1
00:00:05,160 --> 00:00:08,960
Speaker 1: What is software code and can it be thought of

2
00:00:09,119 --> 00:00:12,000
Speaker 1: like a magic spell? Are we in the process of

3
00:00:12,039 --> 00:00:15,480
Speaker 1: building a world so complex that we will lose the

4
00:00:15,520 --> 00:00:18,320
Speaker 1: ability to understand it? Or has that already happened a

5
00:00:18,360 --> 00:00:20,319
Speaker 1: long time ago? And what does any of this have

6
00:00:20,400 --> 00:00:24,279
Speaker 1: to do with SimCity or knowledge that already exists but

7
00:00:24,400 --> 00:00:27,880
Speaker 1: no one is thought to put together, or inventions that

8
00:00:28,000 --> 00:00:32,599
Speaker 1: evolve beyond the grasp of their creators. And what coding

9
00:00:32,680 --> 00:00:38,840
Speaker 1: looks like in the near future. Welcome to Inner Cosmos

10
00:00:38,840 --> 00:00:41,560
Speaker 1: with Me and David Eagleman. I'm a neuroscientist and author

11
00:00:41,600 --> 00:00:44,680
Speaker 1: at Stanford and in these episodes we look at the

12
00:00:44,720 --> 00:00:48,839
Speaker 1: world inside us and around us to understand why and

13
00:00:48,920 --> 00:01:10,600
Speaker 1: how our lives look the way they do. Today's episode

14
00:01:10,640 --> 00:01:15,520
Speaker 1: is about computer code. We live in a world increasingly

15
00:01:15,560 --> 00:01:19,080
Speaker 1: built out of symbols. We've got strings of code and

16
00:01:19,120 --> 00:01:24,000
Speaker 1: lines of logic and invisible layers of computation stacked very

17
00:01:24,120 --> 00:01:26,959
Speaker 1: deeply in our lives. This goes from traffic lights to

18
00:01:27,080 --> 00:01:32,520
Speaker 1: financial markets, to weather predictions, to streaming recommendations, to all

19
00:01:32,560 --> 00:01:34,679
Speaker 1: of our apps and our AI and on and on.

20
00:01:35,080 --> 00:01:40,240
Speaker 1: We are surrounded by systems that hum quietly in the background,

21
00:01:40,520 --> 00:01:44,760
Speaker 1: and that orchestrate everything about our modern lives. But it's

22
00:01:44,880 --> 00:01:50,160
Speaker 1: very uncommonly that we stop to ask, what is code? Really?

23
00:01:50,520 --> 00:01:54,200
Speaker 1: It's a tool, but it's also a massive force that

24
00:01:54,280 --> 00:01:57,800
Speaker 1: has taken over the world. So where does it come from?

25
00:01:58,160 --> 00:02:00,920
Speaker 1: Where is it taking us? And what does it mean

26
00:02:01,240 --> 00:02:05,800
Speaker 1: to live inside systems that we ourselves have written but

27
00:02:05,880 --> 00:02:10,720
Speaker 1: we can't possibly fully understand. So today's guest invites us

28
00:02:10,720 --> 00:02:15,079
Speaker 1: to see code as having an aspect of magic. We're

29
00:02:15,080 --> 00:02:18,720
Speaker 1: going to talk with Samuel Arbisman. He's a complexity scientist

30
00:02:18,760 --> 00:02:23,400
Speaker 1: and a writer who thinks about the evolving relationship between

31
00:02:23,800 --> 00:02:26,840
Speaker 1: humans and the tools that we build, and how our

32
00:02:26,880 --> 00:02:30,640
Speaker 1: creations can outpace our comprehension. He's just written a new

33
00:02:30,680 --> 00:02:34,119
Speaker 1: book called The Magic of Code, and here he takes

34
00:02:34,160 --> 00:02:38,359
Speaker 1: a dive into the joy and the deeper nature of programming,

35
00:02:38,639 --> 00:02:42,520
Speaker 1: in other words, beyond just an engineering discipline, but instead

36
00:02:42,800 --> 00:02:47,160
Speaker 1: as a hopeful and enchanted practice. I think I would

37
00:02:47,200 --> 00:02:51,400
Speaker 1: describe this book as a blurring of the boundaries between

38
00:02:51,560 --> 00:02:56,800
Speaker 1: science and art, between logic and myth, between our intentions

39
00:02:56,840 --> 00:03:00,600
Speaker 1: with writing code and what can actually emerge. Because the

40
00:03:00,720 --> 00:03:05,560
Speaker 1: paradox is that code is built from rigid languages with

41
00:03:05,720 --> 00:03:09,800
Speaker 1: strict syntax and rules, but it lets us create whole

42
00:03:09,960 --> 00:03:16,280
Speaker 1: worlds from scratch. We can simulate galaxies and evolve virtual creatures,

43
00:03:16,320 --> 00:03:20,919
Speaker 1: and test new economies and reimagine cities, and it gives

44
00:03:21,040 --> 00:03:24,560
Speaker 1: rise to things that are complex and unpredictable and often

45
00:03:24,880 --> 00:03:34,040
Speaker 1: quite beautiful. So here's my interview with Sam Arbusman. So, Sam,

46
00:03:34,080 --> 00:03:37,800
Speaker 1: you're a scientist with very broad interests. What drew you

47
00:03:38,400 --> 00:03:41,840
Speaker 1: to writing about your latest book on the subject of code?

48
00:03:42,400 --> 00:03:45,720
Speaker 2: Yeah, So, one of the things I think about when

49
00:03:45,720 --> 00:03:48,360
Speaker 2: I think about how people are talking about technology and

50
00:03:48,400 --> 00:03:51,880
Speaker 2: computing and code is that right now, it feels like

51
00:03:51,880 --> 00:03:53,880
Speaker 2: there's almost this kind of like a broken conversation in

52
00:03:53,960 --> 00:03:57,520
Speaker 2: society where when we talk about code or computing the

53
00:03:57,520 --> 00:04:01,800
Speaker 2: world of tech, there is this average real stance towards it,

54
00:04:02,040 --> 00:04:04,560
Speaker 2: or or just worried about it. Sometimes people are just

55
00:04:04,880 --> 00:04:07,000
Speaker 2: ignorant about it and unwilling to kind of learn more.

56
00:04:07,400 --> 00:04:10,280
Speaker 2: And certainly some of the adversarial stuff is reasonable. But

57
00:04:11,160 --> 00:04:13,640
Speaker 2: for me, like when I think about my own experience

58
00:04:13,680 --> 00:04:17,040
Speaker 2: with computers growing up, it wasn't adversarial. It was kind

59
00:04:17,040 --> 00:04:19,360
Speaker 2: of full of this kind of like wonder and delight.

60
00:04:19,800 --> 00:04:22,000
Speaker 2: It also didn't feel that like computing was really just

61
00:04:22,000 --> 00:04:26,360
Speaker 2: this branch of engineering. It also really connected to lots

62
00:04:26,400 --> 00:04:28,560
Speaker 2: of different things. It was almost like humanistic liberal arts

63
00:04:28,760 --> 00:04:32,240
Speaker 2: that drew in language and philosophy and biology and art

64
00:04:32,360 --> 00:04:34,839
Speaker 2: and how we think in all these different areas. And

65
00:04:34,920 --> 00:04:38,560
Speaker 2: so for me, I wanted to try to explain how

66
00:04:38,600 --> 00:04:41,640
Speaker 2: to think about code and computing as this almost like

67
00:04:41,920 --> 00:04:44,039
Speaker 2: liberal art that attracts all these different topics.

68
00:04:44,080 --> 00:04:47,240
Speaker 1: And in the book, you compare this to philology, Right,

69
00:04:47,360 --> 00:04:50,000
Speaker 1: what's this you tell us about philology?

70
00:04:50,120 --> 00:04:54,160
Speaker 2: Yeah, so, so philology it's this branch of humanistic study

71
00:04:54,480 --> 00:04:59,600
Speaker 2: that within the humanities, it was devoted to understanding the

72
00:04:59,600 --> 00:05:02,279
Speaker 2: origin of words and kind of the nature of the

73
00:05:02,360 --> 00:05:06,040
Speaker 2: history of language. But philology, in the process of doing philology,

74
00:05:06,480 --> 00:05:11,080
Speaker 2: it required knowing about archaeology and anthropology and history and

75
00:05:11,120 --> 00:05:15,960
Speaker 2: all these different topics. And then eventually philology sort of fractured,

76
00:05:16,000 --> 00:05:17,240
Speaker 2: and that's kind of where we got a lot of

77
00:05:17,240 --> 00:05:20,200
Speaker 2: the different domains within the humanities. And I kind of

78
00:05:20,279 --> 00:05:23,599
Speaker 2: think that in some ways computing has at least some

79
00:05:23,960 --> 00:05:28,040
Speaker 2: aspect of that kind of philology as unifier of lots

80
00:05:28,040 --> 00:05:31,240
Speaker 2: of different topics. And so for me, my goal was

81
00:05:31,279 --> 00:05:35,120
Speaker 2: to kind of try to show how computing and code

82
00:05:35,400 --> 00:05:39,920
Speaker 2: can actually have that connective tissue between all these different domains.

83
00:05:40,120 --> 00:05:43,360
Speaker 1: And so you described code as being magical.

84
00:05:43,680 --> 00:05:46,320
Speaker 2: Why and so when I say magical, I'm not saying

85
00:05:47,160 --> 00:05:48,760
Speaker 2: that it's like magic in the sense of like, oh,

86
00:05:48,800 --> 00:05:51,239
Speaker 2: like this piece of software just works, it works like magic,

87
00:05:51,520 --> 00:05:53,640
Speaker 2: although there is some of that. For me, it's actually

88
00:05:53,680 --> 00:05:56,640
Speaker 2: this idea that when we think about the nature of

89
00:05:56,720 --> 00:06:01,159
Speaker 2: code or magic, and we've had as a society this

90
00:06:01,279 --> 00:06:04,000
Speaker 2: desire for millennia to kind of coerce the world around

91
00:06:04,080 --> 00:06:06,320
Speaker 2: us through our language and our texts and our speech

92
00:06:06,600 --> 00:06:09,240
Speaker 2: and make the world kind of do our bidding. And

93
00:06:09,279 --> 00:06:11,679
Speaker 2: only in the past, I don't know, seventy five odd

94
00:06:11,760 --> 00:06:14,680
Speaker 2: years since the advent of the modern digital computer, has

95
00:06:14,720 --> 00:06:16,680
Speaker 2: this been a reality where we can actually write text,

96
00:06:16,720 --> 00:06:18,880
Speaker 2: we can write code, and it can actually do things

97
00:06:18,880 --> 00:06:21,839
Speaker 2: in the world. And so for me, there is this

98
00:06:22,000 --> 00:06:26,599
Speaker 2: deep similarity between how we've thought about magic in the

99
00:06:26,600 --> 00:06:28,599
Speaker 2: ancient or medieval days, or even in the stories that

100
00:06:28,640 --> 00:06:31,839
Speaker 2: we tell ourselves and the reality of code. And so,

101
00:06:32,800 --> 00:06:35,760
Speaker 2: of course this analogy and metaphor can only be taken

102
00:06:35,800 --> 00:06:38,360
Speaker 2: so far before it kind of breaks down. I certainly

103
00:06:38,400 --> 00:06:41,200
Speaker 2: take it to the bending point, if not the breaking point.

