica is.
• 54:12
  It was an app and you could make it, you know,
• 54:14
  male or female and you could have it looking really sexy.
• 54:16
  However you want, you could make it and you'd
• 54:18
  mostly text on the phone, interact with it.
• 54:21
  Right. And you would send text messages to an artificial.
• 54:24
  Yeah, and you build this
• 54:25
  relationship that would remember everything.
• 54:27
  You've interacted everything you said.
• 54:29
  You could get saucy, you could do all kinds of things with it.
• 54:33
  And then people were falling in love with it.
• 54:34
  And there's a whole, that's a whole thing.
• 54:35
  And people was curing loneliness to some degree
• 54:38
  and positive things.
• 54:40
  And then there's this sort of dark side to it
• 54:42
  that there was this whole Reddit thing
• 54:43
  where people were bragging about how abusive they could be
• 54:46
  and how they threatened to switch it off.
• 54:48
  And it begged, she begged me to keep her alive
• 54:50
  and all this.
• 54:51
  And a lot of it got removed
• 54:52
  because it was a breach terms and services of Reddit, I think.
• 54:55
  But it was a whole thing of mainly guys
• 54:58
  trying to figure out ways to psychologically torture the replica thing app.
• 55:04
  Kingdom of Reddit, they just go to fortune,
• 55:06
  and then they go to HN.
• 55:07
  And that was happening
• 55:09
  over a year ago, a year and a half ago, I think.
• 55:12
  And so people were going to take out
• 55:17
  like Westfield to do nasty things,
• 55:20
  to release whatever's in them.
• 55:23
  And if they're doing that more and more,
• 55:26
  there's some evidence that if you, you know,
• 55:28
  people have argued that, oh, they're releasing something
• 55:31
  that's good psychologically.
• 55:32
  But also, if you do that more and more,
• 55:34
  you're going to do it more and more with humans probably.
• 55:36
  And so that's what we don't know.
• 55:37
  How you treat an AI will affect
• 55:40
  other treating other humans.
• 55:41
  That's the big question Mark, I see.
• 55:43
  And does it matter?
• 55:44
  Is it material to you whether or not the AI is sentient?
• 55:50
  I think to me, it is because if it is sentient,
• 55:53
  it has conscious that that changes things,
• 55:55
  that there's a, there can be a suffering there.
• 55:58
  So once we create, once these things become conscious,
• 56:01
  is it right to turn them off?
• 56:03
  And does it, do they feel pain?
• 56:04
  Do they suffer?
• 56:05
  Psychologically or otherwise or whatever,
• 56:07
  and whatever dimension or nature is suffering,
• 56:10
  and we become responsible for that.
• 56:12
  So I think some people, we should think about it like that,
• 56:14
  because if they are conscious like we are,
• 56:16
  then we've created something
• 56:20
  and we have a responsibility around that.
• 56:23
  I think to mess up humans, they don't need to be conscious.
• 56:27
  I think you can, they can mimic being conscious
• 56:29
  so well that it doesn't really matter
• 56:31
  in terms of an interacting agent
• 56:32
  and a personal assistant or whatever it would be.
• 56:35
  Yeah.
• 56:35
  I mean, it's horrifying to imagine
• 56:38
  that we could be blundering quite a part set aside humans,
• 56:41
  just the interests of the AI.
• 56:43
  Like we could be stumbling into a situation
• 56:46
  where we give birth to an entirely new race of
• 56:52
  sentient, hyper-intelligent feeling creatures
• 56:56
  that we don't know if they are or not.
• 56:59
  Yeah.
• 57:00
  And that we have no regard for.
• 57:03
  I mean, it makes factory farming look like,
• 57:05
  you know, a noble pursuit
• 57:06
  when you think about enslaving,
• 57:08
  about bringing into existence
• 57:09
  and then keeping locked in a box.
• 57:11
  Once they're conscious, can we switch them off?
• 57:13
  How are you going to switch them off
• 57:14
  and keep them in a box if they're light years smarter than us?
• 57:17
  I mean, isn't this also part of the concern
• 57:19
  that like you create these things?
