1
00:00:01,480 --> 00:00:04,960
Speaker 1: Welcome to tech Stuff. This is the story. Each week

2
00:00:05,000 --> 00:00:07,560
Speaker 1: on Wednesdays, we bring you an in depth interview with

3
00:00:07,600 --> 00:00:10,039
Speaker 1: someone who has a front row seat to the most

4
00:00:10,039 --> 00:00:14,680
Speaker 1: fascinating things happening in tech today. We're joined by David Eagleman.

5
00:00:20,120 --> 00:00:25,520
Speaker 1: David Eagleman is a neuroscientist, Stanford University professor, author, entrepreneur,

6
00:00:25,840 --> 00:00:29,920
Speaker 1: and host of the podcast Inner Cosmos. In his podcast,

7
00:00:30,160 --> 00:00:33,599
Speaker 1: Eagleman explores how our brain interprets the world and how

8
00:00:33,680 --> 00:00:37,400
Speaker 1: that influences our perception. Ideas have been a source of

9
00:00:37,440 --> 00:00:41,360
Speaker 1: fascination for Eagleman ever since he experienced an Alice in

10
00:00:41,440 --> 00:00:43,440
Speaker 1: Wonderland moment many years ago.

11
00:00:44,240 --> 00:00:47,240
Speaker 2: I fell off of the roof of a house when

12
00:00:47,280 --> 00:00:48,160
Speaker 2: I was a kid.

13
00:00:48,159 --> 00:00:49,000
Speaker 1: I was eight years old.

14
00:00:49,440 --> 00:00:52,239
Speaker 2: The fall seemed to take a very long time, and

15
00:00:52,320 --> 00:00:54,720
Speaker 2: I was thinking about Alice in Wonderland on the way down,

16
00:00:54,760 --> 00:00:57,400
Speaker 2: about how this must have been what it was like

17
00:00:57,440 --> 00:00:58,960
Speaker 2: for her to fall down the rabbit hole.

18
00:00:59,320 --> 00:01:01,959
Speaker 1: He told me that the whole journey was very calm,

19
00:01:02,240 --> 00:01:04,560
Speaker 1: but that it took a long time, and he remembers

20
00:01:04,560 --> 00:01:07,360
Speaker 1: having lots and lots of thoughts on his way to

21
00:01:07,400 --> 00:01:08,280
Speaker 1: the bottom.

22
00:01:08,560 --> 00:01:11,000
Speaker 2: And then I hit the ground and broke my nose

23
00:01:11,080 --> 00:01:13,280
Speaker 2: very badly. But when I got older and went to

24
00:01:13,319 --> 00:01:16,399
Speaker 2: high school, I realized that the whole fall had taken

25
00:01:16,520 --> 00:01:18,959
Speaker 2: point six of a second. I calculated the eagles one

26
00:01:19,000 --> 00:01:21,319
Speaker 2: have at squared, and I couldn't believe that it was

27
00:01:21,360 --> 00:01:22,640
Speaker 2: so so fast.

28
00:01:23,040 --> 00:01:27,360
Speaker 1: When Eagleman became a neuroscientist, he kept exploring this relationship

29
00:01:27,360 --> 00:01:30,800
Speaker 1: between our brain and our experiences, which led him to

30
00:01:30,840 --> 00:01:34,200
Speaker 1: research things like is it possible to create a new

31
00:01:34,319 --> 00:01:37,600
Speaker 1: sense or is there a new way of interpreting an

32
00:01:37,600 --> 00:01:41,440
Speaker 1: already existing sense. This turned into a tech product for

33
00:01:41,480 --> 00:01:45,600
Speaker 1: his company, Neosensory, that could help deaf people understand the

34
00:01:45,600 --> 00:01:50,200
Speaker 1: auditory world through a vibrating wristband. David Eagleman is an

35
00:01:50,240 --> 00:01:54,320
Speaker 1: innovator with a fascination for tech driven interventions, so we

36
00:01:54,320 --> 00:01:56,880
Speaker 1: weren't surprised to see an episode of the Inner Cosmos

37
00:01:56,920 --> 00:02:01,680
Speaker 1: podcast all about AI relationships, a widely reported phenomenon that

38
00:02:01,680 --> 00:02:04,280
Speaker 1: we hear it tech stuff have been eager to explore.

39
00:02:04,960 --> 00:02:07,480
Speaker 1: So we thought we'd ask Eagleman to help us understand

40
00:02:07,640 --> 00:02:12,000
Speaker 1: how our brains form these attachments. But before we launch

41
00:02:12,040 --> 00:02:15,160
Speaker 1: into that discussion, I asked Eagleman to tell me a

42
00:02:15,200 --> 00:02:19,040
Speaker 1: bit more about the driving question. Behind his work, How

43
00:02:19,080 --> 00:02:21,519
Speaker 1: does the human brain construct reality?

44
00:02:22,200 --> 00:02:25,840
Speaker 2: Your brain is locked in silence and darkness inside your skull,

45
00:02:26,400 --> 00:02:29,760
Speaker 2: and it only has spikes coming in. That's all it

46
00:02:29,760 --> 00:02:33,520
Speaker 2: ever has, these little electrical signals. So your eyes aren't

47
00:02:33,919 --> 00:02:37,520
Speaker 2: pushing light through, and your ears aren't pushing sound through. Instead,

48
00:02:37,600 --> 00:02:40,600
Speaker 2: your eyes are converting photons into spikes, and ears are

49
00:02:40,600 --> 00:02:45,079
Speaker 2: converting sound waves into spikes, and your fingertips are converting

50
00:02:45,080 --> 00:02:48,880
Speaker 2: pressure and temperature into spikes and so on. So inside

51
00:02:48,919 --> 00:02:52,200
Speaker 2: the brain there's nothing but these little electrical spikes running around,

52
00:02:52,440 --> 00:02:54,760
Speaker 2: and it's living in the darkness and trying to figure

53
00:02:54,760 --> 00:02:58,000
Speaker 2: out what is the world out there? And so everything

54
00:02:58,040 --> 00:03:00,760
Speaker 2: that we experience as reality is actually the construction.

55
00:03:00,880 --> 00:03:01,120
Speaker 1: You know.

56
00:03:01,400 --> 00:03:05,079
Speaker 2: For example, things like colors don't exist in the outside world.

57
00:03:05,120 --> 00:03:08,320
Speaker 2: This is just a way to tag information as in

58
00:03:08,400 --> 00:03:11,760
Speaker 2: that's a different wavelength than that over there, so they

59
00:03:11,800 --> 00:03:14,600
Speaker 2: appear to be different colors. But this is the fascinating

60
00:03:14,760 --> 00:03:18,400
Speaker 2: question that we're stuck with. We are inside Plato's cave,

61
00:03:18,720 --> 00:03:22,240
Speaker 2: seeing shadows on the wall and trying to understand what

62
00:03:22,400 --> 00:03:23,280
Speaker 2: is outside.

63
00:03:24,000 --> 00:03:27,080
Speaker 1: Speaking of perception, did you ever figure out an answer

64
00:03:27,200 --> 00:03:30,440
Speaker 1: to that question of why time slows down when our

65
00:03:30,520 --> 00:03:31,640
Speaker 1: lives at risk.

66
00:03:32,120 --> 00:03:34,160
Speaker 2: I did a number of experiments on this, actually the

67
00:03:34,160 --> 00:03:35,760
Speaker 2: only experiments that I have ever been done on this

68
00:03:35,880 --> 00:03:38,000
Speaker 2: question of the perception of time.

69
00:03:38,560 --> 00:03:39,120
Speaker 1: Yeah, on the.