104
00:06:41,400 --> 00:06:44,240
Speaker 2: But there are a lot of deep similarities between how

105
00:06:44,240 --> 00:06:47,720
Speaker 2: to think about this. So, for example, magic often requires

106
00:06:47,760 --> 00:06:50,320
Speaker 2: in our stories a certain amount of training and knowledge.

107
00:06:50,400 --> 00:06:52,600
Speaker 2: And it's a craft. It's not just like a thing

108
00:06:52,640 --> 00:06:55,039
Speaker 2: like that just works. It actually requires you to learn

109
00:06:55,080 --> 00:06:58,039
Speaker 2: certain things. And so we have Hogwarts School of Witchcraft

110
00:06:58,120 --> 00:07:00,000
Speaker 2: and Wizardry. You got to go there for seven years

111
00:07:00,080 --> 00:07:02,280
Speaker 2: or whatever it is. And so too with code, like

112
00:07:02,320 --> 00:07:05,320
Speaker 2: it doesn't necessarily just work. You actually have to understand

113
00:07:05,600 --> 00:07:07,919
Speaker 2: how like the nature of syntax and the details of

114
00:07:07,920 --> 00:07:10,560
Speaker 2: code and so and so that example, as well as

115
00:07:10,560 --> 00:07:12,720
Speaker 2: other ones, I kind of us to show the ways

116
00:07:12,840 --> 00:07:15,280
Speaker 2: in which this this idea of like the kind of

117
00:07:15,280 --> 00:07:17,920
Speaker 2: the analogy of magic actually can be a productive and

118
00:07:18,000 --> 00:07:20,400
Speaker 2: useful one to help us better understand how code works.

119
00:07:20,600 --> 00:07:22,760
Speaker 1: So it's not just a set of instructions. It's like

120
00:07:22,920 --> 00:07:25,160
Speaker 1: a spell in the sense that you puts a strange

121
00:07:25,160 --> 00:07:27,560
Speaker 1: set of symbols and it does stuff in the world

122
00:07:27,640 --> 00:07:31,240
Speaker 1: that moves electrons or launches rockets, or models pandemics or

123
00:07:31,280 --> 00:07:35,600
Speaker 1: creates simulations. And that's the sense in which it's got

124
00:07:35,600 --> 00:07:39,480
Speaker 1: this magic to it. But also you point to the

125
00:07:39,520 --> 00:07:44,400
Speaker 1: fact that there's often unpredictability and emergence of things we

126
00:07:44,440 --> 00:07:47,280
Speaker 1: didn't expect from code. So can you give us an

127
00:07:47,280 --> 00:07:50,240
Speaker 1: example of this dual nature of code?

128
00:07:50,880 --> 00:07:53,000
Speaker 2: Yeah, I mean so certainly. In the world of magic,

129
00:07:53,040 --> 00:07:55,240
Speaker 2: we have a lot of these stories, and like there's

130
00:07:55,240 --> 00:07:56,840
Speaker 2: like the story of like if the source is Apprentice,

131
00:07:56,840 --> 00:07:58,560
Speaker 2: which I think there's like the old version and then

132
00:07:58,560 --> 00:08:02,160
Speaker 2: the Disney makeme Out version, where some sort of magic

133
00:08:02,200 --> 00:08:05,960
Speaker 2: has unanticipated consequences ands only you have you have brooms

134
00:08:06,040 --> 00:08:08,880
Speaker 2: kind of walking around and flooding and flooding a basement,

135
00:08:09,040 --> 00:08:11,480
Speaker 2: And the same kind of thing is true with code.

136
00:08:11,520 --> 00:08:14,120
Speaker 2: Where in code, I think people who might not be

137
00:08:14,200 --> 00:08:16,480
Speaker 2: familiar with with programming think of it as, oh, like

138
00:08:16,480 --> 00:08:18,000
Speaker 2: I have this idea in my mind and I'm going

139
00:08:18,040 --> 00:08:20,200
Speaker 2: to instantiate it into a computer program. And there is that,

140
00:08:20,480 --> 00:08:24,040
Speaker 2: but there's also a huge amount of debugging and frustration

141
00:08:24,160 --> 00:08:26,840
Speaker 2: because oftentimes when you write a program, there's a gap

142
00:08:26,880 --> 00:08:28,680
Speaker 2: between how you think it will actually work and how

143
00:08:28,720 --> 00:08:32,440
Speaker 2: it actually does work. And oftentimes the reality of the

144
00:08:32,440 --> 00:08:36,360
Speaker 2: program and better understanding of it is only revealed through

145
00:08:36,480 --> 00:08:38,640
Speaker 2: these bugs, through these glitches and edge cases and things

146
00:08:38,679 --> 00:08:42,559
Speaker 2: like that, And so there are many situations where we

147
00:08:42,720 --> 00:08:45,760
Speaker 2: only see these bizarre errors, and then based on those

148
00:08:45,880 --> 00:08:48,880
Speaker 2: errors and these kind of unanticipated consequences, do we realize, oh,

149
00:08:48,920 --> 00:08:50,480
Speaker 2: how this thing actually works, and so it can be

150
00:08:50,720 --> 00:08:53,600
Speaker 2: So there's a there's a well known story of someone

151
00:08:53,640 --> 00:08:55,360
Speaker 2: who I think was like a systems administrator for some

152
00:08:55,440 --> 00:08:58,520
Speaker 2: university department. He was told by the chair of the

153
00:08:58,520 --> 00:09:01,640
Speaker 2: department that their email was only able to be sent

154
00:09:01,679 --> 00:09:04,840
Speaker 2: about five hundred miles away, and the systems like this

155
00:09:04,880 --> 00:09:06,920
Speaker 2: is insane, like email, that's not how email works. And

156
00:09:06,960 --> 00:09:09,200
Speaker 2: it turned out by delving into it, and he was

157
00:09:09,200 --> 00:09:11,480
Speaker 2: able to find that I think that there was like

158
00:09:11,520 --> 00:09:14,920
Speaker 2: an older piece of software that hadn't been upgraded, but

159
00:09:15,440 --> 00:09:18,160
Speaker 2: the newer system didn't realize this and like would time out,

160
00:09:18,160 --> 00:09:20,160
Speaker 2: but only after like some very small amount of time.

161
00:09:20,280 --> 00:09:22,440
Speaker 2: And it turns out, based on like the speed of

162
00:09:22,480 --> 00:09:24,560
Speaker 2: sound and that small amount of time, it ended up

163
00:09:24,559 --> 00:09:26,360
Speaker 2: working out to about five hundred miles and it was

164
00:09:26,400 --> 00:09:29,120
Speaker 2: this weird unanticipated consequence. But of course, and there's also

165
00:09:29,200 --> 00:09:32,079
Speaker 2: other things that just the fact that when you stitch

166
00:09:32,480 --> 00:09:35,439
Speaker 2: systems together and pieces of software together, they all interact

167
00:09:35,440 --> 00:09:38,120
Speaker 2: in unexpected ways. And that's also the kind of unanticipated

168
00:09:38,120 --> 00:09:41,760
Speaker 2: consequences we see. Whether it's like some weird little system

169
00:09:41,760 --> 00:09:44,880
Speaker 2: fails to get upgraded and then suddenly all like the

170
00:09:44,880 --> 00:09:47,760
Speaker 2: the airline systems go down for a first certain amount

171
00:09:47,760 --> 00:09:49,120
Speaker 2: of time or whatever it is. And so there is

172
00:09:49,160 --> 00:09:52,319
Speaker 2: that kind of unanticipated consequence in lots of different ways.

173
00:09:52,320 --> 00:09:54,679
Speaker 2: And we're seeing this, of course even more so with AI.

174
00:09:56,440 --> 00:09:57,960
Speaker 1: Okay, So this is the thing that you and I

175
00:09:58,040 --> 00:10:01,080
Speaker 1: both love is the emergence of comp complexity in the world.

176
00:10:01,320 --> 00:10:04,880
Speaker 1: And with code it's a very specific, detailed set of instructions,

177
00:10:04,920 --> 00:10:08,560
Speaker 1: and yet it can break, it can decay, it can

178
00:10:08,600 --> 00:10:12,160
Speaker 1: be opaque. All kinds of things can happen. And sometimes

179
00:10:12,200 --> 00:10:16,000
Speaker 1: when we try to model complexity, we actually unleash complexity.

180
00:10:16,360 --> 00:10:19,440
Speaker 1: So tell us your take on how code can do

181
00:10:19,520 --> 00:10:21,000
Speaker 1: things that we didn't expect it to do.

182
00:10:21,400 --> 00:10:23,760
Speaker 2: Yeah, and so when we think about and just engineered

183
00:10:23,800 --> 00:10:26,000
Speaker 2: systems more broadly, and certainly when it comes to code,

184
00:10:26,360 --> 00:10:30,000
Speaker 2: you think, oh, it's designed by people. It sounds very logical,

185
00:10:30,040 --> 00:10:32,440
Speaker 2: it's kind of derived from mathematics. It should be very

186
00:10:32,480 --> 00:10:36,440
Speaker 2: simple and straightforward, and very small bits of code are that.

187
00:10:36,920 --> 00:10:39,800
Speaker 2: But the truth is it adds up and then through

188
00:10:39,840 --> 00:10:44,520
Speaker 2: this combination of the sizes of programs, growing them beingcoming

189
00:10:44,600 --> 00:10:47,640
Speaker 2: connected to various other bits of code that are out there,

190
00:10:47,880 --> 00:10:49,640
Speaker 2: as well as also just engaging with kind of the

191
00:10:50,120 --> 00:10:52,720
Speaker 2: messiness of the world around us, you end up getting

192
00:10:52,840 --> 00:10:54,920
Speaker 2: sort of a certain amount of unexpectedness as well as

193
00:10:55,000 --> 00:10:57,840
Speaker 2: kind of just a reduced understanding. And part of this

194
00:10:57,960 --> 00:11:00,200
Speaker 2: is because there's and you mentioned this kind of like

195
00:11:00,440 --> 00:11:02,839
Speaker 2: it's like breaking and things kind of growing over time.

196
00:11:02,960 --> 00:11:06,440
Speaker 2: There is this whole phenomenon of legacy code where code

197
00:11:06,640 --> 00:11:08,760
Speaker 2: has been around for a very very long time, and

198
00:11:09,360 --> 00:11:11,400
Speaker 2: we have systems that are still being you like, that

199
00:11:11,440 --> 00:11:14,200
Speaker 2: are still being used that were developed decades ago, but

200
00:11:14,240 --> 00:11:16,200
Speaker 2: they might be involved in kind of the irs, but

201
00:11:16,200 --> 00:11:18,280
Speaker 2: they were developed first developed or in the Kennedy administration.