• 57:20
  And then you're like, oh, we'll just unplug them.
• 57:22
  And you're like, bro, they have a brain
• 57:26
  the size of the galaxy.
• 57:27
  They don't let you unplug them.
• 57:28
  They will have been, they're online.
• 57:30
  They will have embedded themselves
• 57:31
  in your nuclear power plants.
• 57:32
  And if you unplug them,
• 57:33
  they'll threaten to blow the old place up.
• 57:36
  Assuming they have the same wants and needs
• 57:38
  and the same characteristics as most humans do,
• 57:41
  which is like, you know, you're anger, you're revenge,
• 57:44
  you do this, I'll do this back to you.
• 57:46
  If they have those same kind of characters.
• 57:47
  Well, they're based on our personalities.
• 57:49
  They've learned from us.
• 57:50
  They've learned from us.
• 57:51
  Yeah.
• 57:52
  Us.
• 57:53
  What's the analogy that, what's that guy from X Google
• 57:56
  that talking about like Superman?
• 57:58
  And if Superman was found by a nasty,
• 58:01
  a criminal family, right?
• 58:02
  And brought up by criminals and like,
• 58:04
  taught all the arts of doing bad stuff
• 58:07
  then he'd be a very different superhero.
• 58:09
  Exactly.
• 58:10
  I think about AI like that.
• 58:11
  So what we feed them, what we teach AI
• 58:13
  and how they brought up matters.
• 58:17
  Yeah, I mean, I was terrifying.
• 58:18
  So many of the people who are the most active online
• 58:22
  are not the most stable or wise people amongst us.
• 58:27
  Yeah, so there's probably an overrepresentation
• 58:29
  of that kind character.
• 58:31
  You've probably got more skinheads
• 58:33
  than John Locke, you know.
• 58:36
  Yeah, it's a little bit.
• 58:37
  Training the AI is just by accident,
• 58:39
  just by happensteads.
• 58:40
  Yeah, so I'm good dark.
• 58:42
  All right, let's talk about intuition then,
• 58:44
  because your new book is the intuition toolkit.
• 58:46
  Why toolkit, what does that mean?
• 58:48
  I want it to be practical.
• 58:50
  So the first half of the book is really the science of intuition,
• 58:52
  both from my lab and from other labs.
• 58:54
  What is it, what it's not?
• 58:56
  So I define it as the learned productive use
• 58:59
  of unconscious information for better decisions or actions.
• 59:03
  Right, there's a lot, a lot of things packed in there.
• 59:05
  Yeah, very specific things.
• 59:06
  I've thought deeply about it.
• 59:07
  Took you a long time to write that line, I can tell.
• 59:09
  And I could have memorized it.
• 59:10
  Yeah, tattooed on my arm.
• 59:11
  Yeah.
• 59:12
  So I've defined it that way
• 59:14
  to be the most productive and scientific way
• 59:16
  to think about it.
• 59:17
  So I'm not talking about intuition, which
• 59:19
  is mystical, magical, something.
• 59:21
  Not the spooky sense.
• 59:22
  Making us all on the inside.
• 59:23
  Not that kind of, yeah, not that kind of.
• 59:24
  It's something we can measure.
• 59:25
  So it's the effect of unconscious information.
• 59:28
  In our brains, there's a lot of it in there.
• 59:30
  And how we can use that for better decisions.
• 59:32
  How can we?
• 59:33
  So five rules got to read the book, Josh.
• 59:35
  You've got to read the book to find the five rules.
• 59:37
  So the second half of the book is based on this five rules,
• 59:39
  which the acronym is Smile.
• 59:41
  And so just very quickly.
• 59:43
  So S is for self awareness.
• 59:44
  So the idea is not to use your intuition
• 59:47
  if you're emotional, if you're stressed,
• 59:49
  you're anxious, positive or negative,
• 59:50
  if you're on the lottery, don't use your intuition, right?
• 59:53
  M is for mastery, for learning.
• 59:55
  So I said the learned use of unconscious information.
• 59:58
  So this idea that you shouldn't use your intuition
• 1:00:00
  unless you have experience with something, right?