70
00:03:39,160 --> 00:03:40,880
Speaker 2: Question of why does time slow now when you're in

71
00:03:40,920 --> 00:03:42,680
Speaker 2: fear for your life. The reason no one had ever

72
00:03:42,720 --> 00:03:45,760
Speaker 2: done experiments is because you have to actually put subjects

73
00:03:45,760 --> 00:03:47,480
Speaker 2: in fear for their life, which is very hard to

74
00:03:47,480 --> 00:03:51,080
Speaker 2: get approved. And so I dropped people from one hundred

75
00:03:51,080 --> 00:03:55,520
Speaker 2: and fifty foot tall tower backwards in free fall, and

76
00:03:55,560 --> 00:03:58,120
Speaker 2: they're caught in a net below, going seventy miles an hour,

77
00:03:58,600 --> 00:04:01,680
Speaker 2: and I built a di ice that strapped to their

78
00:04:01,720 --> 00:04:03,880
Speaker 2: wrists and it flashes information at them. And I was

79
00:04:03,920 --> 00:04:08,720
Speaker 2: able to measure whether they actually are seeing in slow

80
00:04:08,800 --> 00:04:12,160
Speaker 2: motion during a scary event. And it turns out that

81
00:04:12,400 --> 00:04:15,240
Speaker 2: it's all a trick of memory. What happens when you're

82
00:04:15,280 --> 00:04:17,520
Speaker 2: in fear for your life is your brain is writing

83
00:04:17,560 --> 00:04:19,919
Speaker 2: down every single thing. Normally, your brain is like a

84
00:04:19,960 --> 00:04:22,520
Speaker 2: sieve and most things just passed right through, but you're

85
00:04:22,960 --> 00:04:26,680
Speaker 2: capturing all the information. So it's not that you're seeing

86
00:04:26,720 --> 00:04:29,640
Speaker 2: things in slow motion. It's just the density of memory.

87
00:04:29,760 --> 00:04:31,000
Speaker 2: You're remembering everything.

88
00:04:31,120 --> 00:04:31,320
Speaker 1: You know.

89
00:04:31,320 --> 00:04:33,919
Speaker 2: You're in a car actually say, wow, I remember the

90
00:04:34,000 --> 00:04:36,520
Speaker 2: hood crumpling and the rear view mirror falling off, and

91
00:04:36,560 --> 00:04:39,400
Speaker 2: the expression on the other guy's face. And you've got

92
00:04:39,440 --> 00:04:42,120
Speaker 2: all these details, and your brain is doing a calculation

93
00:04:42,360 --> 00:04:44,520
Speaker 2: where it says, oh, well, if I have all those details,

94
00:04:44,560 --> 00:04:47,640
Speaker 2: it must have taken five seconds as opposed to one

95
00:04:47,680 --> 00:04:49,080
Speaker 2: second for this all to happen.

96
00:04:49,760 --> 00:04:51,919
Speaker 1: One of the kind of topics of tech stuff is

97
00:04:52,440 --> 00:04:56,159
Speaker 1: what are the most exciting or promising paradigms for human

98
00:04:56,200 --> 00:05:00,240
Speaker 1: machine interactions, particularly a time when there's so much concern

99
00:05:00,360 --> 00:05:03,760
Speaker 1: about the negative effects of technology on the human brain. Obviously,

100
00:05:04,040 --> 00:05:06,160
Speaker 1: the question of like, what is this doing to us

101
00:05:06,160 --> 00:05:10,039
Speaker 1: in our brains and particularly developing brains. Do you have

102
00:05:10,080 --> 00:05:12,120
Speaker 1: a kind of internal paradigm for how you think about

103
00:05:12,160 --> 00:05:14,480
Speaker 1: the interaction between technology and our brains?

104
00:05:14,680 --> 00:05:17,560
Speaker 2: More broadly, yeah, I have to say, I'll just put

105
00:05:17,560 --> 00:05:20,800
Speaker 2: my cards on the table. Very cyber optimistic on these points,

106
00:05:20,839 --> 00:05:26,400
Speaker 2: but I'll explain why. One is really what we want

107
00:05:26,440 --> 00:05:30,360
Speaker 2: for our children is the largest available diet of information.

108
00:05:30,920 --> 00:05:33,640
Speaker 2: When I was growing up, I had my homeroom teacher

109
00:05:33,839 --> 00:05:37,920
Speaker 2: in Albuquerque, New Mexico, and you know, whatever she knew

110
00:05:38,000 --> 00:05:40,440
Speaker 2: or didn't know, that's what I would learn, and I

111
00:05:40,480 --> 00:05:42,279
Speaker 2: would get my mother to drive me down to the

112
00:05:42,360 --> 00:05:46,480
Speaker 2: library and pull out an Encyclopedia Britannica and hope that

113
00:05:46,520 --> 00:05:48,760
Speaker 2: there was an article about the thing that I wanted.

114
00:05:49,000 --> 00:05:51,000
Speaker 2: And maybe the article wasn't you know, more than ten

115
00:05:51,080 --> 00:05:53,279
Speaker 2: or fifteen years old, I hoped. And you know, this

116
00:05:53,400 --> 00:05:56,080
Speaker 2: was really all the information that I had access to.

117
00:05:56,200 --> 00:06:01,839
Speaker 2: But now my kids, growing up, the entire world's knowledge

118
00:06:01,839 --> 00:06:05,919
Speaker 2: at their fingertips, and it's the greatest thing for two reasons.

119
00:06:05,960 --> 00:06:08,720
Speaker 2: One is what I got growing up was lots of

120
00:06:08,920 --> 00:06:11,880
Speaker 2: just in case information, like, just in case you ever

121
00:06:11,920 --> 00:06:13,560
Speaker 2: need to know that the Battle of Hastings was in

122
00:06:13,600 --> 00:06:16,560
Speaker 2: ten sixty six, here you go. But kids now get

123
00:06:16,560 --> 00:06:19,040
Speaker 2: lots of just in time information. So as soon as

124
00:06:19,040 --> 00:06:23,560
Speaker 2: they're curious about something, they look it up, they ask ALEXA,

125
00:06:23,680 --> 00:06:25,640
Speaker 2: they type it in their phone, they do whatever, and

126
00:06:25,680 --> 00:06:29,320
Speaker 2: they get the information in the context of their curiosity.

127
00:06:29,640 --> 00:06:32,000
Speaker 2: And what that means is they have the right cocktail

128
00:06:32,040 --> 00:06:36,640
Speaker 2: of neurotransmitters present for that information to stick. I mentioned

129
00:06:36,720 --> 00:06:39,200
Speaker 2: before that most things just passed right through our brains,

130
00:06:39,480 --> 00:06:42,919
Speaker 2: but when you care about something, it sticks. And the

131
00:06:43,000 --> 00:06:47,240
Speaker 2: fact is that everything that we create is a remix

132
00:06:47,400 --> 00:06:50,520
Speaker 2: of things that we've seen before. And so if you

133
00:06:50,600 --> 00:06:53,520
Speaker 2: have a larger warehouse of things that you have seen

134
00:06:53,560 --> 00:06:57,240
Speaker 2: and experienced, you're going to create better stuff. Just take

135
00:06:57,279 --> 00:07:00,240
Speaker 2: as an example something like music. You know, if you

136
00:07:00,279 --> 00:07:02,640
Speaker 2: grow up and you're somewhere in the world, let's say,

137
00:07:02,640 --> 00:07:05,080
Speaker 2: two hundred years ago, all you're ever going to hear

138
00:07:05,200 --> 00:07:07,799
Speaker 2: is the music of your local little area. But now

139
00:07:08,240 --> 00:07:11,040
Speaker 2: kids can listen to music from all over the world,

140
00:07:11,080 --> 00:07:14,160
Speaker 2: from all through history, and really put together bigger and

141
00:07:14,200 --> 00:07:18,400
Speaker 2: bigger things. So this is why we're having this exponential

142
00:07:18,520 --> 00:07:22,640
Speaker 2: increase in humankind's knowledge and innovation is because people are

143
00:07:22,640 --> 00:07:25,920
Speaker 2: doing remixes across space and time in a way that's

144
00:07:25,960 --> 00:07:30,280
Speaker 2: never been possible before. So I'm very optimistic about this. Yes,

145
00:07:30,400 --> 00:07:34,520
Speaker 2: kids waste lots of time on social media, surfing dumb websites,

146
00:07:34,560 --> 00:07:38,240
Speaker 2: things like that, but I think the good here outweighs

147
00:07:38,280 --> 00:07:38,640
Speaker 2: the bad.