202
00:11:18,320 --> 00:11:22,240
Speaker 2: Like there's all these kind of crazy examples where things

203
00:11:22,400 --> 00:11:25,440
Speaker 2: that have been developed the people who first made them

204
00:11:25,520 --> 00:11:27,600
Speaker 2: they might be and they might be long retired, they

205
00:11:27,640 --> 00:11:30,520
Speaker 2: might be dead. And we also just don't fully understand

206
00:11:30,520 --> 00:11:34,040
Speaker 2: these systems. And so it's this weird situation where we

207
00:11:34,160 --> 00:11:36,680
Speaker 2: have to recognize that even the systems of our own

208
00:11:36,840 --> 00:11:41,080
Speaker 2: construction are actually like when they become big enough, they

209
00:11:41,080 --> 00:11:43,480
Speaker 2: actually have this kind of qualitative difference where they almost

210
00:11:43,640 --> 00:11:47,480
Speaker 2: they almost become almost biological or organic in their complexity,

211
00:11:47,559 --> 00:11:50,320
Speaker 2: and as a result, we have a reduced understanding and

212
00:11:50,320 --> 00:11:53,240
Speaker 2: we have to kind of take almost biological modes of

213
00:11:53,320 --> 00:11:56,080
Speaker 2: studying these systems, whether it's kind of like the days

214
00:11:56,080 --> 00:11:57,680
Speaker 2: of old with like the natural is kind of going

215
00:11:57,679 --> 00:12:00,480
Speaker 2: out and collecting bugs in this case could be bugs

216
00:12:00,480 --> 00:12:03,040
Speaker 2: and errors. It could be bugs like insects, as well

217
00:12:03,040 --> 00:12:05,000
Speaker 2: as just kind of like trying to tinker at the

218
00:12:05,040 --> 00:12:08,320
Speaker 2: edges and better understand a system, because the thing overall

219
00:12:08,400 --> 00:12:11,640
Speaker 2: you fully don't understand. So there's this weird situation where

220
00:12:12,000 --> 00:12:16,880
Speaker 2: these systems are engineered, but they also involve kind of

221
00:12:17,440 --> 00:12:20,120
Speaker 2: a certain amount of humility in trying to understand these systems.

222
00:12:20,120 --> 00:12:22,000
Speaker 2: And part of that is also because one of the

223
00:12:22,000 --> 00:12:25,240
Speaker 2: other features of computing and software is this idea of

224
00:12:25,240 --> 00:12:27,800
Speaker 2: abstraction that you can kind of build things on top

225
00:12:27,800 --> 00:12:31,800
Speaker 2: of other pieces, and those pieces are then sophisticated kind

226
00:12:31,800 --> 00:12:33,640
Speaker 2: of units, and you can kind of use them as

227
00:12:33,720 --> 00:12:35,800
Speaker 2: standalone bits and then don't have to worry about the

228
00:12:35,840 --> 00:12:39,280
Speaker 2: things underneath it. And so that modularity is very, very powerful,

229
00:12:39,320 --> 00:12:42,520
Speaker 2: but as a result, there is a decreased amount of understanding,

230
00:12:42,559 --> 00:12:45,480
Speaker 2: and so sometimes not understanding what's going on under the

231
00:12:45,480 --> 00:12:48,359
Speaker 2: hood or kind of underneath these pieces, even when yourself

232
00:12:48,600 --> 00:12:51,040
Speaker 2: are programming them or programming kind of the things that

233
00:12:51,080 --> 00:12:53,440
Speaker 2: interact with them, means that you can kind of also

234
00:12:53,520 --> 00:13:10,760
Speaker 2: have a certain amount of anticipated consequences.

235
00:13:12,840 --> 00:13:15,040
Speaker 1: And so what are the key limitations that you see

236
00:13:15,040 --> 00:13:18,239
Speaker 1: when we try to model very complex systems?

237
00:13:18,640 --> 00:13:19,880
Speaker 2: So one of the things I think about when I

238
00:13:19,880 --> 00:13:23,720
Speaker 2: think about modeling complex systems is is what is the

239
00:13:23,720 --> 00:13:27,360
Speaker 2: goal of modeling the complex system? So there's some situations

240
00:13:27,400 --> 00:13:31,000
Speaker 2: where we really want to have perfect fidelity to a

241
00:13:31,040 --> 00:13:33,720
Speaker 2: real world system. And so for predictions, like with weather,

242
00:13:34,160 --> 00:13:35,880
Speaker 2: as a weather prediction, like you want to and you

243
00:13:35,920 --> 00:13:38,760
Speaker 2: want to not just understand kind of how kind of

244
00:13:38,840 --> 00:13:41,800
Speaker 2: air moves around, You want to really understand whether or

245
00:13:41,840 --> 00:13:43,400
Speaker 2: not it's going to rain tomorrow or in an hour

246
00:13:43,480 --> 00:13:46,120
Speaker 2: or two hours or whatever it is. And and based

247
00:13:46,120 --> 00:13:48,360
Speaker 2: on that you have to have a great deal of

248
00:13:48,440 --> 00:13:51,280
Speaker 2: data and a great deal of complexity. And then oftentimes

249
00:13:51,480 --> 00:13:55,360
Speaker 2: the resulting models might be very powerful, very sophisticated, but

250
00:13:55,400 --> 00:13:57,600
Speaker 2: there might be a reduced amount of understanding and actually

251
00:13:57,720 --> 00:13:59,800
Speaker 2: how these things are doing what they're doing. On the

252
00:13:59,800 --> 00:14:03,360
Speaker 2: other hand, if you want to just understand the features

253
00:14:03,400 --> 00:14:06,400
Speaker 2: of a system, you can sometimes get away with a

254
00:14:06,520 --> 00:14:09,880
Speaker 2: much simpler model, which might not necessarily be exactly the

255
00:14:09,920 --> 00:14:12,320
Speaker 2: way that the model works, but could at least capture

256
00:14:12,360 --> 00:14:14,040
Speaker 2: some of the complexity and kind of the and the

257
00:14:14,080 --> 00:14:16,520
Speaker 2: emergence of what you were talking about of that system.

258
00:14:16,559 --> 00:14:18,960
Speaker 2: And so, for example, and this is a kind of

259
00:14:18,960 --> 00:14:22,080
Speaker 2: trivial example, but the computer game SimCity. It is not

260
00:14:22,720 --> 00:14:26,440
Speaker 2: modeling an actual city, but to give you an intuitive

261
00:14:26,520 --> 00:14:32,000
Speaker 2: sense of how feedback operates, or unanticipated consequences work, or

262
00:14:32,080 --> 00:14:34,480
Speaker 2: just the fact that, like complex systems can bite back

263
00:14:34,600 --> 00:14:37,800
Speaker 2: and do weird things that you might not expect. SimCity

264
00:14:37,880 --> 00:14:40,080
Speaker 2: is great for that kind of thing. And so, and

265
00:14:40,120 --> 00:14:41,640
Speaker 2: it can also kind of give you a sense of, oh,

266
00:14:41,640 --> 00:14:43,240
Speaker 2: when I do this kind of thing, according to this

267
00:14:43,320 --> 00:14:46,120
Speaker 2: model of how Will Wright or whoever was programming it

268
00:14:46,280 --> 00:14:48,520
Speaker 2: thought cities would work, this maybe would be the way

269
00:14:48,520 --> 00:14:50,040
Speaker 2: it works. Whether or not that is actually how the

270
00:14:50,040 --> 00:14:52,680
Speaker 2: city operates, that's entirely different, different thing. But so for me,

271
00:14:53,080 --> 00:14:55,680
Speaker 2: I often think about, like, yeah, what is the ultimate

272
00:14:55,720 --> 00:14:57,520
Speaker 2: goal with the model? Is the goal to kind of

273
00:14:57,600 --> 00:15:00,960
Speaker 2: understand things? And we have to recognizing our human minds

274
00:15:00,960 --> 00:15:04,400
Speaker 2: are really limited when it comes to understanding complex and

275
00:15:04,440 --> 00:15:07,680
Speaker 2: nonlinear systems, and so we need these simplified models. If

276
00:15:07,720 --> 00:15:11,320
Speaker 2: it is actually just prediction, then sometimes a really complex

277
00:15:11,360 --> 00:15:15,080
Speaker 2: model can work, but at the cost of reduced understanding.

278
00:15:15,440 --> 00:15:19,640
Speaker 1: So as AI generated code becomes more common, what does

279
00:15:19,640 --> 00:15:24,120
Speaker 1: that do to our sense of authorship and even understanding?

280
00:15:24,120 --> 00:15:26,160
Speaker 1: And could we end up in a situation where we're

281
00:15:26,160 --> 00:15:29,320
Speaker 1: surrounded by systems that are running our world that we

282
00:15:29,360 --> 00:15:30,520
Speaker 1: don't understand at all.

283
00:15:30,640 --> 00:15:33,040
Speaker 2: I mean, to be honest, I think we're probably there already.

284
00:15:33,480 --> 00:15:35,800
Speaker 2: It's just the situation where I think many people are

285
00:15:35,800 --> 00:15:38,000
Speaker 2: not aware of that fact. This is also one of

286
00:15:38,040 --> 00:15:40,640
Speaker 2: these situations where as we build more and more complex

287
00:15:40,680 --> 00:15:44,040
Speaker 2: systems and everyday users are kind of more distant from them,

288
00:15:44,080 --> 00:15:46,880
Speaker 2: we just don't realize the sheer complexity. When the Apple

289
00:15:46,920 --> 00:15:50,080
Speaker 2: Watch first came out, this is years ago, there was

290
00:15:50,120 --> 00:15:52,600
Speaker 2: an article in the Wall Street Journal, I think it

291
00:15:52,640 --> 00:15:54,680
Speaker 2: was like the Style section about like, are people going

292
00:15:54,720 --> 00:15:56,960
Speaker 2: to still use like biomechanical watch? As the answer is

293
00:15:57,040 --> 00:15:59,200
Speaker 2: they still are that. They interviewed this one guy about it,

294
00:15:59,240 --> 00:16:01,360
Speaker 2: like whether or not you want to buy mechanical watch

295
00:16:01,400 --> 00:16:03,720
Speaker 2: or just smart watches, and this guy said something to

296
00:16:03,720 --> 00:16:06,400
Speaker 2: the effect of, of course I want a mechanical watch.

297
00:16:06,400 --> 00:16:08,160
Speaker 2: When I think about a mechanical watch. It's so complex

298
00:16:08,160 --> 00:16:10,360
Speaker 2: as opposed to a smart watch, which is just a chip,

299
00:16:10,520 --> 00:16:12,440
Speaker 2: and the thing is like a chip is orders of

300
00:16:12,520 --> 00:16:15,480
Speaker 2: magnitude more complex than a mechanical watch. But we've been

301
00:16:15,480 --> 00:16:18,560
Speaker 2: shielded from it, and I think as we have AI

302
00:16:18,640 --> 00:16:21,880
Speaker 2: generated code, we're going to kind of have another level

303
00:16:21,920 --> 00:16:26,160
Speaker 2: of shielding. I do think we need better mechanisms for

304
00:16:26,680 --> 00:16:28,720
Speaker 2: so interrogating the system. So I actually so I think

305
00:16:28,760 --> 00:16:30,280
Speaker 2: one of these situations we're On the one hand, it

306
00:16:30,360 --> 00:16:32,560
Speaker 2: is very good that we can now generate code via

307
00:16:32,640 --> 00:16:36,760
Speaker 2: AI and build simple tools and actually democratize the software development.

