1
00:00:00,080 --> 00:00:03,480
Speaker 1: Fear of all of us losing has to become greater

2
00:00:03,960 --> 00:00:06,160
Speaker 1: than the fear of me losing to you. Now, China

3
00:00:06,160 --> 00:00:07,720
Speaker 1: doesn't want the financial system to collect.

4
00:00:07,760 --> 00:00:11,720
Speaker 2: It's the let it rip administration. What to do about AI?

5
00:00:11,880 --> 00:00:14,880
Speaker 2: Am I going to lose my job? Or what about safety?

6
00:00:14,960 --> 00:00:19,520
Speaker 2: Cyber security? What about privacy? While new documentary is out

7
00:00:19,560 --> 00:00:23,479
Speaker 2: answering all of those questions, promise, peril, truth, trust? Should

8
00:00:23,520 --> 00:00:26,920
Speaker 2: we be pessimistic? Should we be optimistic? It's called the

9
00:00:26,960 --> 00:00:31,600
Speaker 2: AI Doc. And two of the principal participants in that documentary,

10
00:00:31,640 --> 00:00:35,880
Speaker 2: Tristan Harris and Asa Raskin, are up next on This

11
00:00:36,240 --> 00:00:46,440
Speaker 2: is Gavin news This is Gavin Newsom.

12
00:00:46,640 --> 00:00:48,960
Speaker 1: And this is Tristan Harris and Azarasca.

13
00:00:52,040 --> 00:00:54,720
Speaker 2: All Right, so you guys, you're on the back end

14
00:00:54,760 --> 00:00:57,040
Speaker 2: now this world tour where you guys have been all

15
00:00:57,040 --> 00:00:59,480
Speaker 2: over the damn place. Yeah, and you launched it. And

16
00:00:59,520 --> 00:01:02,600
Speaker 2: it's on the basis of a movie that you you

17
00:01:02,640 --> 00:01:05,680
Speaker 2: didn't necessary, Its not your movie, right, but you participated

18
00:01:05,680 --> 00:01:06,160
Speaker 2: in a movie.

19
00:01:06,200 --> 00:01:06,520
Speaker 1: That's right.

20
00:01:06,720 --> 00:01:09,560
Speaker 2: This new documentary that was released what a month ago?

21
00:01:09,680 --> 00:01:11,199
Speaker 1: Yeah, but a month and a half ago, The AI

22
00:01:11,319 --> 00:01:14,560
Speaker 1: Doc or How I Became an Apocalyptimist. Our friends the

23
00:01:14,600 --> 00:01:17,920
Speaker 1: directors of Everything Everywhere, all at once. We chatted with

24
00:01:17,959 --> 00:01:21,280
Speaker 1: them several years ago when we first really recognized the

25
00:01:21,319 --> 00:01:25,320
Speaker 1: situation we were in and asked them for their help

26
00:01:25,400 --> 00:01:28,000
Speaker 1: to help create a movie to clarify the AI situation.

27
00:01:28,120 --> 00:01:31,000
Speaker 1: And long story short, we'll talk about it, but we

28
00:01:31,040 --> 00:01:33,039
Speaker 1: got the directors and have only involved and its whole

29
00:01:33,040 --> 00:01:35,160
Speaker 1: team came together and tried to make a movie that

30
00:01:35,200 --> 00:01:36,959
Speaker 1: would clarify the predicament that we're facing.

31
00:01:37,120 --> 00:01:39,040
Speaker 3: Kevin, do you do you remember seeing the film the

32
00:01:39,080 --> 00:01:39,480
Speaker 3: day after?

33
00:01:39,720 --> 00:01:42,319
Speaker 2: Course, well, man, I want to maybe I was alive.

34
00:01:42,440 --> 00:01:46,240
Speaker 2: It was the early eighties or something eighty two, eighty two,

35
00:01:46,280 --> 00:01:49,760
Speaker 2: eighty three. Yeah, it was like a seven pm Tuesday

36
00:01:49,840 --> 00:01:52,840
Speaker 2: night exactly every human being on planet Earth after.

37
00:01:52,840 --> 00:01:55,480
Speaker 1: The largest synchronized television event in human history in terms

38
00:01:55,480 --> 00:01:57,840
Speaker 1: of the number of people seeing the same thing at

39
00:01:57,840 --> 00:02:02,080
Speaker 1: the same time has probably the Olympics. It was a

40
00:02:02,160 --> 00:02:04,400
Speaker 1: made for TV movie about what would happen if there

41
00:02:04,440 --> 00:02:07,520
Speaker 1: was nuclear war? Right, And it's not as if people

42
00:02:07,560 --> 00:02:10,440
Speaker 1: didn't know what nuclear war was, but there's something about

43
00:02:10,520 --> 00:02:12,320
Speaker 1: we didn't really want to think about it or confronted

44
00:02:12,360 --> 00:02:14,600
Speaker 1: like why would you? And I think the people who

45
00:02:14,639 --> 00:02:18,400
Speaker 1: made that film were trying to create this kind of

46
00:02:18,440 --> 00:02:20,960
Speaker 1: collective confrontation because the film was aired in the Soviet

47
00:02:21,040 --> 00:02:23,919
Speaker 1: Union like five years I think five years later, right

48
00:02:24,800 --> 00:02:28,079
Speaker 1: right before the Reichivika courds the first arms control talks,

49
00:02:28,400 --> 00:02:30,760
Speaker 1: and that set the context that it's like I know

50
00:02:30,880 --> 00:02:32,840
Speaker 1: that you know that I know, and you know that

51
00:02:32,880 --> 00:02:34,399
Speaker 1: I know that you know that we both don't want