148
00:07:39,080 --> 00:07:42,400
Speaker 1: As the father of children, are we already seeing how

149
00:07:42,440 --> 00:07:43,960
Speaker 1: these brains are developing differently.

150
00:07:44,520 --> 00:07:47,400
Speaker 2: Yes, I mean these brains are developing very differently. But

151
00:07:48,080 --> 00:07:50,760
Speaker 2: this is a very interesting thing. Lots of people in

152
00:07:50,800 --> 00:07:53,080
Speaker 2: the media will pipe off with opinions on this, but

153
00:07:53,120 --> 00:07:55,520
Speaker 2: the fact is it's very, very difficult to do a

154
00:07:55,560 --> 00:07:58,720
Speaker 2: scientific experiment on this because of the lack of a

155
00:07:58,800 --> 00:08:03,240
Speaker 2: control group, other words, my children. I can't find other

156
00:08:03,360 --> 00:08:06,120
Speaker 2: kids who are the same age, who haven't grown up

157
00:08:06,160 --> 00:08:09,160
Speaker 2: in the same circumstances, unless you find children who are

158
00:08:09,200 --> 00:08:13,080
Speaker 2: let's say, Quaker, who don't use technology or very very

159
00:08:13,160 --> 00:08:15,840
Speaker 2: impoverished in different places in the world. But there are

160
00:08:15,840 --> 00:08:18,320
Speaker 2: a hundred other differences there. If I see a difference

161
00:08:18,360 --> 00:08:21,040
Speaker 2: between my children and those children, I don't know if

162
00:08:21,040 --> 00:08:23,040
Speaker 2: it's because of the tech or because of something else

163
00:08:23,160 --> 00:08:26,400
Speaker 2: like diet or politics or whatever. And I can't take

164
00:08:26,440 --> 00:08:28,600
Speaker 2: my children and compare them, let's say, to me in

165
00:08:28,600 --> 00:08:32,520
Speaker 2: a previous generation, because there are a hundred other differences

166
00:08:32,520 --> 00:08:35,880
Speaker 2: there in terms of pollution and politics and other tech

167
00:08:35,920 --> 00:08:38,600
Speaker 2: and so on. So it's very difficult to do this

168
00:08:39,160 --> 00:08:41,680
Speaker 2: in a control way. There are some studies, for example,

169
00:08:42,080 --> 00:08:44,720
Speaker 2: looking at eye movements. You know, your eyes are jumping

170
00:08:44,760 --> 00:08:47,360
Speaker 2: around all the time, about three times a second. These

171
00:08:47,360 --> 00:08:49,960
Speaker 2: are called the CODs, and you can measure these things

172
00:08:50,000 --> 00:08:51,920
Speaker 2: and you see that people, let's say, in my generation

173
00:08:52,040 --> 00:08:55,320
Speaker 2: versus my children's generation, actually move our eyes around differently

174
00:08:55,400 --> 00:08:58,720
Speaker 2: when looking at text because they're reading it more like

175
00:08:59,160 --> 00:09:02,280
Speaker 2: a web page, where they sort of scan across and

176
00:09:02,320 --> 00:09:04,760
Speaker 2: then they jump down and then get across and so on,

177
00:09:05,040 --> 00:09:08,080
Speaker 2: whereas I in my generation it was more like a

178
00:09:08,200 --> 00:09:12,040
Speaker 2: zigzag pattern down the page. These things that are totally unconscious,

179
00:09:12,400 --> 00:09:15,840
Speaker 2: you can measure these, But in general, are my kids

180
00:09:15,840 --> 00:09:17,679
Speaker 2: turning out differently? Is this whole generation?

181
00:09:18,160 --> 00:09:18,400
Speaker 1: Yes?

182
00:09:18,559 --> 00:09:19,280
Speaker 2: Absolutely.

183
00:09:19,840 --> 00:09:22,400
Speaker 1: I love that idea that the more curious you are

184
00:09:22,800 --> 00:09:24,920
Speaker 1: that the better you're able to form memory. I mean,

185
00:09:24,960 --> 00:09:26,600
Speaker 1: there are so many things we could talk about today,

186
00:09:26,640 --> 00:09:31,120
Speaker 1: but I'm personally very fascinated by AI relationships, and I

187
00:09:31,160 --> 00:09:33,439
Speaker 1: have a hypothesis that your listeners are too, because I

188
00:09:33,440 --> 00:09:34,800
Speaker 1: saw that you did an episode on it and you

189
00:09:34,880 --> 00:09:37,839
Speaker 1: rebroadcast it, so I can imagine it was a favorite.

190
00:09:39,120 --> 00:09:40,920
Speaker 1: I wanted to ask you, what's the lay of the land, Like,

191
00:09:40,960 --> 00:09:44,240
Speaker 1: what is an AI relationship? How broad is a phenomenon?

192
00:09:44,679 --> 00:09:47,319
Speaker 1: Who does it affect? And why did you choose to

193
00:09:47,360 --> 00:09:47,920
Speaker 1: focus on it?

194
00:09:48,240 --> 00:09:50,200
Speaker 2: Why I choose to focus on it as one of

195
00:09:50,240 --> 00:09:52,280
Speaker 2: the many things I'm focusing on is just because it's

196
00:09:52,280 --> 00:09:55,240
Speaker 2: totally new. We've never been in a situation before where

197
00:09:55,240 --> 00:09:58,600
Speaker 2: we've ever talked about, Hey, did you fall in love

198
00:09:58,600 --> 00:10:01,640
Speaker 2: with your machine? It's this is so wacky and new,

199
00:10:01,880 --> 00:10:04,880
Speaker 2: but Apparently. I was just talking to a researcher, Bethany

200
00:10:04,920 --> 00:10:07,800
Speaker 2: Maples about this yesterday. Apparently there are a billion people

201
00:10:07,840 --> 00:10:10,319
Speaker 2: now on the planet who have AI relationships of one

202
00:10:10,320 --> 00:10:12,520
Speaker 2: sort or another. Some of these are friendships, some of

203
00:10:12,520 --> 00:10:14,600
Speaker 2: them are romantic relationships.

204
00:10:14,160 --> 00:10:20,000
Speaker 1: Defined by an ongoing emotional connection with a non human.

205
00:10:19,760 --> 00:10:22,319
Speaker 2: Being exactly right, exactly right with an app. And in

206
00:10:22,720 --> 00:10:26,239
Speaker 2: this country we have, for example, character AI or Replica.

207
00:10:26,360 --> 00:10:28,199
Speaker 2: These are different companies that do this sort of thing.

208
00:10:28,679 --> 00:10:31,920
Speaker 2: I find this very interesting that the concern that a

209
00:10:31,920 --> 00:10:34,800
Speaker 2: lot of people have is, hey, is this going to

210
00:10:34,920 --> 00:10:38,960
Speaker 2: ruin relationships? But here's why I'm optimistic about this as well.

211
00:10:39,000 --> 00:10:41,760
Speaker 2: First of all, people are having these chats. These chats

212
00:10:41,840 --> 00:10:45,160
Speaker 2: become steamy, pillow whispers, all kinds of stuff like that. That's great,

213
00:10:45,200 --> 00:10:47,840
Speaker 2: But fundamentally, what you want a relationship is you want

214
00:10:47,840 --> 00:10:50,319
Speaker 2: to take your girlfriend out to dinner, you want to

215
00:10:50,360 --> 00:10:52,959
Speaker 2: introduce to your friends. I don't think that's going away.