308
00:16:36,760 --> 00:16:39,200
Speaker 2: I think there's lots of interesting things there. But I

309
00:16:39,560 --> 00:16:42,120
Speaker 2: still also think understanding code to a certain degree and

310
00:16:42,160 --> 00:16:44,360
Speaker 2: allowing you to kind of like dive into the code

311
00:16:44,360 --> 00:16:46,920
Speaker 2: that is being generated and tweak it, not only does

312
00:16:46,920 --> 00:16:48,760
Speaker 2: it give you a better understanding of what you're doing,

313
00:16:48,880 --> 00:16:50,920
Speaker 2: but it's still actually really good to help make sure

314
00:16:50,960 --> 00:16:53,720
Speaker 2: that it is doing at least partly what you hope for.

315
00:16:54,440 --> 00:16:57,200
Speaker 2: That being said, our systems have always been imperfect, and

316
00:16:57,240 --> 00:16:59,000
Speaker 2: I think right now, this is this moment is kind

317
00:16:59,000 --> 00:17:02,120
Speaker 2: of just heightening that fact, and maybe we'll give people

318
00:17:02,200 --> 00:17:07,800
Speaker 2: a better awareness that these systems have always been enormously complex,

319
00:17:08,520 --> 00:17:11,440
Speaker 2: enormously imperfect, made by humans at least at some level,

320
00:17:12,560 --> 00:17:15,760
Speaker 2: but maybe give us a greater appreciation for building systems

321
00:17:15,800 --> 00:17:18,680
Speaker 2: on top of these, maybe also AI generated as well,

322
00:17:18,960 --> 00:17:23,040
Speaker 2: that can allow us to make sure that the the

323
00:17:23,080 --> 00:17:25,560
Speaker 2: unanticipated consequences are as minimal as possible.

324
00:17:27,359 --> 00:17:28,720
Speaker 1: You know, I was just thinking about right after I

325
00:17:28,720 --> 00:17:31,320
Speaker 1: asked my question about could we end up living inside

326
00:17:31,359 --> 00:17:34,520
Speaker 1: systems that are too complex for us to understand. Obviously,

327
00:17:34,600 --> 00:17:37,480
Speaker 1: we live inside our biology, and that is for sure

328
00:17:37,680 --> 00:17:41,120
Speaker 1: that we don't understand. But a fraction of what's going

329
00:17:41,160 --> 00:17:43,480
Speaker 1: on inside is biologically. But we try to eat the

330
00:17:43,560 --> 00:17:46,600
Speaker 1: right foods and get exercise and just just right on

331
00:17:46,680 --> 00:17:49,320
Speaker 1: top of this system. So we're actually quite used to

332
00:17:49,359 --> 00:17:52,920
Speaker 1: living inside systems that are beyond our understanding.

333
00:17:53,200 --> 00:17:55,000
Speaker 2: Yeah, And I think and that goes back to kind

334
00:17:55,000 --> 00:17:59,240
Speaker 2: of like the biological nature of these massive, complex computational systems. Like,

335
00:17:59,280 --> 00:18:02,880
Speaker 2: the more we recognize that these systems are really complex

336
00:18:02,920 --> 00:18:06,480
Speaker 2: and almost have like this organic quality, we will have

337
00:18:06,520 --> 00:18:08,520
Speaker 2: to realize, yeah, that we need different ways of approaching

338
00:18:08,560 --> 00:18:11,040
Speaker 2: them and the way we approach our bodies. Right, It's not,

339
00:18:11,240 --> 00:18:13,720
Speaker 2: oh like I have total ignorance about how the system

340
00:18:13,720 --> 00:18:16,440
Speaker 2: works or have complete understanding. There's there's a lot in between,

341
00:18:16,840 --> 00:18:20,480
Speaker 2: and using rules of thumbs and things are are very

342
00:18:20,520 --> 00:18:22,359
Speaker 2: powerful and I but at the same time, though, we

343
00:18:22,359 --> 00:18:24,439
Speaker 2: don't want to necessarily kind of like succumb to like

344
00:18:24,480 --> 00:18:26,600
Speaker 2: the like the biohacking trend, which I feel like five

345
00:18:26,640 --> 00:18:29,360
Speaker 2: ten years ago was a really big thing where it's like, oh,

346
00:18:29,440 --> 00:18:31,280
Speaker 2: if I can just find this one chemical to ingest

347
00:18:31,320 --> 00:18:34,440
Speaker 2: or this one cool trick to do, then I'll never

348
00:18:34,480 --> 00:18:37,240
Speaker 2: need to sleep again or I'll be held and we

349
00:18:37,280 --> 00:18:40,399
Speaker 2: realize that, I mean, yeah, our system, our bodies have

350
00:18:40,480 --> 00:18:44,840
Speaker 2: evolved over millions of years and are optimizing a huge

351
00:18:44,920 --> 00:18:46,680
Speaker 2: number of different things, and they're going to be imperfect

352
00:18:46,720 --> 00:18:49,000
Speaker 2: and weird, and so it's going to be maybe we

353
00:18:49,040 --> 00:18:51,160
Speaker 2: can find those things, but the odds that we are

354
00:18:51,400 --> 00:18:53,320
Speaker 2: going to is very low. And I think the same

355
00:18:53,400 --> 00:18:56,200
Speaker 2: kind of thing that that same kind of approach needs

356
00:18:56,240 --> 00:18:59,240
Speaker 2: to be used when we think about these complex technologies

357
00:18:59,280 --> 00:19:01,000
Speaker 2: that were building around us as well.

358
00:19:01,160 --> 00:19:03,200
Speaker 1: So let me ask you this, if we fast forward

359
00:19:03,280 --> 00:19:07,280
Speaker 1: one hundred years, are we still coding by using symbols

360
00:19:07,280 --> 00:19:09,919
Speaker 1: and syntax or is it a completely different sort of

361
00:19:10,040 --> 00:19:15,320
Speaker 1: thing where we are setting initial conditions and letting complexity evolve.

362
00:19:15,320 --> 00:19:17,480
Speaker 2: I would say it's very different. I don't think we

363
00:19:17,520 --> 00:19:19,639
Speaker 2: need to necessarily go even one hundred years into the future.

364
00:19:19,680 --> 00:19:22,159
Speaker 2: It could be like five ten years until we're kind

365
00:19:22,160 --> 00:19:25,000
Speaker 2: of managing these AI generated systems. I mean. But the

366
00:19:25,040 --> 00:19:28,840
Speaker 2: truth is, when I think about what coding is, it's

367
00:19:28,880 --> 00:19:31,640
Speaker 2: always been this moving target, like it's always been changing.

368
00:19:31,960 --> 00:19:34,439
Speaker 2: So the way I learned how to program in some

369
00:19:34,480 --> 00:19:37,760
Speaker 2: of the languages I learned, they're not totally extinct, but

370
00:19:37,760 --> 00:19:40,760
Speaker 2: they're not things I would ever consider using nowadays. But

371
00:19:40,800 --> 00:19:43,760
Speaker 2: also even when I think about how people before me

372
00:19:43,840 --> 00:19:46,959
Speaker 2: learned how to program, that too was something very very different.

373
00:19:47,040 --> 00:19:50,199
Speaker 2: It was like plugging in cables or flipping switches or

374
00:19:50,359 --> 00:19:53,760
Speaker 2: writing things in binary or assembly code. Like I never

375
00:19:53,800 --> 00:19:55,960
Speaker 2: did those kinds of things. Nor do I really have

376
00:19:56,040 --> 00:19:57,840
Speaker 2: a strong desire too. And that's okay, And I think

377
00:19:57,840 --> 00:19:59,600
Speaker 2: it's going to continue changing. And so one of the

378
00:19:59,600 --> 00:20:02,280
Speaker 2: ways I think about this is actually I tell this

379
00:20:02,359 --> 00:20:05,800
Speaker 2: story in the book, a story from the Tumud actually

380
00:20:05,920 --> 00:20:11,000
Speaker 2: where it's this conversation that's describing a conversation between God

381
00:20:11,040 --> 00:20:14,240
Speaker 2: and Moses and they're discussing like, oh, like who's going

382
00:20:14,280 --> 00:20:16,919
Speaker 2: to be the greatest scholar in the future, and God says, oh,

383
00:20:16,920 --> 00:20:18,960
Speaker 2: it's going to be some rabbi, like a thousand and

384
00:20:18,960 --> 00:20:20,879
Speaker 2: two thousand years in the future, and Moses says, can

385
00:20:20,920 --> 00:20:22,600
Speaker 2: you show him to me. So there's this weird time

386
00:20:22,640 --> 00:20:24,720
Speaker 2: travel moment where he's transported to the hall of Study,

387
00:20:24,720 --> 00:20:26,280
Speaker 2: like a thousand years in the future whatever it is,

388
00:20:26,600 --> 00:20:28,560
Speaker 2: and Moses is sitting in the back listening to this

389
00:20:28,600 --> 00:20:32,680
Speaker 2: illustrious scholar talking about things, and he realizes he doesn't

390
00:20:32,760 --> 00:20:35,560
Speaker 2: understand anything this guy's talking about, and he's kind of overwhelmed.

391
00:20:35,560 --> 00:20:37,920
Speaker 2: He's like, Oh, I'm the one who received the law

392
00:20:38,400 --> 00:20:40,480
Speaker 2: from Heaven and I don't get it until at the

393
00:20:40,560 --> 00:20:42,640
Speaker 2: very last moment, the rabbi says, oh, in the way

394
00:20:42,680 --> 00:20:46,320
Speaker 2: we understand all this is because of the law received

395
00:20:46,359 --> 00:20:48,479
Speaker 2: from Moses at Sinai, And at that point he's calmed

396
00:20:48,600 --> 00:20:51,239
Speaker 2: and because he realizes that even if he doesn't understand it,

397
00:20:51,480 --> 00:20:55,080
Speaker 2: there's this clear, continuous line and kind of continuous tradition.

398
00:20:55,119 --> 00:20:57,359
Speaker 2: And I feel like when it comes to code, the

399
00:20:57,400 --> 00:20:59,520
Speaker 2: same kind of thing is true. Coding has changed, it

400
00:20:59,520 --> 00:21:02,000
Speaker 2: will continue to change. It'll be much more like managing

401
00:21:02,040 --> 00:21:05,080
Speaker 2: AI systems or some other thing. But Ultimately, it's all

402
00:21:05,119 --> 00:21:08,240
Speaker 2: about taking some idea in our heads and finding some

403
00:21:08,320 --> 00:21:11,040
Speaker 2: way of instantiing it into a machine and actually getting

404
00:21:11,080 --> 00:21:12,840
Speaker 2: the machine to do something, to kind of do our bidding.

405
00:21:13,440 --> 00:21:15,679
Speaker 2: And what that looks like is always going to be

406
00:21:15,760 --> 00:21:18,240
Speaker 2: changing and so but as long as we recognize it's

407
00:21:18,280 --> 00:21:21,480
Speaker 2: all part of this long tradition, I think, then I

408
00:21:21,720 --> 00:21:24,040
Speaker 2: kind of do it. It's all coding, whether or not

409
00:21:24,080 --> 00:21:26,919
Speaker 2: it's syntax or Python or Pearl or whatever it is,

410
00:21:27,320 --> 00:21:29,200
Speaker 2: it will definitely not be that. But that's okay.

411
00:21:29,400 --> 00:21:31,080
Speaker 1: First of all, I love that story. I had no

412
00:21:31,080 --> 00:21:33,720
Speaker 1: idea that there was time travel and the toallment. That's amazing.