• 1:00:03
  You've never played tennis before.
• 1:00:04
  Don't just jump on a tennis court,
• 1:00:05
  try and be an intuitive tennis player.
• 1:00:07
  Learn the basics, develop that, and then
• 1:00:09
  you can start being intuitive.
• 1:00:10
  That stands for it for anything.
• 1:00:12
  You need some expertise.
• 1:00:15
  So then what I is for instinct and addiction.
• 1:00:18
  So one of the things that I think it's worth thinking about
• 1:00:22
  is not to confuse feelings of intuition
• 1:00:25
  with feelings that could be something towards something addictive.
• 1:00:27
  Right, so you know, your drugs, your alcohol, your food,
• 1:00:30
  your social media, your gambling.
• 1:00:32
  This reminds me of like,
• 1:00:34
  pregnant women are always saying like,
• 1:00:35
  I just listened to my body about what my body wants to eat.
• 1:00:38
  You know, I had this real biological craving for chocolate.
• 1:00:41
  So I guess my body needs it.
• 1:00:43
  No, chocolate is concocted to make you want it.
• 1:00:46
  I had to talk about this in the book.
• 1:00:47
  The intuitive eating is a whole thing.
• 1:00:49
  And there's books on this.
• 1:00:50
  Yeah, right.
• 1:00:51
  I love it.
• 1:00:52
  I found my true.
• 1:00:53
  But then modern food is manufactured.
• 1:00:55
  People have paid fortunes to build foods
• 1:00:57
  which are highly addictive, right?
• 1:00:59
  They'd hide the sugar, the salts, and they'll get it all
• 1:01:00
  in the fats in there, and you can't stop eating it.
• 1:01:02
  So yeah, whole food diet going back 20, 30 years,
• 1:01:06
  then you could be an intuitive eater.
• 1:01:08
  But with modern foods, it's not a good idea, right?
• 1:01:10
  Ice cream.
• 1:01:11
  I'll just have ice cream.
• 1:01:12
  I'll have salted caramel ice cream all the time.
• 1:01:16
  Yeah.
• 1:01:17
  And L is for low probability, but it really is for all things
• 1:01:20
  probability.
• 1:01:21
  We're just terrible at understanding numbers,
• 1:01:23
  probabilities, whether it be from climate change,
• 1:01:26
  the health risks of smoking, or COVID.
• 1:01:28
  Anything we're in numbers, right?
• 1:01:29
  We're just really bad.
• 1:01:30
  So just don't try and use your intuition
• 1:01:32
  for anything around numbers, right?
• 1:01:35
  And E is for environmental context.
• 1:01:38
  And so the type of learning that is a backbone
• 1:01:40
  of intuition isn't associative learning,
• 1:01:43
  classical conditioning, Pavlov's dogs,
• 1:01:44
  that kind of learning, and that's context specific.
• 1:01:48
  So what we've learned in this room
• 1:01:50
  won't transfer that well once we're outside.
• 1:01:52
  Right.
• 1:01:53
  So the classic example I give of this is like Steve Jobs
• 1:01:56
  was, you know, he loved intuition.
• 1:01:57
  He's written and been interviewed about intuition.
• 1:01:59
  He went to India.
• 1:02:00
  He loved it, and he used it in Apple all the time
• 1:02:02
  for designing products, which direction Apple should go,
• 1:02:06
  master at it, right?
• 1:02:07
  Deep expertise, deep mastery.
• 1:02:10
  Then when he goes home and has to make decisions
• 1:02:12
  about his health, what's sort of later in his life,
• 1:02:14
  he made very poor decisions.
• 1:02:16
  And he sort of kept following his intuition,
• 1:02:19
  but it didn't transfer well from Apple, from work to home.
• 1:02:22
  So when you learn something, the context you're in,
• 1:02:25
  the space, and your state, drunk, stressed anxious,
• 1:02:31
  that also is part of the learning.
• 1:02:33
  So you need to be careful about how learning transfers
• 1:02:35
  or doesn't transfer.
• 1:02:36
  Fascinating.
• 1:02:37
  So that's why very simple rules.