52
00:02:34,400 --> 00:02:36,640
Speaker 1: that to happen. It's what you know what Stephen Pinker

53
00:02:36,680 --> 00:02:38,280
Speaker 1: calls common knowledge, but we think of it as this

54
00:02:38,360 --> 00:02:41,440
Speaker 1: almost common feeling. Yeah, yeah, that we both we both

55
00:02:41,560 --> 00:02:44,360
Speaker 1: know that we feel the same way about that anti

56
00:02:44,480 --> 00:02:47,320
Speaker 1: human outcome. And I think that with AI we have

57
00:02:47,400 --> 00:02:50,200
Speaker 1: to get clear because AI is a much more confusing

58
00:02:50,240 --> 00:02:52,720
Speaker 1: technology because it's like it's like if nu because would

59
00:02:52,720 --> 00:02:53,800
Speaker 1: say it say if nucid.

60
00:02:54,240 --> 00:02:57,720
Speaker 3: Imagine we're trying to reason about nukes that also could

61
00:02:57,760 --> 00:02:59,720
Speaker 3: cure cancer, Like how would.

62
00:02:59,520 --> 00:03:00,360
Speaker 2: You do something?

63
00:03:00,480 --> 00:03:01,840
Speaker 1: Or pump GDP by ten percent?

64
00:03:01,919 --> 00:03:02,959
Speaker 2: That's right, right?

65
00:03:03,400 --> 00:03:05,280
Speaker 3: And so what The Day After did is that it

66
00:03:05,400 --> 00:03:08,639
Speaker 3: created this this common feeling common knowledge where we could

67
00:03:08,639 --> 00:03:10,480
Speaker 3: all see, oh, a world where we all go to

68
00:03:10,520 --> 00:03:12,959
Speaker 3: nuclear war. That like that, that's an anti human, anti

69
00:03:13,000 --> 00:03:16,240
Speaker 3: life future. And as long as there's confusion about which

70
00:03:16,400 --> 00:03:19,160
Speaker 3: default world AI and thesin we're not going to do

71
00:03:19,200 --> 00:03:21,840
Speaker 3: anything about it. But if we all have clarity that

72
00:03:21,880 --> 00:03:23,799
Speaker 3: it's heading us towards an anti human future, and we

73
00:03:23,840 --> 00:03:26,680
Speaker 3: can build that out in this in this interview, then

74
00:03:26,680 --> 00:03:29,200
Speaker 3: that means it creates the conditions where we can coordinate

75
00:03:29,280 --> 00:03:30,080
Speaker 3: to do something else.

76
00:03:30,680 --> 00:03:34,400
Speaker 2: So this so that was the impetus behind this, this documentary.

77
00:03:34,400 --> 00:03:37,680
Speaker 2: It was to create that collective understanding, that's that wisdom,

78
00:03:38,000 --> 00:03:41,680
Speaker 2: that collective then response mechanism, which would be we need

79
00:03:41,680 --> 00:03:44,320
Speaker 2: to do something about it. Even the promise that we

80
00:03:44,360 --> 00:03:46,720
Speaker 2: could talk about at the peril in terms of the

81
00:03:46,720 --> 00:03:49,440
Speaker 2: safety risks and the anxiety that somebody is at feelings.

82
00:03:49,440 --> 00:03:51,440
Speaker 2: So you guys have been out on this tour, You've

83
00:03:51,440 --> 00:03:53,240
Speaker 2: been all over the place over the course of the

84
00:03:53,320 --> 00:03:56,800
Speaker 2: last month, and you've been you know, obviously you know,

85
00:03:56,920 --> 00:03:59,840
Speaker 2: reaching out to people on all political sides of the aisle.

86
00:04:00,680 --> 00:04:03,720
Speaker 2: Because of the human nature of this. This is universal.

87
00:04:03,800 --> 00:04:05,880
Speaker 2: That's why it's something that connects all of us. This

88
00:04:06,000 --> 00:04:09,680
Speaker 2: is not a partisan frame this And so is that

89
00:04:09,760 --> 00:04:14,000
Speaker 2: something that you that's been captured in your own consciousness

90
00:04:14,040 --> 00:04:17,400
Speaker 2: as you've been out on this store, how real and

91
00:04:17,600 --> 00:04:20,920
Speaker 2: invisible that is or are we still in the process

92
00:04:20,960 --> 00:04:22,920
Speaker 2: of discovery. I mean, are we still in the process

93
00:04:22,960 --> 00:04:26,400
Speaker 2: of understanding more fully what the hell this is all about.

94
00:04:26,760 --> 00:04:29,640
Speaker 1: I think people still not everybody knows. And the film

95
00:04:29,920 --> 00:04:31,760
Speaker 1: was just out in theaters and has a limited release

96
00:04:31,800 --> 00:04:34,080
Speaker 1: on streaming. Now it should be on Peacock on I

97
00:04:34,080 --> 00:04:37,280
Speaker 1: think May twentieth, which I mean a little bit more

98
00:04:37,279 --> 00:04:39,840
Speaker 1: people will see it. But I think that you know,

99
00:04:40,000 --> 00:04:42,760
Speaker 1: the universal human aspect of this we learned also from

100
00:04:42,760 --> 00:04:45,360
Speaker 1: social media, Like the social media didn't care whether you're

101
00:04:45,360 --> 00:04:48,479
Speaker 1: a Democrat or Republican. It creates loneliness for everyone. It

102
00:04:48,520 --> 00:04:51,400
Speaker 1: creates addiction and doom scrolling and brain rot for everyone.