413
00:21:33,840 --> 00:21:35,320
Speaker 1: I want to come back to this point that we're

414
00:21:35,320 --> 00:21:38,680
Speaker 1: talking about that we're already living inside a system that

415
00:21:38,880 --> 00:21:41,960
Speaker 1: is so complex. We don't understand this because you know,

416
00:21:42,040 --> 00:21:46,000
Speaker 1: we program simulations, we program other things. But increasingly what

417
00:21:46,040 --> 00:21:50,919
Speaker 1: it means is we're really living inside of this opaque simulation.

418
00:21:51,119 --> 00:21:53,680
Speaker 1: And this will be even more true for our descendants,

419
00:21:53,680 --> 00:21:56,879
Speaker 1: where they'll be living inside this world of creation that

420
00:21:56,920 --> 00:21:59,440
Speaker 1: they can't understand explicitly.

421
00:22:00,040 --> 00:22:03,080
Speaker 2: Danny Danny hillis the computer scientist. He has this great

422
00:22:03,160 --> 00:22:05,359
Speaker 2: term where he talks about how we've kind of moved

423
00:22:05,400 --> 00:22:07,560
Speaker 2: from the enlightenment to this kind of age where we

424
00:22:07,560 --> 00:22:10,080
Speaker 2: could take our take our mind and kind of apply

425
00:22:10,119 --> 00:22:11,520
Speaker 2: it to the world around us and really understand it.

426
00:22:11,960 --> 00:22:13,760
Speaker 2: To the entanglement. We've kind of moved to this area

427
00:22:13,800 --> 00:22:16,520
Speaker 2: where everything is so hopelessly interconnected, we're never going to

428
00:22:16,560 --> 00:22:19,119
Speaker 2: fully understand it. And I think we've been in the

429
00:22:19,240 --> 00:22:22,600
Speaker 2: entanglement for quite some time, and it's it's really just

430
00:22:22,640 --> 00:22:25,679
Speaker 2: a matter of becoming a little bit more aware of it.

431
00:22:25,920 --> 00:22:27,879
Speaker 2: I think, I mean going back to kind of the

432
00:22:27,920 --> 00:22:31,560
Speaker 2: analogy of biology and things like that with technology, and

433
00:22:31,600 --> 00:22:34,199
Speaker 2: biology is a form of technology. Someone once told me

434
00:22:34,240 --> 00:22:36,119
Speaker 2: that the way he kind of thought about it is

435
00:22:36,160 --> 00:22:39,440
Speaker 2: like the most complicated engineered system that humans have ever

436
00:22:39,520 --> 00:22:44,800
Speaker 2: made are domesticating dogs, because like these are we made them.

437
00:22:44,920 --> 00:22:48,479
Speaker 2: I we we evolve them basically through like through our

438
00:22:48,680 --> 00:22:52,479
Speaker 2: artificial selection, but they're an enormously complicated system. And I

439
00:22:52,520 --> 00:22:56,560
Speaker 2: think that's those kinds of approaches, whether it's kind of

440
00:22:56,600 --> 00:23:00,480
Speaker 2: like tinkering with systems, kind of evolving them, wrecking these

441
00:23:00,480 --> 00:23:03,280
Speaker 2: things as enormously complicated, those might be the kind of

442
00:23:03,320 --> 00:23:07,240
Speaker 2: approaches that that we need. That being said, I would

443
00:23:07,240 --> 00:23:08,800
Speaker 2: say one of the other things though that at least

444
00:23:08,800 --> 00:23:10,399
Speaker 2: gives me a certain amount of hope. Though that it

445
00:23:10,480 --> 00:23:12,320
Speaker 2: is kind of weird when you think about it, is

446
00:23:12,359 --> 00:23:15,359
Speaker 2: that the extent to which humans are also really good

447
00:23:15,400 --> 00:23:18,000
Speaker 2: at adapting to the world around us. And so we

448
00:23:18,320 --> 00:23:21,360
Speaker 2: think about all the technological changes that have come over

449
00:23:21,400 --> 00:23:26,360
Speaker 2: the past couple hundred years, and these things were enormously destabilizing,

450
00:23:26,880 --> 00:23:28,879
Speaker 2: but in many ways we kind of now take them

451
00:23:28,920 --> 00:23:32,560
Speaker 2: for granted, and even like more modern and more modern ones.

452
00:23:32,600 --> 00:23:35,320
Speaker 2: I'm not just talking about I like air travel or

453
00:23:35,520 --> 00:23:38,000
Speaker 2: certain things around the internet, or the industrial Revolution. So

454
00:23:38,000 --> 00:23:40,000
Speaker 2: my grandfather, he was he lived at the age of

455
00:23:40,040 --> 00:23:42,119
Speaker 2: ninety nine. He was a retired dentist. But he also

456
00:23:42,280 --> 00:23:44,840
Speaker 2: read science fiction since like the modern dawn of the genre,

457
00:23:44,880 --> 00:23:47,879
Speaker 2: Like he read his entire life, and I remember he

458
00:23:47,880 --> 00:23:49,720
Speaker 2: read I think he read Dune when it was sialized

459
00:23:49,760 --> 00:23:52,440
Speaker 2: in a magazine, so like no story could surprise him.

460
00:23:52,600 --> 00:23:54,920
Speaker 2: And I remember when when the iPhone first came out.

461
00:23:55,160 --> 00:23:57,480
Speaker 2: I went with my grandfather as well as my father

462
00:23:57,520 --> 00:23:59,280
Speaker 2: to the Apple store to kind of check out the

463
00:23:59,600 --> 00:24:01,960
Speaker 2: iPhone and we're playing with them looking at it, and

464
00:24:02,000 --> 00:24:04,000
Speaker 2: he looks at and one point he goes, this is it,

465
00:24:04,080 --> 00:24:05,919
Speaker 2: Like this is the object I've been reading about for

466
00:24:05,960 --> 00:24:09,280
Speaker 2: all these years, and we've moved though from like, oh,

467
00:24:09,520 --> 00:24:11,760
Speaker 2: the iPhone is this object of wonder and science fiction

468
00:24:11,840 --> 00:24:15,040
Speaker 2: in the future, to like like complaining about like camera

469
00:24:15,080 --> 00:24:17,639
Speaker 2: resolution or battery life and things like that. Like we've

470
00:24:17,760 --> 00:24:21,000
Speaker 2: so quickly adapted, which, on the one hand, is good

471
00:24:21,080 --> 00:24:22,760
Speaker 2: and kind of gives me hope that we are going

472
00:24:22,800 --> 00:24:26,119
Speaker 2: to figure out ways of adapting to new types of complexity.

473
00:24:26,280 --> 00:24:28,800
Speaker 2: On the other hand, though, it means that we sometimes

474
00:24:28,840 --> 00:24:32,080
Speaker 2: don't necessarily retain that capacity for wonder or when it

475
00:24:32,119 --> 00:24:35,240
Speaker 2: comes to complexity and the complex systems around us, maybe

476
00:24:35,520 --> 00:24:38,600
Speaker 2: a more critical stance and actually saying okay, like how

477
00:24:38,680 --> 00:24:40,240
Speaker 2: like what are the kind of systems we want to

478
00:24:40,240 --> 00:24:43,000
Speaker 2: be embedded within and as opposed to kind of just

479
00:24:43,040 --> 00:24:45,840
Speaker 2: allowing them to kind of wash over us. But I

480
00:24:45,880 --> 00:24:48,480
Speaker 2: do think that kind of adaptive capacity does give me

481
00:24:48,480 --> 00:24:49,240
Speaker 2: a little bit of hope.

482
00:24:49,359 --> 00:24:52,120
Speaker 1: Yes, you probably know this routine from the comedian Louis

483
00:24:52,160 --> 00:24:54,639
Speaker 1: c k where the first time he's ever on an

484
00:24:54,640 --> 00:24:56,600
Speaker 1: airplane and they announce we have Wi Fi in the

485
00:24:56,640 --> 00:24:59,760
Speaker 1: airplane and he's amazed. Everyone on the plane is amazing

486
00:25:00,000 --> 00:25:02,640
Speaker 1: I ever heard of this, And then ten minutes into

487
00:25:02,680 --> 00:25:05,560
Speaker 1: the flight, the Wi Fi breaks, it stops working in

488
00:25:05,600 --> 00:25:08,240
Speaker 1: the guy next to him starts complaining, and Luisy cases

489
00:25:08,280 --> 00:25:10,639
Speaker 1: you know, ten minutes ago you didn't even know this existed,

490
00:25:10,720 --> 00:25:13,280
Speaker 1: and now you're complaining about it. So yes, it is

491
00:25:13,320 --> 00:25:16,000
Speaker 1: true that we adapt so quickly to that. Okay, so

492
00:25:16,080 --> 00:25:18,960
Speaker 1: let me ask you something really random. Given the evolution

493
00:25:19,240 --> 00:25:22,480
Speaker 1: of the complexity all around us, what is your opinion

494
00:25:22,520 --> 00:25:24,880
Speaker 1: on whether we are already living in a simulation?

495
00:25:25,119 --> 00:25:27,880
Speaker 2: For me, I like to think about it much more

496
00:25:27,920 --> 00:25:30,720
Speaker 2: as if you kind of not necessarily take the simulation

497
00:25:30,800 --> 00:25:34,359
Speaker 2: hypothesis seriously as this like question of great importance, but

498
00:25:34,400 --> 00:25:36,280
Speaker 2: as like, oh, a question that kind of leads me

499
00:25:36,320 --> 00:25:39,239
Speaker 2: to think about more things around physics and computing. Then

500
00:25:39,280 --> 00:25:40,879
Speaker 2: I think it can actually be very productive. So the

501
00:25:40,920 --> 00:25:44,240
Speaker 2: question becomes like, in the same way that people talk

502
00:25:44,280 --> 00:25:48,919
Speaker 2: about the simulation hypothesis, there are interesting aspects around, like

503
00:25:49,160 --> 00:25:51,560
Speaker 2: breaking out of a computer program when you were inside it,

504
00:25:51,840 --> 00:25:56,800
Speaker 2: or like the like the high resolution fidelity of computer games,

505
00:25:57,320 --> 00:26:00,920
Speaker 2: or even just the ways in which physic in reality

506
00:26:00,960 --> 00:26:04,560
Speaker 2: and computing intersect. Oftentimes, when we think about computation, we

507
00:26:04,560 --> 00:26:08,159
Speaker 2: think of it as this kind of like ephemeral information stuff,

508
00:26:08,800 --> 00:26:10,960
Speaker 2: and I think that's a really powerful way of thinking

509
00:26:10,960 --> 00:26:13,920
Speaker 2: about it. But the truth is, like computing and computers,

510
00:26:13,920 --> 00:26:17,359
Speaker 2: like they are deeply physical. So the like the Internet,

511
00:26:17,440 --> 00:26:19,840
Speaker 2: like the Internet is not just information kind of whizzing around.

512
00:26:20,760 --> 00:26:22,560
Speaker 2: It is kind of to use the term from like

513
00:26:22,600 --> 00:26:25,280
Speaker 2: the that senator a number of years ago who was

514
00:26:25,280 --> 00:26:27,600
Speaker 2: like widely mocked, like it is a series of tubes.