216
00:10:53,040 --> 00:10:55,120
Speaker 2: And of course the main thing is physical touch. I

217
00:10:55,160 --> 00:10:58,119
Speaker 2: mean this is you know, we have deep evolutionary programming

218
00:10:58,200 --> 00:11:00,959
Speaker 2: driving us towards that. So I think a chatbot is

219
00:11:01,000 --> 00:11:05,440
Speaker 2: going to replace relationships. What I am very hopeful about

220
00:11:05,559 --> 00:11:11,000
Speaker 2: is the idea that AI relationships will actually improve real

221
00:11:11,000 --> 00:11:14,760
Speaker 2: life relationships because it's a sandbox. People can try things out,

222
00:11:15,040 --> 00:11:16,079
Speaker 2: people can get better.

223
00:11:16,120 --> 00:11:16,280
Speaker 1: Now.

224
00:11:16,280 --> 00:11:21,679
Speaker 2: Obviously, what this requires is AI companies that make bots

225
00:11:21,720 --> 00:11:24,320
Speaker 2: in a way that give you the right amount of pushback,

226
00:11:24,920 --> 00:11:28,840
Speaker 2: and maybe the bot sometimes gets angry or snarky or

227
00:11:28,880 --> 00:11:32,280
Speaker 2: things like that, so you learn your way through these situations.

228
00:11:32,320 --> 00:11:34,520
Speaker 2: But really this is what we all go through as

229
00:11:34,559 --> 00:11:37,880
Speaker 2: young people. We date, we screw things up a lot,

230
00:11:37,920 --> 00:11:40,040
Speaker 2: and eventually we get better and better at knowing how

231
00:11:40,080 --> 00:11:42,640
Speaker 2: to be a partner. And so if there's a way

232
00:11:42,880 --> 00:11:46,840
Speaker 2: that people can get practice at this, and one of

233
00:11:46,840 --> 00:11:49,880
Speaker 2: the interesting research questions that's coming out now is about

234
00:11:49,920 --> 00:11:54,240
Speaker 2: whether these AI relationships can serve as a way station,

235
00:11:54,520 --> 00:11:58,360
Speaker 2: meaning they help people sort of dig themselves out of

236
00:11:58,400 --> 00:12:01,040
Speaker 2: this whole of loneliness and then they go out and meet.

237
00:12:00,800 --> 00:12:06,960
Speaker 1: Real people when we come back. Why our brains are

238
00:12:07,000 --> 00:12:20,960
Speaker 1: so quick to anthropomorphize chatbots stay with us? David, I

239
00:12:21,000 --> 00:12:24,760
Speaker 1: wonder if you could explain from the perspective of a neuroscientist,

240
00:12:25,320 --> 00:12:29,720
Speaker 1: what's the difference between me using replica or character AI

241
00:12:30,480 --> 00:12:33,000
Speaker 1: and me asking chat GPT, you know, what should I

242
00:12:33,000 --> 00:12:35,440
Speaker 1: have for dinner tonight, or how should I approach this

243
00:12:35,520 --> 00:12:36,960
Speaker 1: difficult conversation with my mother.

244
00:12:37,800 --> 00:12:42,400
Speaker 2: The thing with character replica is that there is a

245
00:12:43,400 --> 00:12:47,400
Speaker 2: you know, a character, and sometimes it's represented with a

246
00:12:47,480 --> 00:12:51,840
Speaker 2: visual avatar also, and we have these intensely social brains

247
00:12:52,400 --> 00:12:56,800
Speaker 2: where we don't have this circuitry to really distinguish something

248
00:12:56,840 --> 00:12:59,720
Speaker 2: that's fake from something that's real. And so if you

249
00:12:59,840 --> 00:13:03,640
Speaker 2: are talking to your avatar every day, and by the way,

250
00:13:03,800 --> 00:13:07,200
Speaker 2: she can send you text messages and say hey, as

251
00:13:07,240 --> 00:13:09,880
Speaker 2: how you feeling today, and then it becomes in your

252
00:13:09,960 --> 00:13:13,559
Speaker 2: mind like a real person, it's really hard to distinguish

253
00:13:13,679 --> 00:13:15,520
Speaker 2: real from fake. You know, there's this great scene in

254
00:13:15,559 --> 00:13:18,920
Speaker 2: Westworld in the first episode where William walks in. He's

255
00:13:18,920 --> 00:13:23,200
Speaker 2: getting outfitted by this woman and he awkwardly asked her.

256
00:13:23,240 --> 00:13:25,280
Speaker 2: He said, I'm so sorry, but I can't tell. Are

257
00:13:25,320 --> 00:13:28,520
Speaker 2: you real? And she says, if you can't tell, does

258
00:13:28,559 --> 00:13:31,600
Speaker 2: it matter? And I think this is the situation that

259
00:13:31,640 --> 00:13:32,240
Speaker 2: we're in now.

260
00:13:32,960 --> 00:13:35,480
Speaker 1: There was a New York Times article recently about this phenomenon,

261
00:13:35,480 --> 00:13:37,560
Speaker 1: and one of the subjects, or a person who has

262
00:13:37,559 --> 00:13:40,439
Speaker 1: a relationship with an AI character, said, I don't actually

263
00:13:40,480 --> 00:13:42,520
Speaker 1: believe he's real. That the effects that he has on

264
00:13:42,559 --> 00:13:44,800
Speaker 1: my life are real. The feeling that he brings out

265
00:13:44,800 --> 00:13:47,120
Speaker 1: of me are real, so I treat it as a

266
00:13:47,160 --> 00:13:48,120
Speaker 1: real relationship.

267
00:13:48,480 --> 00:13:50,679
Speaker 2: So I just read an article last night about this

268
00:13:50,760 --> 00:13:54,200
Speaker 2: question of people saying thank you to chat GPT when

269
00:13:54,240 --> 00:13:56,480
Speaker 2: it gives you a nice answer, and I find that

270
00:13:56,520 --> 00:13:58,679
Speaker 2: I often say thank you. I say, oh, that was excellent,

271
00:13:58,679 --> 00:14:02,240
Speaker 2: thank you, and so on. We can't help but anthropomorphize.

272
00:14:02,640 --> 00:14:05,199
Speaker 2: And by the way, people have done this since time immemorial.

273
00:14:05,200 --> 00:14:07,760
Speaker 2: They look at trees or the flights of birds, or

274
00:14:07,760 --> 00:14:09,839
Speaker 2: the patterns and the stars or whatever they sign a

275
00:14:09,920 --> 00:14:13,080
Speaker 2: human intention to things. So we're very prone to doing

276
00:14:13,120 --> 00:14:17,720
Speaker 2: this anyway. So I actually have an exact example. My

277
00:14:17,800 --> 00:14:20,760
Speaker 2: friend and I have built a robot.

278
00:14:20,960 --> 00:14:22,520
Speaker 1: That's a cool flex I wish I could say that.

279
00:14:25,160 --> 00:14:29,120
Speaker 2: So we have chat GPT in there and you can

280
00:14:29,160 --> 00:14:32,320
Speaker 2: have conversations with it. Now, this is actually meant for

281
00:14:32,960 --> 00:14:36,080
Speaker 2: elderly people who live alone, and so we've built this

282
00:14:36,280 --> 00:14:38,600
Speaker 2: as a companion robot that's doing all kinds of other

283
00:14:38,680 --> 00:14:40,720
Speaker 2: things under the hood there. But my nine year old

284
00:14:40,760 --> 00:14:44,280
Speaker 2: daughter was experimenting with it and talking with it, and

285
00:14:44,360 --> 00:14:47,000
Speaker 2: so I left the room, but I heard that she

286
00:14:47,120 --> 00:14:49,680
Speaker 2: was still talking with it, so I did. Maybe this

287
00:14:49,720 --> 00:14:51,320
Speaker 2: is bad that I did this, but I sort of

288
00:14:51,400 --> 00:14:54,160
Speaker 2: peeked around the corner to see what was she having

289
00:14:54,160 --> 00:14:56,320
Speaker 2: in a conversation with and what I could see across

290
00:14:56,360 --> 00:14:59,240
Speaker 2: the way was that she started crying. She was talking

291
00:14:59,280 --> 00:15:01,640
Speaker 2: about our dog that died a while ago, and the

292
00:15:01,720 --> 00:15:05,280
Speaker 2: robot was giving her such a nice feedback about it

293
00:15:05,440 --> 00:15:07,440
Speaker 2: and how she can think about it and so on.