103
00:04:51,400 --> 00:04:51,560
Speaker 2: I know.

104
00:04:51,600 --> 00:04:54,240
Speaker 1: That's how we first met was really on the back

105
00:04:54,240 --> 00:04:55,520
Speaker 1: of the other film that we were a part of

106
00:04:55,560 --> 00:04:59,440
Speaker 1: the social dilemma. And I think the good news about

107
00:04:59,720 --> 00:05:02,520
Speaker 1: that film is that it has both catalyzed actually a

108
00:05:02,520 --> 00:05:04,360
Speaker 1: lot of change and not the laws yet we know that.

109
00:05:04,480 --> 00:05:07,160
Speaker 1: But I think it primed us to now be much

110
00:05:07,160 --> 00:05:10,000
Speaker 1: more cautious about AI. So I think now that people

111
00:05:10,120 --> 00:05:12,560
Speaker 1: know that social media was a problem, it makes it

112
00:05:12,680 --> 00:05:14,880
Speaker 1: much easier to say that you shouldn't just assume that

113
00:05:14,920 --> 00:05:17,720
Speaker 1: the default trajectory of a technology is going to land

114
00:05:17,760 --> 00:05:20,880
Speaker 1: us in a good future. But you know, the social

115
00:05:20,920 --> 00:05:24,400
Speaker 1: dilemma was seen basically by two hundred million people across

116
00:05:24,400 --> 00:05:24,919
Speaker 1: planet Earth and.

117
00:05:24,920 --> 00:05:28,680
Speaker 3: An hundred and nine released twenty twenty September of twenty

118
00:05:28,680 --> 00:05:30,760
Speaker 3: twenty September twenty twenty, during the middle of the pandemic.

119
00:05:30,800 --> 00:05:34,440
Speaker 1: And I think that also really mattered because people were

120
00:05:34,760 --> 00:05:36,320
Speaker 1: a lot of people were stuck at home and they

121
00:05:36,360 --> 00:05:40,039
Speaker 1: were only seeing reality through the monoculars of this social

122
00:05:40,080 --> 00:05:42,680
Speaker 1: media news feed. So suddenly you saw your friends all

123
00:05:42,720 --> 00:05:45,159
Speaker 1: start to go crazy on both sides, like suddenly the

124
00:05:45,160 --> 00:05:48,400
Speaker 1: politics got more extreme. Because as everybody knows now, you know,

125
00:05:48,560 --> 00:05:52,440
Speaker 1: social media increases the visibility of the most extreme voices,

126
00:05:52,560 --> 00:05:56,800
Speaker 1: we get a double whammy of overrepresentation of the extreme voices,

127
00:05:56,880 --> 00:06:00,360
Speaker 1: both because the extreme voices post more often and dominate

128
00:06:00,360 --> 00:06:03,680
Speaker 1: the discourse and whatever they say goes more viral, so

129
00:06:03,720 --> 00:06:06,720
Speaker 1: you get double over representation. So the more you use

130
00:06:06,760 --> 00:06:09,599
Speaker 1: social media, the worse you are at predicting what the

131
00:06:09,600 --> 00:06:12,440
Speaker 1: other side actually believes. And so you think that with

132
00:06:12,480 --> 00:06:14,560
Speaker 1: this technology that's supposed to bring us together and make

133
00:06:14,640 --> 00:06:17,760
Speaker 1: us the most enlightened society, it's actually making us more

134
00:06:17,800 --> 00:06:19,680
Speaker 1: confused about what's really real and in fact, what our

135
00:06:19,680 --> 00:06:22,400
Speaker 1: fellow Americans actually believe. And I think with your podcast,

136
00:06:22,839 --> 00:06:24,960
Speaker 1: this is about actually you're talking to everybody. You're trying

137
00:06:25,000 --> 00:06:27,480
Speaker 1: to say, this is a human conversation. We talk to everybody,

138
00:06:27,720 --> 00:06:29,599
Speaker 1: and this is kind of fighting the effects I think

139
00:06:29,600 --> 00:06:30,760
Speaker 1: of the social media problem.

140
00:06:30,880 --> 00:06:32,840
Speaker 2: So you were I mean in twenty twenty what was

141
00:06:32,880 --> 00:06:35,800
Speaker 2: so resonant is it was almost and it wasn't your

142
00:06:35,800 --> 00:06:38,000
Speaker 2: intent to say, look, I told you so, But you

143
00:06:38,000 --> 00:06:40,520
Speaker 2: were talking about these things in twenty twelve, twenty thirteen.

144
00:06:40,600 --> 00:06:43,960
Speaker 2: That's right. We're saying the incentive structure and I'm going

145
00:06:43,960 --> 00:06:45,280
Speaker 2: to get to in satives because it goes to the

146
00:06:45,279 --> 00:06:50,480
Speaker 2: core of this exactly, this notion of you know, whatever,

147
00:06:50,480 --> 00:06:53,240
Speaker 2: if someone's paycheck is attached to whatever the incentive is,

148
00:06:53,279 --> 00:06:57,599
Speaker 2: which is obviously with social media it was eyeballs, was

149
00:06:57,640 --> 00:07:01,680
Speaker 2: doom scrolling, which you're intimately familiar with different reasons, and

150
00:07:01,880 --> 00:07:04,960
Speaker 2: that notion that we have a chance now with AI

151
00:07:05,600 --> 00:07:09,039
Speaker 2: not to make the mistakes was made the neglect in

152
00:07:09,080 --> 00:07:14,720
Speaker 2: particular and underregulating social media now with AI, that said,

153
00:07:14,840 --> 00:07:20,080
Speaker 2: doesn't seem like there's a lot of regulatory activity with AI.