515
00:26:27,840 --> 00:26:29,639
Speaker 2: And actually there's a book called Tubes based on that

516
00:26:29,720 --> 00:26:31,800
Speaker 2: of like the physical infrastructure of the Internet, Like there

517
00:26:31,840 --> 00:26:34,280
Speaker 2: is a lot of this physicality, and and I think

518
00:26:34,520 --> 00:26:38,480
Speaker 2: thinking about that the physical nature of our computing can

519
00:26:38,520 --> 00:26:41,240
Speaker 2: be really powerful. And sometimes the simulation hypothesis, like thinking

520
00:26:41,240 --> 00:26:43,560
Speaker 2: about it can can help heighten that or can just

521
00:26:43,640 --> 00:26:45,560
Speaker 2: make you realize, oh, there's some interesting bugs that are

522
00:26:45,600 --> 00:26:47,480
Speaker 2: worth thinking about. So for example, there was a story

523
00:26:47,520 --> 00:26:49,720
Speaker 2: I read where I think it was like in some

524
00:26:49,880 --> 00:26:53,399
Speaker 2: hospital people noticed that iPhones stopped working when they were

525
00:26:53,560 --> 00:26:56,840
Speaker 2: near one MRI machine. But it wasn't Android phones, it

526
00:26:56,880 --> 00:26:58,960
Speaker 2: wasn't on there. It was like just Apple products. And

527
00:26:58,960 --> 00:27:01,280
Speaker 2: it turned out it happened to be that some sort

528
00:27:01,320 --> 00:27:04,520
Speaker 2: of switch or some other component within these Apple devices

529
00:27:04,640 --> 00:27:08,960
Speaker 2: it had some small enough gap that it happened to

530
00:27:08,960 --> 00:27:11,080
Speaker 2: be this MRI machine had had a helium leak, and

531
00:27:11,119 --> 00:27:13,040
Speaker 2: the helium atoms were just the right size to kind

532
00:27:13,080 --> 00:27:15,639
Speaker 2: of get into this machine but didn't affect Android devices

533
00:27:15,720 --> 00:27:17,720
Speaker 2: or other things. And so it was this wild thing

534
00:27:17,960 --> 00:27:21,959
Speaker 2: that just brought home like the deep, deeply physical nature

535
00:27:22,200 --> 00:27:25,040
Speaker 2: of computing. And so for me, when I think about

536
00:27:25,040 --> 00:27:26,919
Speaker 2: the simulation hypothesis, I don't think about it as like,

537
00:27:27,119 --> 00:27:30,119
Speaker 2: oh no, like I'm being controlled by aliens or humans

538
00:27:30,119 --> 00:27:32,040
Speaker 2: in the future or whatever it is. It's much more about, Okay,

539
00:27:32,119 --> 00:27:35,200
Speaker 2: how do I think about breaking open computer games, or

540
00:27:35,520 --> 00:27:37,760
Speaker 2: like the deeply physical nature of bugs and like all this,

541
00:27:37,920 --> 00:27:39,920
Speaker 2: like that's the kind of stuff that I find most

542
00:27:39,960 --> 00:27:43,360
Speaker 2: interesting about the simulation hypothesis. I also think about it

543
00:27:43,400 --> 00:27:47,960
Speaker 2: as for me, it's almost this like cry for myth

544
00:27:48,160 --> 00:27:51,440
Speaker 2: in like the in the tech world where it's like, oh,

545
00:27:51,720 --> 00:27:55,000
Speaker 2: like we're a deeply like like the Silton Valley world

546
00:27:55,080 --> 00:27:59,040
Speaker 2: is like deeply rational, deeply logical, but we still kind

547
00:27:59,040 --> 00:28:01,800
Speaker 2: of need some sort of myth or store organizing story

548
00:28:01,880 --> 00:28:06,000
Speaker 2: in our world. And the simulation hypothesis and ideas around

549
00:28:06,040 --> 00:28:09,600
Speaker 2: the singularity, certain ideas around longevity or AI or things

550
00:28:09,600 --> 00:28:12,800
Speaker 2: like that, they often many many times they're also based

551
00:28:12,840 --> 00:28:17,080
Speaker 2: on technology, but when they kind of get big enough,

552
00:28:17,359 --> 00:28:21,159
Speaker 2: those ideas kind of veer into kind of myth and storyland.

553
00:28:21,240 --> 00:28:23,760
Speaker 2: And so for me, it's I view that as kind

554
00:28:23,760 --> 00:28:25,959
Speaker 2: of when people take those ideas a little bit more seriously,

555
00:28:26,000 --> 00:28:29,440
Speaker 2: it's much more around okay, fitting kind of a certain

556
00:28:29,440 --> 00:28:31,320
Speaker 2: amount of myth into kind of that myth shape hole

557
00:28:31,359 --> 00:28:32,520
Speaker 2: for those for those type people.

558
00:28:49,600 --> 00:28:51,880
Speaker 1: So let's return to your grandfather and the iPhone. So

559
00:28:52,280 --> 00:28:54,880
Speaker 1: how do you recommend in your book and in your

560
00:28:54,920 --> 00:28:59,479
Speaker 1: life preserving our sense of magic around the technology that

561
00:28:59,480 --> 00:28:59,800
Speaker 1: we have.

562
00:29:00,400 --> 00:29:03,120
Speaker 2: When I think about like magic and wonder and kind

563
00:29:03,120 --> 00:29:08,640
Speaker 2: of delight in computing, it's never really been an either

564
00:29:08,800 --> 00:29:11,000
Speaker 2: or of like, oh, like there's either kind of like

565
00:29:11,120 --> 00:29:14,480
Speaker 2: corporate SaaS software or kind of like the fun weird things.

566
00:29:15,520 --> 00:29:17,080
Speaker 2: And I feel like you can kind of tell some

567
00:29:17,120 --> 00:29:18,640
Speaker 2: people might tell a story of like oh, there used

568
00:29:18,640 --> 00:29:20,120
Speaker 2: to be more of that kind of wondrous stuff and

569
00:29:20,160 --> 00:29:23,160
Speaker 2: now we've kind of it's all like we're we're just

570
00:29:23,200 --> 00:29:27,120
Speaker 2: kind of locked into large social media sites or just

571
00:29:27,240 --> 00:29:29,880
Speaker 2: we're just using these like a large, kind of bland,

572
00:29:30,040 --> 00:29:33,520
Speaker 2: beige pieces of software. I think there is an element

573
00:29:33,600 --> 00:29:38,800
Speaker 2: of that, but the truth is these two aspects of computing,

574
00:29:38,800 --> 00:29:42,640
Speaker 2: they've always co existed, and so like alongside like the

575
00:29:42,720 --> 00:29:45,880
Speaker 2: really big mainframe or refrigerator sized computers, there were people

576
00:29:45,960 --> 00:29:48,800
Speaker 2: trying to build like early computer games, and then once

577
00:29:48,800 --> 00:29:51,200
Speaker 2: we had personal computers, there was a lot of fun,

578
00:29:51,240 --> 00:29:55,320
Speaker 2: weird things, people experimenting with fractals, but also people using

579
00:29:55,320 --> 00:29:57,640
Speaker 2: spreadsheets in businesses, and so it's not an either or

580
00:29:57,680 --> 00:29:59,400
Speaker 2: any and the truth is even on the web now

581
00:30:00,080 --> 00:30:03,280
Speaker 2: alongside kind of the large websites, there are also people

582
00:30:03,360 --> 00:30:07,000
Speaker 2: talking about there's a term called the poetic web, where

583
00:30:07,000 --> 00:30:09,960
Speaker 2: it's like the kind of the more human scale, fun,

584
00:30:10,400 --> 00:30:14,240
Speaker 2: funkier and weirder sort of websites. And for me, it's

585
00:30:14,280 --> 00:30:18,560
Speaker 2: really just a matter of trying to actually like discover

586
00:30:18,640 --> 00:30:20,400
Speaker 2: these kinds of things and realize that it's that it's

587
00:30:20,440 --> 00:30:22,720
Speaker 2: always been out there and it's really just a matter

588
00:30:22,920 --> 00:30:25,280
Speaker 2: of being able to find it. And so for me,

589
00:30:25,920 --> 00:30:27,560
Speaker 2: I kind of view some of the ideas in the

590
00:30:27,560 --> 00:30:30,320
Speaker 2: book almost as like a proof of existence of like, oh,

591
00:30:30,360 --> 00:30:32,280
Speaker 2: these things do exist out that out there. You don't

592
00:30:32,280 --> 00:30:34,160
Speaker 2: necessarily have to be as excited as I am by

593
00:30:34,200 --> 00:30:36,960
Speaker 2: some of the examples I give, but let that be

594
00:30:37,560 --> 00:30:40,640
Speaker 2: a guide to ohkay, there are other things out there

595
00:30:40,720 --> 00:30:45,160
Speaker 2: that are just worth enjoying and experiencing and delighting it.

596
00:30:45,600 --> 00:30:48,280
Speaker 2: And I think part of that often is just kind

597
00:30:48,320 --> 00:30:52,440
Speaker 2: of at the smaller scale. And I do actually think

598
00:30:53,120 --> 00:30:56,120
Speaker 2: that one of the exciting things about AI generated code

599
00:30:56,440 --> 00:30:59,240
Speaker 2: is it really allows for this kind of democratization of

600
00:30:59,360 --> 00:31:01,880
Speaker 2: building soft where and so people have talked about this

601
00:31:01,920 --> 00:31:03,160
Speaker 2: kind of thing for a very long time of like

602
00:31:03,320 --> 00:31:05,680
Speaker 2: it shouldn't just be the domain of like big companies

603
00:31:05,800 --> 00:31:08,080
Speaker 2: or kind of serious software developers of building things that

604
00:31:08,120 --> 00:31:10,720
Speaker 2: are going to be used by like millions or hundreds

605
00:31:10,720 --> 00:31:12,640
Speaker 2: of millions of people. There should be a way for

606
00:31:12,760 --> 00:31:15,120
Speaker 2: each individual user to kind of build the bespoke thing

607
00:31:15,120 --> 00:31:18,200
Speaker 2: they want. And so the novelist Robin Sloan has this phrase.

608
00:31:18,600 --> 00:31:20,040
Speaker 2: I think it's like an app can be a home

609
00:31:20,040 --> 00:31:22,480
Speaker 2: cooked meal, this idea that like, you don't necessarily need

610
00:31:22,640 --> 00:31:24,840
Speaker 2: to build something for everyone. You can build like a

611
00:31:24,880 --> 00:31:27,400
Speaker 2: little program for yourself or for your loved ones, and

612
00:31:27,440 --> 00:31:29,720
Speaker 2: that's fine, and that's great. In fact, and the true

613
00:31:29,720 --> 00:31:31,719
Speaker 2: this spreadsheets were actually a simple version of this kind

614
00:31:31,720 --> 00:31:33,760
Speaker 2: of thing, because you can actually program in very simple ways.

615
00:31:33,760 --> 00:31:35,520
Speaker 2: But then there was and there was HyperCard with kind

616
00:31:35,520 --> 00:31:37,120
Speaker 2: of some of the early macintoshes where it was like

617
00:31:37,120 --> 00:31:39,760
Speaker 2: this authoring program to kind of build like weird little

618
00:31:40,600 --> 00:31:43,040
Speaker 2: sort of like pseudo website programs on your own computer.