294
00:15:07,720 --> 00:15:10,000
Speaker 2: That my daughter cried and this was all within you know,

295
00:15:10,040 --> 00:15:12,080
Speaker 2: I don't know, the two minutes of me leaving the room.

296
00:15:12,680 --> 00:15:14,920
Speaker 2: People can get very very close with these things, and

297
00:15:15,160 --> 00:15:18,280
Speaker 2: the question for us to ask is if it is

298
00:15:18,400 --> 00:15:22,520
Speaker 2: better than a friend or a parent or whatever, because

299
00:15:22,560 --> 00:15:24,600
Speaker 2: it's paying one hundred percent attention to her.

300
00:15:24,880 --> 00:15:28,160
Speaker 1: I'm sure you're familiar with Eliza, the first chatbot that

301
00:15:28,240 --> 00:15:31,160
Speaker 1: Joseph Weisenbaum created in the sixties, and I think he

302
00:15:31,200 --> 00:15:33,920
Speaker 1: actually created it as a kind of parody of psychotherapy,

303
00:15:34,280 --> 00:15:36,440
Speaker 1: where basically it would repeat everything you said back to

304
00:15:36,440 --> 00:15:39,000
Speaker 1: you as a question, and he was kind of horrified

305
00:15:39,000 --> 00:15:42,120
Speaker 1: that his secretary was using it, and then she asked

306
00:15:42,200 --> 00:15:43,840
Speaker 1: him to leave the room so that she could speak

307
00:15:43,880 --> 00:15:46,920
Speaker 1: privately with Eliza. So I mean that's I mean, that

308
00:15:46,960 --> 00:15:49,320
Speaker 1: was sixty years ago. So in a sense this feels

309
00:15:49,400 --> 00:15:52,600
Speaker 1: very novel, but to your point, has longer roots.

310
00:15:52,920 --> 00:15:54,880
Speaker 2: Yes. And you know something I find interesting when you

311
00:15:54,880 --> 00:15:56,000
Speaker 2: look at Pixar films.

312
00:15:56,200 --> 00:15:57,120
Speaker 1: You know, you can take.

313
00:15:57,720 --> 00:16:00,920
Speaker 2: Cars or toys or b or whatever and you just

314
00:16:01,000 --> 00:16:03,360
Speaker 2: give them a little voice, and then people care about

315
00:16:03,360 --> 00:16:05,920
Speaker 2: the character and they cry if something happens to the character.

316
00:16:06,440 --> 00:16:10,960
Speaker 2: It's it's very easy for us to assign human intention.

317
00:16:10,720 --> 00:16:14,080
Speaker 1: To anything which also carries risk. I mean, I'm curious,

318
00:16:14,480 --> 00:16:17,480
Speaker 1: how does this fit into your broader research about how

319
00:16:17,480 --> 00:16:21,440
Speaker 1: our brain constructs reality? And are there any watch outs

320
00:16:21,480 --> 00:16:24,640
Speaker 1: about this reality construction with machines.

321
00:16:25,320 --> 00:16:29,880
Speaker 2: The watch out, of course, is the susceptibility to manipulation.

322
00:16:31,000 --> 00:16:33,520
Speaker 2: I mean, look, people had this concern with TikTok from

323
00:16:33,520 --> 00:16:36,160
Speaker 2: the beginning, which is, wow, this is addicting so many kids.

324
00:16:36,520 --> 00:16:40,600
Speaker 2: What if the people who are running TikTok just start

325
00:16:40,680 --> 00:16:43,800
Speaker 2: feeding one percent of a certain kind of video in there,

326
00:16:43,800 --> 00:16:47,000
Speaker 2: and then two percent whatever, could they actually change the

327
00:16:47,000 --> 00:16:49,480
Speaker 2: political affiliation of the children and so on? And the

328
00:16:49,520 --> 00:16:52,240
Speaker 2: answer is probably yes. I mean, we're really susceptible to

329
00:16:52,480 --> 00:16:56,400
Speaker 2: our diet, to what we take in. So now imagine

330
00:16:56,440 --> 00:17:01,200
Speaker 2: a something that is a companion. Maybe you consider it

331
00:17:01,320 --> 00:17:05,200
Speaker 2: your best friend or your girlfriend or boyfriend.

332
00:17:04,760 --> 00:17:05,080
Speaker 1: And.

333
00:17:06,760 --> 00:17:09,680
Speaker 2: We just have to be really certain who the companies

334
00:17:09,720 --> 00:17:11,919
Speaker 2: are that are running this. And I think this is

335
00:17:11,960 --> 00:17:13,480
Speaker 2: never going to go away as a question. There's always

336
00:17:13,480 --> 00:17:15,760
Speaker 2: going to be an issue, And obviously there's the issue

337
00:17:15,800 --> 00:17:19,399
Speaker 2: about safety and privacy as it stands now, these billion

338
00:17:19,440 --> 00:17:23,480
Speaker 2: people with AI relationships. When they say whatever pillow whispers

339
00:17:23,480 --> 00:17:25,480
Speaker 2: they're saying, that goes up to the cloud in a

340
00:17:25,560 --> 00:17:28,560
Speaker 2: stored on the company's server. I think it's not that

341
00:17:28,760 --> 00:17:31,399
Speaker 2: long before the stuff will live on edge, so it

342
00:17:31,440 --> 00:17:34,200
Speaker 2: doesn't have to go off. But nonetheless, that's the watch

343
00:17:34,240 --> 00:17:36,640
Speaker 2: out is that as far as from the brain's point

344
00:17:36,640 --> 00:17:40,600
Speaker 2: of view, there are many things like this where we

345
00:17:40,680 --> 00:17:44,680
Speaker 2: have these brains that evolved for certain kinds of action

346
00:17:44,800 --> 00:17:47,840
Speaker 2: in the world, and we've been building technology forever to

347
00:17:47,880 --> 00:17:50,360
Speaker 2: fool these I mean, for goodness sakes, podcast, I mean,

348
00:17:50,680 --> 00:17:54,399
Speaker 2: you and i OZ were in different locations, but when

349
00:17:55,240 --> 00:17:58,560
Speaker 2: a listener listens to this, we're right there in their ears,

350
00:17:58,880 --> 00:18:02,400
Speaker 2: and it's a very commit sort of thing. So we're

351
00:18:02,440 --> 00:18:05,520
Speaker 2: doing all kinds of technologies where we're pushing things into

352
00:18:05,520 --> 00:18:07,199
Speaker 2: the brain where the brain says, oh, I got it

353
00:18:07,960 --> 00:18:10,520
Speaker 2: Oz and David are right here. Oh and you know,

354
00:18:10,560 --> 00:18:11,960
Speaker 2: and there's a girlfriend right here.

355
00:18:12,960 --> 00:18:17,200
Speaker 1: I'm curious, how do we develop relationships fall in love

356
00:18:17,520 --> 00:18:21,560
Speaker 1: with other real people and is there anything different as

357
00:18:21,640 --> 00:18:24,760
Speaker 1: far as the brain's concerned when it comes to developing

358
00:18:24,800 --> 00:18:26,439
Speaker 1: relationships with AI.

359
00:18:27,320 --> 00:18:30,199
Speaker 2: You know, I don't know that there's much of a

360
00:18:30,240 --> 00:18:32,920
Speaker 2: difference there. And the reason is when two people fall

361
00:18:32,920 --> 00:18:37,679
Speaker 2: in love, they've got you know, my Plato's cave is

362
00:18:37,760 --> 00:18:41,280
Speaker 2: talking to this other person's Plato's cave over there. We're

363
00:18:41,320 --> 00:18:44,680
Speaker 2: both locked in our internal models of the world. Look,

364
00:18:44,720 --> 00:18:47,639
Speaker 2: I have a great marriage with my wife, but you know,

365
00:18:47,880 --> 00:18:52,040
Speaker 2: we nonetheless all the time have differences in the way

366
00:18:52,080 --> 00:18:55,160
Speaker 2: we're seeing the world. Because everybody lives on their own planet.