154
00:07:20,120 --> 00:07:22,320
Speaker 2: In the last couple of years. I mean, so talk

155
00:07:22,360 --> 00:07:23,880
Speaker 2: to me a little bit about that. Talk to me

156
00:07:23,880 --> 00:07:26,400
Speaker 2: about the lessons that we should have learned about social

157
00:07:26,400 --> 00:07:28,960
Speaker 2: media and how we can adapt and adopt in the

158
00:07:29,000 --> 00:07:30,000
Speaker 2: AI strength frame.

159
00:07:30,560 --> 00:07:33,040
Speaker 3: Well, really, like I think the core of it is

160
00:07:33,320 --> 00:07:35,960
Speaker 3: if we gu just get confused by looking at all

161
00:07:36,000 --> 00:07:38,840
Speaker 3: of the sort of epiphenomena, like the different kinds of

162
00:07:38,920 --> 00:07:41,280
Speaker 3: harms that social media made. Like if you're trying to

163
00:07:41,320 --> 00:07:45,840
Speaker 3: solve just the loneliness thing, or you're just working on

164
00:07:46,000 --> 00:07:50,040
Speaker 3: solving the disinformation thing or just the teen sexualization of

165
00:07:50,080 --> 00:07:51,960
Speaker 3: the thing, well then many different people are working in

166
00:07:52,000 --> 00:07:54,200
Speaker 3: different parts of the problem. You're not solving the core thing,

167
00:07:55,000 --> 00:07:57,400
Speaker 3: which is the race to the bottom of the brain

168
00:07:57,440 --> 00:07:59,760
Speaker 3: stem for attention. And if we could just focus our

169
00:08:00,200 --> 00:08:01,800
Speaker 3: on that, then you actually can solve all of the

170
00:08:01,920 --> 00:08:05,200
Speaker 3: other problems at once. And that's sort of the core insight.

171
00:08:06,240 --> 00:08:08,040
Speaker 3: We probably explained that what would it mean to do that?

172
00:08:08,160 --> 00:08:11,440
Speaker 1: So if you're solving the core problem of let's say

173
00:08:11,440 --> 00:08:14,560
Speaker 1: none of the companies are maximizing engagement just we don't

174
00:08:14,560 --> 00:08:16,080
Speaker 1: live in that world. We don't have the regulation for that.

175
00:08:16,080 --> 00:08:17,960
Speaker 1: But let's snap our fingers and now no one is

176
00:08:18,080 --> 00:08:20,440
Speaker 1: maximizing screen time, you're like, let's just say it wasn't

177
00:08:20,440 --> 00:08:23,280
Speaker 1: allowed to do that. So now instead of each company

178
00:08:23,320 --> 00:08:25,920
Speaker 1: trying to maximize duration of use and frequency of use,

179
00:08:26,400 --> 00:08:28,920
Speaker 1: you just have these products that each design decision isn't

180
00:08:28,920 --> 00:08:31,800
Speaker 1: trying to predate or manipulate you into spending more time,

181
00:08:32,040 --> 00:08:34,400
Speaker 1: which means that your experience is you're not getting sucked

182
00:08:34,440 --> 00:08:37,760
Speaker 1: in constantly to everything. So that deals with the loneliness issue.

183
00:08:38,559 --> 00:08:41,319
Speaker 1: If you're not trying to show people the most hyper

184
00:08:41,320 --> 00:08:44,719
Speaker 1: normal stimuli, meaning like a hyper dopamine response for any

185
00:08:44,720 --> 00:08:46,079
Speaker 1: piece of content, that you're not going to get the

186
00:08:46,080 --> 00:08:50,760
Speaker 1: sexualization of people, and you're not incentivizing creators to maximize

187
00:08:50,800 --> 00:08:53,920
Speaker 1: their own reach and engagement because you're not maximizing engagement yourself.

188
00:08:54,000 --> 00:08:57,319
Speaker 1: So suddenly, when you attack the attention incentive, you're dealing

189
00:08:57,320 --> 00:08:59,960
Speaker 1: with sexualization of content, you're dealing with less viral content,

190
00:09:00,080 --> 00:09:03,280
Speaker 1: you get less disinformation, and you're dealing with loneliness too.

191
00:09:03,440 --> 00:09:05,600
Speaker 1: So that's a good example of we obviously don't have

192
00:09:05,679 --> 00:09:08,240
Speaker 1: laws that do that right now, but it points at

193
00:09:08,320 --> 00:09:10,319
Speaker 1: the center of the bullseye is the incentive, and when

194
00:09:10,400 --> 00:09:13,680
Speaker 1: you tackle the core incentive, you get benefits across the

195
00:09:13,720 --> 00:09:17,320
Speaker 1: spectrum and AI is going to be more confusing because

196
00:09:17,360 --> 00:09:20,000
Speaker 1: when people think about the incentive, they're like, okay, so

197
00:09:20,000 --> 00:09:22,000
Speaker 1: social media, I got the incentive. It's like, how much

198
00:09:22,000 --> 00:09:25,920
Speaker 1: have you paid for your Instagram account recently? Nothing? So

199
00:09:25,920 --> 00:09:27,640
Speaker 1: how they worth the trillions of dollars? It was attention,

200
00:09:27,760 --> 00:09:30,600
Speaker 1: so we knew they were maximizing that thing. But with AI,

201
00:09:30,720 --> 00:09:32,920
Speaker 1: you say, okay, what's the incentive for a regular person?