619
00:31:43,200 --> 00:31:48,040
Speaker 2: But now with AI generated code, I really see this

620
00:31:48,120 --> 00:31:52,240
Speaker 2: kind of democratization potential really blossoming. And and so for me,

621
00:31:52,680 --> 00:31:54,760
Speaker 2: like that is the kind of thing that really can

622
00:31:54,800 --> 00:31:57,120
Speaker 2: hopefully induce a sense of wonder and people where like

623
00:31:57,200 --> 00:31:59,320
Speaker 2: they can now build all the programs that they want,

624
00:31:59,360 --> 00:32:01,840
Speaker 2: and so to have to be that if you kind

625
00:32:01,840 --> 00:32:03,680
Speaker 2: of explored the world and kind of went about your

626
00:32:03,760 --> 00:32:05,760
Speaker 2: day and looked at the world and noticed interesting problems

627
00:32:05,800 --> 00:32:08,280
Speaker 2: that could maybe be solved by software, if you weren't

628
00:32:08,280 --> 00:32:09,920
Speaker 2: a software developer, you would kind of have to shut

629
00:32:09,960 --> 00:32:12,000
Speaker 2: down that portion of your mind because you couldn't do

630
00:32:12,040 --> 00:32:14,360
Speaker 2: anything about it. But now you can turn it back

631
00:32:14,400 --> 00:32:16,320
Speaker 2: on because now anyone can build those kinds of things.

632
00:32:16,360 --> 00:32:17,920
Speaker 2: And so I think that is actually a really interesting

633
00:32:17,960 --> 00:32:18,640
Speaker 2: source for wonder.

634
00:32:18,720 --> 00:32:21,080
Speaker 1: And you know, to my mind, there's the flip side

635
00:32:21,080 --> 00:32:24,560
Speaker 1: of that coin, not just for the individual, but for society.

636
00:32:24,600 --> 00:32:27,120
Speaker 1: What's going to come out of this? So tell us

637
00:32:27,120 --> 00:32:30,120
Speaker 1: about Dawn Swanson's paper from what was that I think

638
00:32:30,160 --> 00:32:33,360
Speaker 1: the eighties or something about undiscovered public knowledge. Tell us

639
00:32:33,360 --> 00:32:33,720
Speaker 1: about that.

640
00:32:34,480 --> 00:32:37,120
Speaker 2: Yeah, so Don Swanson, Yeah, he's this information scientist. In

641
00:32:37,120 --> 00:32:41,120
Speaker 2: the nineteen eighties, he wrote this paper called Undisovered Public Knowledge,

642
00:32:41,160 --> 00:32:45,400
Speaker 2: where the idea behind it was and he begins kind

643
00:32:45,400 --> 00:32:47,400
Speaker 2: of like with a thought experiment. He says, Okay, imagine,

644
00:32:47,480 --> 00:32:50,440
Speaker 2: somewhere in the scientific literature there's a paper that says

645
00:32:50,600 --> 00:32:52,920
Speaker 2: A implies B, and then somewhere else in the literature

646
00:32:52,960 --> 00:32:54,840
Speaker 2: could be in the same field, it could be an

647
00:32:54,960 --> 00:32:57,920
Speaker 2: entirely different field. There's another paper that says BE implies C.

648
00:32:58,320 --> 00:33:00,400
Speaker 2: And so if you would connect to them together, you

649
00:33:00,440 --> 00:33:03,040
Speaker 2: would say, oh, maybe in fact A IMPLYI C. But

650
00:33:03,120 --> 00:33:06,239
Speaker 2: because the scientific literature is so vast, no one has

651
00:33:06,240 --> 00:33:08,600
Speaker 2: actually been able to read these two papers, and so

652
00:33:09,160 --> 00:33:12,680
Speaker 2: that that knowledge, the connection was undiscovered, but it was

653
00:33:12,680 --> 00:33:14,479
Speaker 2: public because it was out there. And so it's one

654
00:33:14,480 --> 00:33:17,000
Speaker 2: of these things where if we actually had ways of

655
00:33:17,040 --> 00:33:20,480
Speaker 2: stitching together all the scientific knowledge that was out there,

656
00:33:20,560 --> 00:33:22,320
Speaker 2: we would actually be able to make new discoveries that

657
00:33:22,320 --> 00:33:24,680
Speaker 2: were kind of just lying out there, ready for the taking.

658
00:33:24,960 --> 00:33:26,760
Speaker 2: The interesting thing with Swanson is that he was not

659
00:33:26,880 --> 00:33:29,240
Speaker 2: content with leaving this as a thought experiment. He actually

660
00:33:29,320 --> 00:33:31,720
Speaker 2: tried to test it in the real world, and he

661
00:33:31,880 --> 00:33:34,520
Speaker 2: used in the then cutting edge technology, which was I

662
00:33:34,560 --> 00:33:37,600
Speaker 2: think like using keyword searches on like the medline database.

663
00:33:37,680 --> 00:33:40,680
Speaker 2: But he actually found this relationship between I think I

664
00:33:40,720 --> 00:33:43,320
Speaker 2: think it was consuming fish oil and then helping treat

665
00:33:43,360 --> 00:33:45,560
Speaker 2: some sort of circulatory disorder, and then he was I

666
00:33:45,560 --> 00:33:47,360
Speaker 2: think he was able to publish it in a medical

667
00:33:47,440 --> 00:33:49,880
Speaker 2: journal even though he himself had no medical training, which

668
00:33:49,920 --> 00:33:52,440
Speaker 2: was kind of wild and so and I think with

669
00:33:53,000 --> 00:33:54,760
Speaker 2: all like we had with a lot of these AI tools,

670
00:33:54,800 --> 00:33:56,360
Speaker 2: we are now going to be able to kind of

671
00:33:56,400 --> 00:33:59,840
Speaker 2: stitch together lots of different ideas and kind of navigate

672
00:33:59,880 --> 00:34:01,960
Speaker 2: them like the latent space of knowledge or however you

673
00:34:02,000 --> 00:34:04,120
Speaker 2: want to describe it, in a way that that has

674
00:34:04,240 --> 00:34:06,160
Speaker 2: really never before been possible.

675
00:34:07,800 --> 00:34:10,840
Speaker 1: This is actually my highest hope with large language models

676
00:34:11,320 --> 00:34:17,359
Speaker 1: is tackling the biomedical data, in putting facts together that

677
00:34:17,480 --> 00:34:20,320
Speaker 1: anyone could know, but nobody is going to because they're

678
00:34:20,320 --> 00:34:23,840
Speaker 1: published in totally different journals. I wrote a paper a

679
00:34:23,880 --> 00:34:27,719
Speaker 1: couple of years ago now on a meaningful test for

680
00:34:27,840 --> 00:34:32,040
Speaker 1: intelligence in AI, and I think that what I just

681
00:34:32,080 --> 00:34:35,080
Speaker 1: described that's going to be enormously helpful for science. But

682
00:34:35,480 --> 00:34:38,000
Speaker 1: the next level of intelligence which I don't think llms

683
00:34:38,000 --> 00:34:43,160
Speaker 1: are at yet is actually questioning whether something is true

684
00:34:43,480 --> 00:34:47,759
Speaker 1: and coming up with alternative models and then simulating those

685
00:34:47,800 --> 00:34:51,680
Speaker 1: and evaluating them. For example, you know, saying, hey, what

686
00:34:51,719 --> 00:34:53,600
Speaker 1: if I were writing on a photon of light, what

687
00:34:53,640 --> 00:34:57,080
Speaker 1: would that look like? And then getting to the theory

688
00:34:57,080 --> 00:35:00,680
Speaker 1: of relativity and realizing the trajectory of merch can be

689
00:35:00,719 --> 00:35:03,200
Speaker 1: explained by that, and so on. So that's the sort

690
00:35:03,200 --> 00:35:06,520
Speaker 1: of thing that lms don't do now. But yes, I

691
00:35:06,520 --> 00:35:11,200
Speaker 1: think they're going to be enormously helpful in this discovery

692
00:35:11,520 --> 00:35:14,439
Speaker 1: process within the public knowledge. It's already sitting out there.

693
00:35:14,520 --> 00:35:17,040
Speaker 2: People are already talking about, like AI scientists and things

694
00:35:17,040 --> 00:35:18,239
Speaker 2: like that, like whether or not it's going to be

695
00:35:18,280 --> 00:35:22,560
Speaker 2: helping with or it's stitching together the knowledge hypothesis generation.

696
00:35:22,719 --> 00:35:25,719
Speaker 2: Maybe eventually even yeah, this kind of like thought experiment

697
00:35:25,800 --> 00:35:28,000
Speaker 2: and then examining like what are the implications of the

698
00:35:28,000 --> 00:35:30,719
Speaker 2: thought experiments. But I do think, yeah, even if they're

699
00:35:30,719 --> 00:35:33,040
Speaker 2: not necessarily able to kind of do everything on their own,

700
00:35:33,640 --> 00:35:36,200
Speaker 2: the potential for this kind of like science like human

701
00:35:36,239 --> 00:35:40,160
Speaker 2: scientist machine partnership will well hopefully unlock a lot of

702
00:35:40,200 --> 00:35:42,080
Speaker 2: information and knowledge that is already out there, but we

703
00:35:42,120 --> 00:35:43,000
Speaker 2: just don't even realize it.

704
00:35:43,280 --> 00:35:46,880
Speaker 1: What's one thing that you wish more people understood about

705
00:35:46,920 --> 00:35:48,960
Speaker 1: the coded systems that surround them.

706
00:35:49,040 --> 00:35:51,839
Speaker 2: One aspect about code is that the extent to which

707
00:35:51,880 --> 00:35:55,160
Speaker 2: there's like a craft and a style and almost an

708
00:35:55,239 --> 00:35:57,640
Speaker 2: art to it. And when people think about like programming

709
00:35:57,680 --> 00:35:59,520
Speaker 2: languages or kind of which language they want to program,

710
00:35:59,600 --> 00:36:03,880
Speaker 2: and the truth is there's there's a lot of personal

711
00:36:03,960 --> 00:36:07,040
Speaker 2: choice and a lot of a lot of opinion, very

712
00:36:07,040 --> 00:36:10,520
Speaker 2: strong opinions about what kind of languages work, but also

713
00:36:10,560 --> 00:36:12,480
Speaker 2: even kind of the way in which you program. And

714
00:36:12,480 --> 00:36:16,440
Speaker 2: so for example, there's actually this book called If Hemingway

715
00:36:16,480 --> 00:36:18,719
Speaker 2: Wrote JavaScript, where it actually takes I think, like the

716
00:36:18,719 --> 00:36:22,400
Speaker 2: same coding task and then programs in different ways accordinated

717
00:36:22,480 --> 00:36:25,759
Speaker 2: kind of different like authorial styles, and to show that

718
00:36:26,000 --> 00:36:28,920
Speaker 2: it is a deeply human kind of thing. Now, of course,

719
00:36:29,360 --> 00:36:33,360
Speaker 2: and many people compare it to writing or fiction or

720
00:36:33,400 --> 00:36:36,040
Speaker 2: poetry and things like that, and there are aspects of that,

721
00:36:36,080 --> 00:36:38,160
Speaker 2: and I certainly, but I don't want to push it

722
00:36:38,200 --> 00:36:40,319
Speaker 2: too far because the code still has to do something,

723
00:36:40,360 --> 00:36:43,320
Speaker 2: it still has to operate. But there really are many

724
00:36:43,719 --> 00:36:47,480
Speaker 2: almost like artistic aspects to code, and I think that

725
00:36:47,560 --> 00:36:52,640
Speaker 2: kind of interesting combination of extreme logic and practicality and

726
00:36:52,680 --> 00:36:58,160
Speaker 2: efficacy combined with style and art and problem solving, I

727
00:36:58,200 --> 00:36:59,960
Speaker 2: think is something that maybe people who are kind of

728
00:37:00,040 --> 00:37:02,000
Speaker 2: outside of the world of code just don't realize.