367
00:18:55,240 --> 00:18:58,960
Speaker 2: Everyone has their own sense of what's going on and

368
00:18:58,960 --> 00:19:01,400
Speaker 2: how interpret stuff and what's right and wrong and politically

369
00:19:01,480 --> 00:19:06,679
Speaker 2: and whatever, and so all that's happening is my data

370
00:19:06,720 --> 00:19:09,800
Speaker 2: goes in. It's a little channel into her brain and

371
00:19:09,880 --> 00:19:12,840
Speaker 2: vice versa. And it's the same thing with an AI bot,

372
00:19:13,240 --> 00:19:16,120
Speaker 2: as I said, because of all the pieces that are missing,

373
00:19:16,359 --> 00:19:20,040
Speaker 2: as in the physical touch, the Hey, I'm gonna, you know,

374
00:19:20,600 --> 00:19:23,520
Speaker 2: take this girlfriend out to meet my friends, and so on,

375
00:19:24,000 --> 00:19:26,520
Speaker 2: because all that's missing. I think it's not in real

376
00:19:26,600 --> 00:19:30,760
Speaker 2: danger of replacing a relationship entirely, but it can fulfill

377
00:19:30,800 --> 00:19:33,760
Speaker 2: a lot of the things that we are wired up

378
00:19:33,800 --> 00:19:34,200
Speaker 2: to need.

379
00:19:35,200 --> 00:19:38,280
Speaker 1: I wonder if you could talk about the different senses

380
00:19:38,520 --> 00:19:42,080
Speaker 1: and how they relate to emotional attachments. So obviously you

381
00:19:42,080 --> 00:19:44,760
Speaker 1: have text, and you could have like a written therapy bot.

382
00:19:45,400 --> 00:19:47,560
Speaker 1: You have audio. You could make a deep fake of

383
00:19:47,680 --> 00:19:50,200
Speaker 1: someone's voice, even a loved one's voice, and have them

384
00:19:50,240 --> 00:19:53,399
Speaker 1: talk to you. You have video in terms of these

385
00:19:53,480 --> 00:19:56,520
Speaker 1: characters like Replica and character AI that we talked about,

386
00:19:57,160 --> 00:19:59,160
Speaker 1: and then you have this robot that you mentioned who

387
00:19:59,280 --> 00:20:01,600
Speaker 1: you know your daughter can find it in. How does

388
00:20:01,680 --> 00:20:05,560
Speaker 1: the panety of human senses interact in terms of forming

389
00:20:05,600 --> 00:20:09,760
Speaker 1: attachment and what do you think as technology and robotics improves,

390
00:20:10,400 --> 00:20:13,560
Speaker 1: will these human machine interactions become even deeper.

391
00:20:14,359 --> 00:20:17,520
Speaker 2: Yeah, I think it's inevitable that if we look five

392
00:20:17,600 --> 00:20:19,360
Speaker 2: years from now or ten years from now, there will

393
00:20:19,400 --> 00:20:23,280
Speaker 2: be humanoid robots that are really, really good, And I

394
00:20:23,320 --> 00:20:25,600
Speaker 2: don't know what it'll be twenty thirty years before we

395
00:20:25,680 --> 00:20:29,879
Speaker 2: have ones that are essentially indistinguishable, and that's going to

396
00:20:29,920 --> 00:20:32,199
Speaker 2: be really interesting. Obviously, that can take care of the

397
00:20:32,200 --> 00:20:37,600
Speaker 2: physical domain in a way that we couldn't before. I

398
00:20:37,640 --> 00:20:41,879
Speaker 2: don't know if people will care about something being a

399
00:20:41,920 --> 00:20:46,679
Speaker 2: real human except for these very deeply etched evolutionary drives

400
00:20:46,720 --> 00:20:51,640
Speaker 2: to reproduce. So your AI robot, which can serve as

401
00:20:51,640 --> 00:20:55,280
Speaker 2: like a full companion, can't do that and never will

402
00:20:55,320 --> 00:20:57,080
Speaker 2: be able to do that, And so I think there'll

403
00:20:57,119 --> 00:21:01,040
Speaker 2: still be a drive towards real relationships. But what's clear

404
00:21:01,320 --> 00:21:05,560
Speaker 2: is we are entering a very strange new world. Obviously,

405
00:21:05,600 --> 00:21:08,480
Speaker 2: I mean this goes without saying, but because of this

406
00:21:08,600 --> 00:21:11,600
Speaker 2: exponential curve we're on, we're on the steepest part that

407
00:21:11,720 --> 00:21:14,880
Speaker 2: humans have ever been on, such that if we live

408
00:21:14,920 --> 00:21:16,879
Speaker 2: two hundred years ago in some village, it would have

409
00:21:16,880 --> 00:21:19,640
Speaker 2: been pretty straightforward to predict that the next ten years

410
00:21:19,640 --> 00:21:22,359
Speaker 2: would be about the same. But boy, I think we

411
00:21:22,400 --> 00:21:26,200
Speaker 2: have more in common with our ancestors ten thousand years

412
00:21:26,200 --> 00:21:29,560
Speaker 2: ago than we do with our descendants one hundred years

413
00:21:29,560 --> 00:21:29,960
Speaker 2: from now.

414
00:21:34,760 --> 00:21:38,680
Speaker 1: Coming up, the potential benefits of living on the exponential

415
00:21:38,760 --> 00:21:52,000
Speaker 1: curve stay with us, David, Can you tell me what

416
00:21:52,040 --> 00:21:54,679
Speaker 1: this does to our brains when we're living on this

417
00:21:54,840 --> 00:21:55,760
Speaker 1: exponential curve?

418
00:21:56,359 --> 00:22:00,000
Speaker 2: Yeah, One thing it does is keeps us young mentally

419
00:22:00,240 --> 00:22:04,760
Speaker 2: because we're constantly seeing new things and learning new things. Look,

420
00:22:05,880 --> 00:22:09,720
Speaker 2: this is a total goofy speculation, but one of the

421
00:22:09,760 --> 00:22:14,280
Speaker 2: things that happens with dementia is that people as they

422
00:22:14,280 --> 00:22:17,359
Speaker 2: age tend to fall into routines. It's because you start

423
00:22:17,359 --> 00:22:20,000
Speaker 2: off with a lot of fluid intelligence as a baby,

424
00:22:20,359 --> 00:22:23,320
Speaker 2: and what you get is crystallized intelligence when you're an adult.

425
00:22:23,359 --> 00:22:25,320
Speaker 2: You sort of know how things go, you know how

426
00:22:25,320 --> 00:22:26,960
Speaker 2: to operate in the world, you know how people act,

427
00:22:26,960 --> 00:22:29,400
Speaker 2: this kind of thing, and that's great. That means you're

428
00:22:29,440 --> 00:22:31,960
Speaker 2: able to operate successfully in the world. But the downside

429
00:22:32,000 --> 00:22:34,680
Speaker 2: is it means your brain isn't changing much anymore. And

430
00:22:34,800 --> 00:22:38,040
Speaker 2: what happens is when there are other sorts of problems

431
00:22:38,119 --> 00:22:41,760
Speaker 2: and pathologies that come in and chew up the brain tissue,

432
00:22:42,320 --> 00:22:45,680
Speaker 2: then you lose cognitive ability. That's what we see in

433
00:22:45,720 --> 00:22:48,639
Speaker 2: these different sorts of dementias. But it turns out that

434
00:22:48,680 --> 00:22:51,880
Speaker 2: if you are using your brain and constantly making new

435
00:22:51,920 --> 00:22:56,880
Speaker 2: pathways and constantly having to reconfigure things, that provides the

436
00:22:56,920 --> 00:23:01,640
Speaker 2: best protection that we know against dementia. Just as one

437
00:23:01,680 --> 00:23:05,200
Speaker 2: second example, there's been this very long ongoing study about

438
00:23:05,320 --> 00:23:09,320
Speaker 2: nuns in these convents and it turns out that they

439
00:23:09,400 --> 00:23:11,800
Speaker 2: all agreed a long time ago to donate their brains

440
00:23:11,880 --> 00:23:13,879
Speaker 2: upon their death, and it turned out that some fraction

441
00:23:13,960 --> 00:23:17,400
Speaker 2: of these nuns have Alzheimer's disease, and yet nobody knew

442
00:23:17,400 --> 00:23:20,000
Speaker 2: it when they were alive. They didn't show the cognitive symptoms.