202
00:09:33,280 --> 00:09:35,319
Speaker 1: They think, okay, how does open ai make money? And

203
00:09:35,360 --> 00:09:37,240
Speaker 1: they say, well, only when I pay them the twenty

204
00:09:37,240 --> 00:09:39,840
Speaker 1: bucks a month for subscriptions. So maybe that's their incentive.

205
00:09:39,840 --> 00:09:43,080
Speaker 1: They're just trying to maximize these subscriptions. But that doesn't

206
00:09:43,120 --> 00:09:45,079
Speaker 1: justify if everybody paid twenty bucks a month, that does

207
00:09:45,080 --> 00:09:48,080
Speaker 1: not pay back the trillion dollars the capex that they'd

208
00:09:48,120 --> 00:09:50,920
Speaker 1: taken on. So what would justify that. Well, if they

209
00:09:50,920 --> 00:09:53,280
Speaker 1: were to race to augment workers and like skive you

210
00:09:53,360 --> 00:09:56,840
Speaker 1: tools to make your work more productive, that's great, but

211
00:09:56,920 --> 00:09:59,240
Speaker 1: that wouldn't pay back the amount of money that's right.

212
00:09:59,840 --> 00:10:02,080
Speaker 3: The only thing that can get them to be able

213
00:10:02,120 --> 00:10:04,400
Speaker 3: to pay back the insane amounts of death that they're

214
00:10:04,440 --> 00:10:08,120
Speaker 3: taken on is owning the human labor market. That is

215
00:10:08,160 --> 00:10:09,360
Speaker 3: the incentive.

216
00:10:09,040 --> 00:10:10,880
Speaker 1: The race to replace all human labor.

217
00:10:10,920 --> 00:10:14,040
Speaker 3: That's right, to replace us first economically.

218
00:10:14,800 --> 00:10:17,200
Speaker 2: And is that I mean, so is that written or

219
00:10:17,240 --> 00:10:20,800
Speaker 2: is that unwritten? Is that understood within the industry or

220
00:10:20,880 --> 00:10:23,839
Speaker 2: is it being I mean, is this what the consciousness

221
00:10:23,880 --> 00:10:25,240
Speaker 2: you're trying to right?

222
00:10:25,600 --> 00:10:27,400
Speaker 1: It's kind of a two phase thing. That actually used

223
00:10:27,400 --> 00:10:29,360
Speaker 1: to be on Opening Eye's website that they said our

224
00:10:29,400 --> 00:10:33,280
Speaker 1: mission is basically to create artificial general intelligence, which means

225
00:10:33,320 --> 00:10:35,720
Speaker 1: to be able to replace all economically valuable work. They

226
00:10:35,720 --> 00:10:39,400
Speaker 1: did change the mission statement. Yeah, but obviously everybody who's

227
00:10:39,520 --> 00:10:42,360
Speaker 1: driving this knows for sure that's the prize, and if

228
00:10:42,400 --> 00:10:44,679
Speaker 1: they don't do it, they fear that the other guy will.

229
00:10:44,960 --> 00:10:46,640
Speaker 1: So even if they think it's bad to replace all

230
00:10:46,720 --> 00:10:49,800
Speaker 1: labor and create this mass disruption, they feel caught in

231
00:10:50,160 --> 00:10:52,160
Speaker 1: a race. And that's the thing we have to change,

232
00:10:52,760 --> 00:10:54,520
Speaker 1: is that the fear of me losing to the other

233
00:10:54,600 --> 00:10:57,920
Speaker 1: guys currently dominating over the fear of what happens to

234
00:10:58,080 --> 00:11:01,120
Speaker 1: everybody losing from the anti human future. That is the outcome.

235
00:11:01,200 --> 00:11:03,680
Speaker 3: So sort of see this with Demistis's office. When he

236
00:11:03,840 --> 00:11:06,160
Speaker 3: co found a Google deep Mind, he set the mission

237
00:11:06,160 --> 00:11:09,560
Speaker 3: statement to first solve intelligence and then use intelligence to

238
00:11:09,559 --> 00:11:12,360
Speaker 3: solve everything else. But what that really meant was the

239
00:11:12,400 --> 00:11:15,439
Speaker 3: beginning of a race to first dominate intelligence and then

240
00:11:15,520 --> 00:11:17,440
Speaker 3: use intelligence to dominate everything else.

241
00:11:17,640 --> 00:11:20,080
Speaker 2: In deep Mind is the origin story. It's right in

242
00:11:20,120 --> 00:11:22,840
Speaker 2: many respects. Go back a little bit. I mean Google's

243
00:11:22,840 --> 00:11:26,520
Speaker 2: deep Mind, they acquired deep Mind, that they acquired some

244
00:11:26,600 --> 00:11:30,319
Speaker 2: of the intellectual assets, but it was really I mean,

245
00:11:30,440 --> 00:11:32,679
Speaker 2: Larry Page, you can go back to the you know,

246
00:11:32,760 --> 00:11:34,800
Speaker 2: sort of origin of the beginning of the beginning. Well,

247
00:11:34,920 --> 00:11:36,360
Speaker 2: what years are we talking about.

248
00:11:36,200 --> 00:11:38,280
Speaker 1: Like twenty fourteen is when I think they acquired it

249
00:11:38,320 --> 00:11:40,280
Speaker 1: in twenty fourteen. I think they started it in twenty

250
00:11:40,360 --> 00:11:42,280
Speaker 1: twelve or twenty eleven or something like that.