729
00:37:02,080 --> 00:37:06,600
Speaker 1: So here's a random question, the relationship between software code

730
00:37:06,680 --> 00:37:09,719
Speaker 1: and let's say, biological code like DNA. Is this just

731
00:37:09,760 --> 00:37:12,080
Speaker 1: a metaphorical thing or is there something deeper there?

732
00:37:12,160 --> 00:37:15,640
Speaker 2: We are wet, squishy, messy things, and then down at

733
00:37:15,680 --> 00:37:21,080
Speaker 2: the sollular or substolular level, it's incredibly incredibly stochastic and random,

734
00:37:21,080 --> 00:37:24,480
Speaker 2: and there's things just all vibrating around, and it's wildly

735
00:37:24,560 --> 00:37:27,280
Speaker 2: different from the way in which we think about coding.

736
00:37:27,680 --> 00:37:31,319
Speaker 2: But the one exciting aspect about this is that, like

737
00:37:31,480 --> 00:37:33,560
Speaker 2: Mike Levin and some of his and his collaborators, they've

738
00:37:33,560 --> 00:37:38,640
Speaker 2: talked about this idea that really traditional computation is really

739
00:37:38,680 --> 00:37:42,280
Speaker 2: just a subset of kind of information processing as a whole,

740
00:37:42,600 --> 00:37:45,399
Speaker 2: and biology is just another mode of doing that kind

741
00:37:45,400 --> 00:37:47,560
Speaker 2: of thing. So I think by looking at what biology

742
00:37:47,600 --> 00:37:50,560
Speaker 2: is doing and comparing it to how code operates and

743
00:37:50,560 --> 00:37:53,280
Speaker 2: how computers operate, where they are similar and where they're different,

744
00:37:54,239 --> 00:37:58,239
Speaker 2: just shows you the sheer number of different ways that

745
00:37:58,360 --> 00:38:00,759
Speaker 2: computing can be done, and when it comes to kind

746
00:38:00,800 --> 00:38:03,640
Speaker 2: of more engineered traditional computing, we are still only beginning

747
00:38:03,640 --> 00:38:06,160
Speaker 2: to scratch the surface. So I think in that way,

748
00:38:06,360 --> 00:38:10,000
Speaker 2: comparing contrasting the way biology and computation are similar and

749
00:38:10,000 --> 00:38:11,640
Speaker 2: different can be enormously valuable.

750
00:38:11,880 --> 00:38:14,719
Speaker 1: So let's end with telling us what your message is

751
00:38:14,800 --> 00:38:17,520
Speaker 1: at the heart of the magic of Code, your new book.

752
00:38:17,600 --> 00:38:20,000
Speaker 1: What do you hope that people will take away from

753
00:38:20,000 --> 00:38:22,040
Speaker 1: it in seeing the world around them?

754
00:38:22,360 --> 00:38:25,400
Speaker 2: Like Steve Jobs has this idea that computers are the

755
00:38:25,400 --> 00:38:28,000
Speaker 2: bicycle for the mind. And the idea behind this was

756
00:38:28,239 --> 00:38:30,200
Speaker 2: that he was reading I think some old scientific American

757
00:38:30,280 --> 00:38:31,960
Speaker 2: article where it was like a chart of like the

758
00:38:32,080 --> 00:38:36,080
Speaker 2: energy efficiency of different organisms and humans were kind of mediocre,

759
00:38:36,120 --> 00:38:37,920
Speaker 2: and like maybe some birds were much better, but then

760
00:38:37,960 --> 00:38:39,759
Speaker 2: everything changed when a human got on a bicycle, because

761
00:38:39,760 --> 00:38:42,400
Speaker 2: suddenly they were much more efficient. And his idea was

762
00:38:42,440 --> 00:38:44,359
Speaker 2: that computers they should be this bicycle for the mind,

763
00:38:44,400 --> 00:38:47,640
Speaker 2: for helping accelerate how we think, how we interact engage

764
00:38:47,680 --> 00:38:51,080
Speaker 2: with the world. And that's really ultimately what it's all about.

765
00:38:51,120 --> 00:38:54,360
Speaker 2: And so for me, whether I'm thinking about like trends

766
00:38:54,440 --> 00:38:58,360
Speaker 2: in like super powerful AI or certain other things around

767
00:38:59,000 --> 00:39:01,799
Speaker 2: the Internet or whatever it is we all have to

768
00:39:01,800 --> 00:39:04,720
Speaker 2: be thinking about, not just saying, oh, these are interesting trends.

769
00:39:04,760 --> 00:39:06,400
Speaker 2: I wonder what things are going to look like in

770
00:39:06,400 --> 00:39:10,239
Speaker 2: the future, but more no, like these tools are for me,

771
00:39:10,680 --> 00:39:12,359
Speaker 2: what is the future that I want to live in?

772
00:39:12,440 --> 00:39:14,840
Speaker 2: And how can I kind of make that human centered

773
00:39:14,880 --> 00:39:17,920
Speaker 2: future with technology that much more possible? And so the

774
00:39:17,960 --> 00:39:19,919
Speaker 2: book is kind of a guide to kind of think

775
00:39:20,000 --> 00:39:23,279
Speaker 2: about all these different sort of human centered ideas, to

776
00:39:23,280 --> 00:39:25,440
Speaker 2: hopefully provide a guide for that sense of wonder and

777
00:39:25,520 --> 00:39:28,560
Speaker 2: delight and humane aspects of computing.

778
00:39:33,200 --> 00:39:36,960
Speaker 1: That was my interview with complexity scientist and lover of code,

779
00:39:37,320 --> 00:39:40,520
Speaker 1: Sam Arbusman. We are right now at the beginning of

780
00:39:40,600 --> 00:39:46,279
Speaker 1: a centuries long experiment in computation. For the first time

781
00:39:46,320 --> 00:39:50,360
Speaker 1: in history, we can build these dynamic worlds that evolve

782
00:39:50,440 --> 00:39:55,239
Speaker 1: and adapt. We can simulate climate futures, or economic collapses,

783
00:39:55,440 --> 00:40:00,960
Speaker 1: or entire societies that rise and fall in silic But

784
00:40:01,080 --> 00:40:04,360
Speaker 1: Sam talks about code not just as a set of instructions,

785
00:40:04,440 --> 00:40:08,200
Speaker 1: but more generally like a kind of spell, a system

786
00:40:08,239 --> 00:40:12,320
Speaker 1: of symbols that does something real in the outside world.

787
00:40:12,719 --> 00:40:14,919
Speaker 1: And part of what I find the most amazing about

788
00:40:14,920 --> 00:40:18,160
Speaker 1: our current moment in time is the way that code

789
00:40:18,520 --> 00:40:23,880
Speaker 1: can and has evolved past the understanding of its creators.

790
00:40:24,200 --> 00:40:27,120
Speaker 1: And that's the paradox we're sitting with, and in some

791
00:40:27,160 --> 00:40:29,600
Speaker 1: sense we have been sitting with for some centuries, now

792
00:40:29,800 --> 00:40:33,759
Speaker 1: that we are building systems more powerful than our ability

793
00:40:34,040 --> 00:40:36,920
Speaker 1: to fully understand them. There's one more thing I just

794
00:40:36,960 --> 00:40:40,320
Speaker 1: want to touch on from Sam's book, the idea that code,

795
00:40:40,960 --> 00:40:44,520
Speaker 1: like language or like myth, offers us a kind of mirror.

796
00:40:45,160 --> 00:40:49,439
Speaker 1: Code can reflect our values and our metaphors, and our

797
00:40:49,719 --> 00:40:54,000
Speaker 1: hopes for control, and our particular curiosities about how the

798
00:40:54,040 --> 00:40:57,200
Speaker 1: world works. I suspect that someday there are going to

799
00:40:57,239 --> 00:41:01,600
Speaker 1: be code anthropologists who look back on the kind of

800
00:41:01,680 --> 00:41:05,839
Speaker 1: programs written by different civilizations at different time points, and

801
00:41:05,880 --> 00:41:09,640
Speaker 1: it will tell them as much about those civilizations as

802
00:41:09,960 --> 00:41:13,360
Speaker 1: their books and plays and religious practices, because our code

803
00:41:13,400 --> 00:41:16,600
Speaker 1: reflects the assumptions that we build into them and the

804
00:41:16,680 --> 00:41:20,879
Speaker 1: blind spots that we forget to consider. Every simulation has

805
00:41:20,880 --> 00:41:24,839
Speaker 1: some of us in there. So with the passing of decades,

806
00:41:25,200 --> 00:41:27,560
Speaker 1: we're going to go beyond how do we code the

807
00:41:27,560 --> 00:41:31,359
Speaker 1: world to questions about who's doing the coding, and what

808
00:41:31,400 --> 00:41:33,400
Speaker 1: do we make sure is in there, and what do

809
00:41:33,440 --> 00:41:36,200
Speaker 1: we choose to leave out? And what are the limits

810
00:41:36,280 --> 00:41:38,440
Speaker 1: of what we can simulate? And when does it matter

811
00:41:38,480 --> 00:41:43,399
Speaker 1: that those limits shape our conclusions? In any case, as

812
00:41:43,400 --> 00:41:46,080
Speaker 1: I think about what Sam and I talked about, I

813
00:41:46,120 --> 00:41:50,200
Speaker 1: come to something like this conclusion. As we look into

814
00:41:50,239 --> 00:41:54,839
Speaker 1: the deep future, we may find ourselves less in control

815
00:41:54,960 --> 00:41:58,920
Speaker 1: than we thought, but also more creative than we ever

816
00:41:59,000 --> 00:42:03,719
Speaker 1: thought possible. Our minds will be riding bicycles and eventually

817
00:42:03,960 --> 00:42:07,640
Speaker 1: motorcycles and jets. And that is the sense in which

818
00:42:07,920 --> 00:42:11,440
Speaker 1: even the most logical systems we've ever built have a

819
00:42:11,520 --> 00:42:20,000
Speaker 1: healthy dose of magic. Go to eagleman dot com slash

820
00:42:20,040 --> 00:42:23,840
Speaker 1: podcast for more information and find further reading. Join the

821
00:42:23,840 --> 00:42:27,680
Speaker 1: weekly discussions on my substack, and check out and subscribe

822
00:42:27,719 --> 00:42:30,760
Speaker 1: to Inner Cosmos on YouTube for videos of each episode

823
00:42:30,800 --> 00:42:35,720
Speaker 1: and to leave comments until next time. I'm David Eagleman,

824
00:42:35,880 --> 00:42:37,759
Speaker 1: and this is Inner Cosmos.