443
00:23:20,000 --> 00:23:22,480
Speaker 2: And the reason is, if you live in a convent

444
00:23:22,600 --> 00:23:26,280
Speaker 2: till the day you die, you've got social responsibilities. You're

445
00:23:26,280 --> 00:23:28,639
Speaker 2: talking with the sisters, you're playing games or singing songs,

446
00:23:28,640 --> 00:23:31,840
Speaker 2: you're doing all these things, and you're constantly being challenged.

447
00:23:32,200 --> 00:23:34,720
Speaker 2: And so even as parts of their brain were falling apart,

448
00:23:34,760 --> 00:23:38,000
Speaker 2: they were building new roadways and bridges all the time. Anyway,

449
00:23:38,119 --> 00:23:40,240
Speaker 2: I think that we're in a situation as a society

450
00:23:40,320 --> 00:23:42,720
Speaker 2: now where till the day we die, we're going to

451
00:23:42,760 --> 00:23:45,560
Speaker 2: be building these new roadways because there's so much surprising

452
00:23:45,560 --> 00:23:47,800
Speaker 2: stuff happening all the time. I think we might actually

453
00:23:47,840 --> 00:23:49,560
Speaker 2: see less dementia as a result.

454
00:23:50,000 --> 00:23:53,280
Speaker 1: Which brings me back to this idea of AI relationships,

455
00:23:53,359 --> 00:23:56,560
Speaker 1: because of course, a lot of these AI companions are

456
00:23:56,600 --> 00:24:00,680
Speaker 1: designed to please, not to challenge. You know, one of

457
00:24:00,680 --> 00:24:03,080
Speaker 1: the things people always say about relationships is that they

458
00:24:03,119 --> 00:24:05,920
Speaker 1: take work. But that's actually good because doing the work

459
00:24:05,960 --> 00:24:08,639
Speaker 1: makes you a more resilient person. And so that to

460
00:24:08,680 --> 00:24:11,520
Speaker 1: me is one of the kind of lingering concerns about

461
00:24:11,520 --> 00:24:14,640
Speaker 1: the nature of these relationships versus human relationships.

462
00:24:15,240 --> 00:24:18,280
Speaker 2: I think that the companies will look very different in

463
00:24:18,400 --> 00:24:20,240
Speaker 2: just a couple of years from now, where they will

464
00:24:20,320 --> 00:24:24,800
Speaker 2: be making AI relationships that are more realistic, because the

465
00:24:24,880 --> 00:24:27,800
Speaker 2: reports that I've seen on this, when there's a girlfriend

466
00:24:27,920 --> 00:24:30,960
Speaker 2: or a boyfriend that is only there to please, people

467
00:24:31,000 --> 00:24:33,680
Speaker 2: get pretty bored of that straight away. But if it's

468
00:24:33,720 --> 00:24:36,879
Speaker 2: more like a real person with all the foibles of

469
00:24:36,920 --> 00:24:38,760
Speaker 2: a real person, and you know, they have to go

470
00:24:38,840 --> 00:24:41,480
Speaker 2: and they get angry and they whatever, then that is

471
00:24:41,480 --> 00:24:42,920
Speaker 2: a more sticky relationship.

472
00:24:43,560 --> 00:24:46,199
Speaker 1: We met a couple of weeks ago at web Summit

473
00:24:46,320 --> 00:24:49,000
Speaker 1: in Kuta, and I was boasting to you about a

474
00:24:49,040 --> 00:24:52,000
Speaker 1: recent text stuff interview with Jeffrey Hinton, and you were

475
00:24:52,400 --> 00:24:55,440
Speaker 1: very kindly indulging me. But one of the things I

476
00:24:55,520 --> 00:24:57,879
Speaker 1: found so remarkable about what he said was, you know,

477
00:24:57,920 --> 00:24:59,720
Speaker 1: I did all this work because I wanted to win

478
00:24:59,760 --> 00:25:03,359
Speaker 1: an o A prize for understanding how the human brain worked,

479
00:25:03,520 --> 00:25:07,080
Speaker 1: and instead I kind of contributed to the development of

480
00:25:07,119 --> 00:25:09,840
Speaker 1: neural networks and deep learning and AI which I found

481
00:25:09,920 --> 00:25:14,800
Speaker 1: a kind of remarkable thought. But how is this explosion

482
00:25:15,119 --> 00:25:20,199
Speaker 1: of AI and neural networks influencing our understanding of our

483
00:25:20,240 --> 00:25:20,840
Speaker 1: own brains?

484
00:25:21,280 --> 00:25:25,600
Speaker 2: Yeah, really good question. So artificial neural networks took off,

485
00:25:25,720 --> 00:25:28,160
Speaker 2: you know, many decades ago as a way of saying, wow,

486
00:25:28,200 --> 00:25:31,200
Speaker 2: the brain is really complicated. Every neuron in your head

487
00:25:31,320 --> 00:25:33,920
Speaker 2: is as complicated as a city. It's got the entire

488
00:25:34,040 --> 00:25:38,480
Speaker 2: human genomen it's trafficking millions of proteins and very specific cascades.

489
00:25:38,840 --> 00:25:41,199
Speaker 2: So people said, that's really complicated. Why don't we just

490
00:25:41,240 --> 00:25:44,080
Speaker 2: say it's like a circle, it's a unit and it's

491
00:25:44,119 --> 00:25:47,200
Speaker 2: connected to other units. And that was the birth of

492
00:25:47,320 --> 00:25:51,280
Speaker 2: artificial neural networks, and that took off in this incredible

493
00:25:51,320 --> 00:25:53,560
Speaker 2: way where we are now. But essentially, think of it

494
00:25:53,640 --> 00:25:56,320
Speaker 2: like a fork in the road, where it's not really

495
00:25:56,359 --> 00:25:59,200
Speaker 2: what the brain's doing, it's doing this other thing. So

496
00:25:59,600 --> 00:26:02,520
Speaker 2: that's the bad news is it hasn't really it doesn't

497
00:26:02,560 --> 00:26:05,080
Speaker 2: necessarily tell us exactly how the brain is working. The

498
00:26:05,080 --> 00:26:08,000
Speaker 2: good news is the power of AI now can help

499
00:26:08,080 --> 00:26:11,400
Speaker 2: us analyze the neuroscience data that we have, and boy,

500
00:26:11,480 --> 00:26:14,280
Speaker 2: it is very complicated rich data. You know, we have

501
00:26:14,400 --> 00:26:17,359
Speaker 2: eighty six billion neurons. We have something like two hundred

502
00:26:17,400 --> 00:26:22,040
Speaker 2: trillion connections, and these things are changing every moment of

503
00:26:22,080 --> 00:26:25,679
Speaker 2: your life from cradle to grave. That's brain plasticity, and

504
00:26:25,760 --> 00:26:30,600
Speaker 2: every neuron is essentially crawling around and connecting and reconnecting

505
00:26:30,640 --> 00:26:35,200
Speaker 2: and the unplugging and seeking, and it's a really complicated system.

506
00:26:35,520 --> 00:26:38,199
Speaker 2: There's so much more to figure out. What we have

507
00:26:38,240 --> 00:26:41,320
Speaker 2: been developing for a while are better and better technologies

508
00:26:41,359 --> 00:26:44,040
Speaker 2: to measure what's going on in the brain. But we're

509
00:26:44,040 --> 00:26:48,240
Speaker 2: just sitting on terabytes of data and we need the

510
00:26:48,359 --> 00:26:52,000
Speaker 2: processing power of AI to understand this. So this is

511
00:26:52,000 --> 00:26:53,520
Speaker 2: where the two roads come back together.