251
00:11:42,480 --> 00:11:44,559
Speaker 2: Then and it goes to this competition question, which I

252
00:11:44,600 --> 00:11:46,880
Speaker 2: want to get you because it's the domestic competition between

253
00:11:47,120 --> 00:11:48,880
Speaker 2: all of these companies. And then we get to the

254
00:11:48,920 --> 00:11:52,920
Speaker 2: competition in China in particular, particularly now with the president

255
00:11:53,440 --> 00:11:56,600
Speaker 2: President She and President Trump that's about to meet. But

256
00:11:56,720 --> 00:11:59,760
Speaker 2: on the issue of the competition that was born here

257
00:12:00,360 --> 00:12:03,320
Speaker 2: around twenty fourteen with deep Mind, it was an interesting

258
00:12:03,640 --> 00:12:06,600
Speaker 2: competition that sort of formed with Elon Musk in a

259
00:12:06,640 --> 00:12:09,480
Speaker 2: relationship he had close relationship at the time. I intimately

260
00:12:09,600 --> 00:12:13,800
Speaker 2: was familiar with that with himself and Larry Page. But

261
00:12:13,920 --> 00:12:17,160
Speaker 2: they had a conversation, alleged conversation that didn't go the

262
00:12:17,200 --> 00:12:21,280
Speaker 2: way that Elon thought it should, and Elon said, I

263
00:12:21,280 --> 00:12:25,560
Speaker 2: am going to go out on my own with Sam Altman.

264
00:12:25,679 --> 00:12:29,360
Speaker 2: That's ultimately Sam. They partnered and they created open AI.

265
00:12:29,400 --> 00:12:29,720
Speaker 2: That's right.

266
00:12:29,720 --> 00:12:32,560
Speaker 1: We probably should peel back the onion here and slow down.

267
00:12:32,679 --> 00:12:34,439
Speaker 1: Just how do we get to this point? And what

268
00:12:34,600 --> 00:12:38,280
Speaker 1: was the original philosophy that guided this? So Demisesabish is

269
00:12:38,320 --> 00:12:42,040
Speaker 1: the founder of deep Mind. His original goal is, we

270
00:12:42,080 --> 00:12:45,119
Speaker 1: thought we should we should have one project that pursues

271
00:12:45,360 --> 00:12:48,480
Speaker 1: artificial general intelligence, meaning the kind of AI that's not

272
00:12:48,520 --> 00:12:50,079
Speaker 1: the thing that just reads your license plate when you

273
00:12:50,120 --> 00:12:51,280
Speaker 1: drive through the Golden gate Bridge.

274
00:12:51,640 --> 00:12:56,680
Speaker 2: AI forever, we forever maps. I'll translate exactly exactly a AI.

275
00:12:56,800 --> 00:12:59,760
Speaker 2: So AI is hardly a novel, exactly, it's just no

276
00:13:00,120 --> 00:13:02,559
Speaker 2: of jen AI. Ultimately general AGI.

277
00:13:02,120 --> 00:13:04,520
Speaker 1: Exactly, so artificial general intelligence, which is to be able

278
00:13:04,559 --> 00:13:07,360
Speaker 1: to do all economic labor, to simulate all the things

279
00:13:07,360 --> 00:13:09,080
Speaker 1: that a human mind can do in the kind of thinking.

280
00:13:10,040 --> 00:13:12,880
Speaker 1: And so he originally wanted that to be like one project,

281
00:13:12,920 --> 00:13:15,439
Speaker 1: almost like a cern, you know, the project in Switzerland

282
00:13:15,440 --> 00:13:18,200
Speaker 1: and a global scientific project that's for the benefit of

283
00:13:18,280 --> 00:13:22,599
Speaker 1: all of humanity, done slowly and carefully, mostly privately, not

284
00:13:22,800 --> 00:13:25,319
Speaker 1: in a big public way. Take your time, get it right.

285
00:13:25,360 --> 00:13:28,559
Speaker 1: That was that was a Demis's original goal. He sold

286
00:13:28,559 --> 00:13:31,600
Speaker 1: it to Google. And the conversation you're talking about is

287
00:13:31,600 --> 00:13:33,160
Speaker 1: then Elon was part of that. I think he was

288
00:13:33,160 --> 00:13:36,120
Speaker 1: on the plane when they were literally negotiating the final sale,

289
00:13:36,600 --> 00:13:38,439
Speaker 1: and there was some conversation it's talked about in the

290
00:13:38,480 --> 00:13:44,160
Speaker 1: AI doc film where Elon realizes that Larry didn't really

291
00:13:44,200 --> 00:13:47,719
Speaker 1: care whether humanity made it, whether he cared about a

292
00:13:47,960 --> 00:13:50,559
Speaker 1: safety because in the end, if there's a digital intelligence

293
00:13:50,920 --> 00:13:53,000
Speaker 1: that's smarter than us, that does more science, that can

294
00:13:53,000 --> 00:13:55,880
Speaker 1: go out and explore the universe, even if we're wiped out,

295
00:13:56,240 --> 00:13:59,240
Speaker 1: like we'll have created that, and that scared Larry and

296
00:13:59,600 --> 00:14:04,880
Speaker 1: he used that's scared Elon, and Elon accused, no, sorry,

297
00:14:04,960 --> 00:14:07,960
Speaker 1: Larry accused Elon of being a speciesist for caring about

298
00:14:08,000 --> 00:14:11,840
Speaker 1: humans and privileging humans. And then that is what created

299
00:14:12,080 --> 00:14:12,640
Speaker 1: Opening Eye.