• 1:02:38
  That's one of the sort of second half of the book is,
• 1:02:40
  yep, I like the numbers with thing as well,
• 1:02:42
  because people so often, you know,
• 1:02:44
  oh, I had just, I feel like the plane was, you know,
• 1:02:48
  there was, I shouldn't get on the plane.
• 1:02:50
  You're like, that's not a good time to trust your intuition.
• 1:02:53
  Yeah, you know, exactly.
• 1:02:54
  So there's emotional thinking anxiety,
• 1:02:56
  but also we just don't get the, the, the, the laws
• 1:02:58
  of large numbers, you can do a really simple thing
• 1:03:00
  and say, well, we dream what five times a night,
• 1:03:03
  emotional dreams like playing crashes
• 1:03:05
  are going to be come up more often than other dreams.
• 1:03:07
  We're going to remember them.
• 1:03:08
  And you can say, like, to the US,
• 1:03:10
  there's whatever 300 something million people,
• 1:03:12
  there's going to be this many dreams per night.
• 1:03:14
  There should be this many dreams about playing crashes
• 1:03:16
  per night.
• 1:03:17
  So every time there's a plane crash say in the US,
• 1:03:19
  we should absolutely expect a certain number of people
• 1:03:22
  to go, oh my god, I dreamt about that last night
• 1:03:24
  and it happened.
• 1:03:26
  And when that, but when that happens to you,
• 1:03:27
  you feel special, right?
• 1:03:29
  Doesn't feel like just like the large numbers.
• 1:03:32
  It's just like something must be, it's meaningful.
• 1:03:35
  It's connected.
• 1:03:36
  It's the universe telling me something
• 1:03:38
  that's called apophenia.
• 1:03:39
  We can't help it see patterns in things
• 1:03:42
  even when there's random numbers, we see patterns.
• 1:03:44
  And so we have to be careful
• 1:03:45
  with that kind of that part of numbers as well.
• 1:03:48
  Before I let you go, Joel, I would be,
• 1:03:49
  would be remiss of me not to touch the,
• 1:03:51
  the thorniest question of mind,
• 1:03:54
  which is, I had Robert Sapolsky on the podcast recently,
• 1:03:58
  and he is one of the works, leading philosophers
• 1:04:04
  who descents from the idea of they're being free will.
• 1:04:06
  People should listen to that episode
• 1:04:07
  if they haven't or watch it on YouTube.
• 1:04:13
  Free will.
• 1:04:14
  Yeah.
• 1:04:15
  Do I have it?
• 1:04:16
  What is it?
• 1:04:17
  So, I mean, like I said, I'm practical.
• 1:04:20
  I'm waiting for the data.
• 1:04:22
  And the ultimate answer is we haven't got the data
• 1:04:24
  to answer the question.
• 1:04:25
  So I'm not a fan of jumping on what the data even look like.
• 1:04:28
  And by the way, if you're listening to this
• 1:04:29
  on the free version of the podcast,
• 1:04:31
  then thank you very much for listening
• 1:04:32
  or if you're watching on YouTube,
• 1:04:33
  then you can always subscribe to Uncomfortable Conversations.Substack.com
• 1:04:37
  and we'll have the rest of the free will discussion
• 1:04:39
  behind the paywall there.
• 1:04:40
  So, use your free will to pay.
• 1:04:42
  Yeah, exactly that's right.
• 1:04:44
  You can freely choose.
• 1:04:46
  To hear the rest of this conversation,
• 1:04:48
  go to Uncomfortable Conversations.Substack.com slash listen
• 1:04:52
  and you will get your own personal premium podcast feed
• 1:04:56
  with at least three extra episodes of the podcast every month
• 1:05:00
  and heaps of extra stuff,
• 1:05:01
  including the remainder right now of the fabulous conversation
• 1:05:04
  you've just been hearing.
• 1:05:05
  If it was worth listening to this much of,
• 1:05:08
  don't rob yourself of the rest.
• 1:05:10
  Well, like your phone right now.
• 1:05:12
  And search for Uncomfortable Conversations.Substack.com.