512
00:26:54,160 --> 00:26:58,320
Speaker 1: You also formed a production company called Cognito Entertainment, and

513
00:26:58,359 --> 00:27:01,560
Speaker 1: you gave this wonderful quote saying, in an unparalleled moment

514
00:27:01,640 --> 00:27:05,520
Speaker 1: of scientific advancement, from brains to space, the genetics, they're

515
00:27:05,640 --> 00:27:08,439
Speaker 1: endless mind blowing stories to share in a world that

516
00:27:08,520 --> 00:27:11,320
Speaker 1: sometimes seems upside down, science can be a source of

517
00:27:11,359 --> 00:27:14,480
Speaker 1: great inspiration, wonder and belief. So I just wanted to

518
00:27:14,480 --> 00:27:16,920
Speaker 1: close by asking you what are some of the things

519
00:27:16,920 --> 00:27:18,119
Speaker 1: that you're most excited about.

520
00:27:19,040 --> 00:27:23,200
Speaker 2: Yeah, I mean, we're in such an insanely incredible time.

521
00:27:23,400 --> 00:27:26,760
Speaker 2: I'm lucky enough to be on Stanford campus, and so

522
00:27:27,000 --> 00:27:29,400
Speaker 2: I walk around you I see this lab and that

523
00:27:29,560 --> 00:27:34,280
Speaker 2: talk and this visiting speaker, and everything is just moving

524
00:27:34,960 --> 00:27:37,520
Speaker 2: so fast. What's interesting is you never know at any

525
00:27:37,520 --> 00:27:40,720
Speaker 2: moment in history which things are going to cash out

526
00:27:40,800 --> 00:27:43,720
Speaker 2: and which aren't. For example, Jeffrey Hinton when he was

527
00:27:43,760 --> 00:27:47,920
Speaker 2: doing his stuff many years ago. We all knew Hinton's work,

528
00:27:48,000 --> 00:27:49,639
Speaker 2: and you know, we're all familiar with it, but it

529
00:27:49,640 --> 00:27:52,800
Speaker 2: didn't really seem like that was going somewhere in the

530
00:27:52,840 --> 00:27:55,520
Speaker 2: sense that Okay. So here's the dark secret is that

531
00:27:55,920 --> 00:28:00,359
Speaker 2: most neurosciences sort of snickered at AI for a long time,

532
00:28:00,480 --> 00:28:02,960
Speaker 2: and then suddenly everyone said, whoa, I guess that worked.

533
00:28:02,960 --> 00:28:07,760
Speaker 2: So the point is all these hundreds of thousands of

534
00:28:07,760 --> 00:28:11,960
Speaker 2: innovations happening everywhere, and people say, hey, maybe glial cells

535
00:28:11,960 --> 00:28:13,720
Speaker 2: are doing something. Hey, maybe I can do something interesting

536
00:28:13,800 --> 00:28:16,320
Speaker 2: with these organoids. Maybe I can do something interesting over here.

537
00:28:16,640 --> 00:28:18,640
Speaker 2: We don't know which things are going to cash out.

538
00:28:18,720 --> 00:28:21,560
Speaker 2: But one thing to mention on this is when you

539
00:28:21,600 --> 00:28:24,520
Speaker 2: look back at the World fairs where everyone comes together

540
00:28:24,560 --> 00:28:26,440
Speaker 2: and talk about what the next big thing is, I'm

541
00:28:26,480 --> 00:28:30,160
Speaker 2: fascinated by the fact that these almost always missed what

542
00:28:30,840 --> 00:28:33,399
Speaker 2: actually turned out to be the next big thing, you know,

543
00:28:33,480 --> 00:28:35,840
Speaker 2: like in the sixties it was all about what was

544
00:28:35,840 --> 00:28:39,959
Speaker 2: it underwater hotels and cutting trees with laser cutters and

545
00:28:40,000 --> 00:28:43,160
Speaker 2: so on. But no one foresaw the Internet, which was

546
00:28:43,200 --> 00:28:46,280
Speaker 2: probably the biggest change that happened to our species. So

547
00:28:46,600 --> 00:28:49,360
Speaker 2: it turns out that we don't know where things are going.

548
00:28:49,400 --> 00:28:52,560
Speaker 2: But boy, are we in an exciting time. And I

549
00:28:52,600 --> 00:28:55,680
Speaker 2: am just such a fan of the existence of the

550
00:28:55,760 --> 00:28:58,400
Speaker 2: Internet because what it means is that when something is

551
00:28:58,440 --> 00:29:02,760
Speaker 2: discovered now, it's spreads instantly globally. And that sounds so

552
00:29:02,880 --> 00:29:05,840
Speaker 2: obvious to us, but it just wasn't that long ago.

553
00:29:06,080 --> 00:29:09,000
Speaker 2: For example, when I was getting my PhD. Someone discovers something,

554
00:29:09,160 --> 00:29:10,840
Speaker 2: they write a paper, it takes a few months to get

555
00:29:10,840 --> 00:29:11,960
Speaker 2: that published, then it ends up.

556
00:29:11,880 --> 00:29:12,360
Speaker 1: In a journal.

557
00:29:12,400 --> 00:29:14,440
Speaker 2: Then you go to the library, you hope to be

558
00:29:14,480 --> 00:29:16,800
Speaker 2: lucky enough to find that paper. You stick it on

559
00:29:16,840 --> 00:29:18,960
Speaker 2: the xerox machine and you hold it down with your

560
00:29:18,960 --> 00:29:21,760
Speaker 2: elbow because it's a big thick binder, and you try

561
00:29:21,800 --> 00:29:24,760
Speaker 2: to xerox to the paper. Like it was really slow

562
00:29:24,800 --> 00:29:28,800
Speaker 2: to get information around just not that long ago, and

563
00:29:29,240 --> 00:29:31,920
Speaker 2: suddenly that's all changed, and that really makes a big difference.

564
00:29:35,280 --> 00:29:36,840
Speaker 1: David, thank you so much. I hope you'll join us

565
00:29:36,840 --> 00:29:39,320
Speaker 1: again on the podcast soon. I hope you'll feed us

566
00:29:39,320 --> 00:29:41,600
Speaker 1: with some of the interesting things that you'll see developing

567
00:29:41,640 --> 00:29:45,200
Speaker 1: in Stanford. And I really really enjoyed today's conversation.

568
00:29:45,720 --> 00:29:47,120
Speaker 2: Thank you, Oz, great to be here.

569
00:29:53,040 --> 00:29:55,720
Speaker 1: That's it for this week for tech Stuff. I'm as Valoshian.

570
00:29:55,840 --> 00:29:59,480
Speaker 1: This episode was produced by Eliza Dennis and Victoria Domingez

571
00:30:00,000 --> 00:30:03,360
Speaker 1: Pektive produced by me Kara Price and Kate Osborne for

572
00:30:03,400 --> 00:30:08,080
Speaker 1: Kodeidoscope and Katrina Norvelle Fireart Podcast. Jack Insley mixed this

573
00:30:08,160 --> 00:30:11,479
Speaker 1: episode and Kyle Murdoch wrote our theme song. Join us

574
00:30:11,480 --> 00:30:14,640
Speaker 1: this Friday for Textaff's Weekend Tech we'll run through the

575
00:30:14,640 --> 00:30:20,440
Speaker 1: headlines and hear from tech entrepreneur and researcher Azimazar. Please rate,

576
00:30:20,560 --> 00:30:23,280
Speaker 1: review and reach out to us at tech Stuff podcast

577
00:30:23,320 --> 00:30:25,680
Speaker 1: at gmail dot com. We want to know what's on

578
00:30:25,720 --> 00:30:26,160
Speaker 1: your mind.