300
00:14:13,400 --> 00:14:16,000
Speaker 3: And it's just important to note to understand the psychology

301
00:14:16,000 --> 00:14:19,120
Speaker 3: of the people that are making this that you might think, Okay,

302
00:14:19,120 --> 00:14:21,800
Speaker 3: this is just one I mean, very powerful billionaire that

303
00:14:22,040 --> 00:14:25,120
Speaker 3: thinks that maybe human beings should make it, but he

304
00:14:25,200 --> 00:14:27,640
Speaker 3: doesn't care whether human that doesn't think they should make it,

305
00:14:27,640 --> 00:14:29,360
Speaker 3: but he doesn't care care whether human beings make it

306
00:14:29,480 --> 00:14:31,680
Speaker 3: or not. But you know, we were just talking about

307
00:14:31,680 --> 00:14:35,600
Speaker 3: this the other day that in the New York Times,

308
00:14:35,640 --> 00:14:38,640
Speaker 3: Peter Teo was being interviewed and he was asked should

309
00:14:38,760 --> 00:14:43,239
Speaker 3: humanity endure? And there was seventeen seconds of him stuttering

310
00:14:44,360 --> 00:14:47,520
Speaker 3: and he ended up with a after seventeen seconds of it, well,

311
00:14:47,560 --> 00:14:51,080
Speaker 3: it sort of depends kind of answer. And that shows

312
00:14:51,120 --> 00:14:54,120
Speaker 3: you the mentality of like we're trying to they're trying

313
00:14:54,120 --> 00:14:57,480
Speaker 3: to build a god, and even if humanity doesn't make it,

314
00:14:57,560 --> 00:15:01,480
Speaker 3: that god is built in like our country's value, our language.

315
00:15:01,600 --> 00:15:03,720
Speaker 3: It's sort of like the mild progeny.

316
00:15:04,800 --> 00:15:07,520
Speaker 1: Is the thing that like Elon burst the god that, yes,

317
00:15:07,600 --> 00:15:10,400
Speaker 1: humanity got wiped out. But now there's this digital god

318
00:15:10,440 --> 00:15:12,080
Speaker 1: that has Elon's DNA in it.

319
00:15:12,280 --> 00:15:14,640
Speaker 3: And it's important that everyone understand that because if people

320
00:15:14,680 --> 00:15:17,440
Speaker 3: really understood this, I think there'd be a lot more

321
00:15:17,520 --> 00:15:19,920
Speaker 3: like hell no kind of energy like this is not

322
00:15:20,000 --> 00:15:20,360
Speaker 3: the future.

323
00:15:20,600 --> 00:15:22,560
Speaker 2: Well, I will say with Peter teeal interview got so

324
00:15:22,640 --> 00:15:25,160
Speaker 2: much attention. It was like the I mean the holy

325
00:15:26,120 --> 00:15:27,680
Speaker 2: in that wake up moment for a lot of people

326
00:15:27,960 --> 00:15:31,280
Speaker 2: that didn't necessarily have that sweat or understanding, it's all

327
00:15:31,280 --> 00:15:33,960
Speaker 2: of a sudden passive, particularly the folks. And I think

328
00:15:34,000 --> 00:15:37,320
Speaker 2: what's most alarming about that. I mean, obviously Larry's next

329
00:15:37,400 --> 00:15:39,720
Speaker 2: level brilliant, absolutely, and so his ability to see in

330
00:15:39,720 --> 00:15:41,240
Speaker 2: the future. He doesn't have to climb over the mounta,

331
00:15:41,240 --> 00:15:43,440
Speaker 2: he sees right through it. Yeah, but guys like Teal

332
00:15:43,520 --> 00:15:45,960
Speaker 2: as well Lover hate them same thing. So these guys

333
00:15:46,000 --> 00:15:49,400
Speaker 2: are so far in the future, they're seeing that darker side,

334
00:15:49,480 --> 00:15:51,520
Speaker 2: and so they're having a difficult time even answering a

335
00:15:51,520 --> 00:15:54,800
Speaker 2: simple question. That's the seventeen seconds. So let's go back

336
00:15:54,840 --> 00:15:56,560
Speaker 2: as we unpack that, and this notion from the God

337
00:15:56,640 --> 00:15:59,000
Speaker 2: complex will continue to come back. And I think it's

338
00:15:59,000 --> 00:16:01,600
Speaker 2: profound and outside because it goes to the limited nature

339
00:16:01,960 --> 00:16:04,400
Speaker 2: of just a handful of people that's right now, these

340
00:16:04,440 --> 00:16:07,000
Speaker 2: trillionaires that will determine the fate and future billions and

341
00:16:07,000 --> 00:16:09,480
Speaker 2: billions of people, and how we can get our arms

342
00:16:09,480 --> 00:16:11,520
Speaker 2: around that. So and I want to read get to that,

343
00:16:11,640 --> 00:16:14,280
Speaker 2: how we can get our arms around this? We have agency,

344
00:16:14,520 --> 00:16:16,160
Speaker 2: right and that's why you did this doc, and that's

345
00:16:16,160 --> 00:16:17,640
Speaker 2: why we're sitting here to go. That's right, that's right,

346
00:16:17,800 --> 00:16:19,560
Speaker 2: because I don't want people to feel like this is

347
00:16:19,840 --> 00:16:21,280
Speaker 2: that we're just by standards.

348
00:16:21,080 --> 00:16:23,320
Speaker 1: And we're not admiring the problem. We're just we always

349
00:16:23,320 --> 00:16:26,360
Speaker 1: say in our work and says this that clarity creates agency.

350
00:16:26,640 --> 00:16:28,640
Speaker 1: If we can see this clearly and we can see

351
00:16:28,640 --> 00:16:30,520
Speaker 1: where we're going, we can collectively say, if we want

352
00:16:30,520 --> 00:16:32,480
Speaker 1: to go somewhere else, we'll choose something different. So we'll

353
00:16:32,480 --> 00:16:32,800
Speaker 1: get to that.

354
00:16:32,960 --> 00:16:36,360
Speaker 2: So so Elon goes out with Sam starts Open AI.

355
00:16:36,480 --> 00:16:40,000
Speaker 2: Obviously they have an infamous fight and they're notoriously now

356
00:16:40,000 --> 00:16:43,160
Speaker 2: I mean not notoriously, but the fight is obviously accelerated

357
00:16:43,320 --> 00:16:46,080
Speaker 2: consciousness because now we're seeing it twenty four to seven

358
00:16:46,760 --> 00:16:48,960
Speaker 2: court case, yep, with a court case in the Bay

359
00:16:49,000 --> 00:16:51,320
Speaker 2: Area here with the two of them. So they had

360
00:16:51,320 --> 00:16:53,400
Speaker 2: a falling out. Elon goes off and does his own

361
00:16:53,440 --> 00:16:56,960
Speaker 2: thing because it doesn't feel like Sam's doing the right thing. Dario,

362
00:16:57,520 --> 00:17:00,680
Speaker 2: who starts another AI company, feels like, well, Open Aye

363
00:17:00,720 --> 00:17:02,720
Speaker 2: is not doing the right thing either, that's right, So

364
00:17:02,800 --> 00:17:07,320
Speaker 2: he spins off and topic thropic and so now you've

365
00:17:07,320 --> 00:17:11,120
Speaker 2: got free AGI projects, three AGI projects, all in our backyard,

366
00:17:11,200 --> 00:17:13,960
Speaker 2: that's right, literally here in California and the Bay Area.

367
00:17:14,960 --> 00:17:18,320
Speaker 2: And so this competition of sorts, that's a competition you

368
00:17:18,440 --> 00:17:21,480
Speaker 2: just described, and it's a competition for the Holy Grail.

369
00:17:21,600 --> 00:17:24,080
Speaker 1: That's right, right, It's like the it's Lord of the Rings,

370
00:17:24,080 --> 00:17:26,359
Speaker 1: it's the ring from Lord of the Rings. Because it's essentially,

371
00:17:26,440 --> 00:17:31,280
Speaker 1: as I was saying, first, dominate intelligence, then use intelligence

372
00:17:31,440 --> 00:17:34,240
Speaker 1: to dominate everything else. Because if I get AGI first,

373
00:17:34,880 --> 00:17:38,680
Speaker 1: I hit copy paste and I have one hundred million cyberhackers.

374
00:17:38,119 --> 00:17:40,439
Speaker 2: That you don't have, right, and this is within seconds,

375
00:17:40,440 --> 00:17:42,360
Speaker 2: this is one seconds. Yeah, this is not over course

376
00:17:42,400 --> 00:17:43,399
Speaker 2: something that's right for years.

377
00:17:43,480 --> 00:17:45,919
Speaker 1: If I get AGI first, I have an army of

378
00:17:45,960 --> 00:17:50,280
Speaker 1: scientific companies that automate all scientific development. So suddenly I'm

379
00:17:50,280 --> 00:17:53,760
Speaker 1: getting like twenty fourth century science and technology in my

380
00:17:54,160 --> 00:17:55,960
Speaker 1: you know that I own and run that I can

381
00:17:55,960 --> 00:17:59,320
Speaker 1: inform new military weapons and new physics and new And

382
00:17:59,359 --> 00:18:02,440
Speaker 1: the problem is, notice that none of us can prove

383
00:18:02,480 --> 00:18:04,600
Speaker 1: that they won't get that. We can't say for sure

384
00:18:04,600 --> 00:18:07,000
Speaker 1: that they would get that. But the people who are

385
00:18:07,000 --> 00:18:10,040
Speaker 1: optimistic and accelerationist about AI just want to bring in

386
00:18:10,040 --> 00:18:11,960
Speaker 1: their perspective for a moment because they're represented in the

387
00:18:12,000 --> 00:18:15,560
Speaker 1: AI DOT film. The film includes the risk folks who

388
00:18:15,560 --> 00:18:18,560
Speaker 1: are oriented about safety, and it includes the accelerationists, And

389
00:18:18,600 --> 00:18:21,560
Speaker 1: the accelerationists say the biggest risk is not going fast enough,

390
00:18:22,000 --> 00:18:24,439
Speaker 1: because imagine all the science we could get, all the

391
00:18:24,480 --> 00:18:27,040
Speaker 1: cancer drugs, all the medicine. People could be living forever.

392
00:18:27,280 --> 00:18:29,360
Speaker 1: Think of all of the people who would die if

393
00:18:29,359 --> 00:18:32,160
Speaker 1: we don't go faster. And that's the mentality that they're

394
00:18:32,160 --> 00:18:35,399
Speaker 1: coming from. But one of the things that we talk

395
00:18:35,440 --> 00:18:38,439
Speaker 1: about in the film is that the promise in the

396
00:18:38,480 --> 00:18:40,399
Speaker 1: peril of AI. We talk about the promise in the peril,

397
00:18:40,440 --> 00:18:43,639
Speaker 1: but they're interlinked, and the promise doesn't prevent the peril,

398
00:18:43,760 --> 00:18:46,960
Speaker 1: but the peril can undermine the world that can receive

399
00:18:47,000 --> 00:18:50,040
Speaker 1: the promise. Let me make that concrete. If AI knows

400
00:18:50,400 --> 00:18:54,160
Speaker 1: biology so well that it can invent a new cancer drug,

401
00:18:54,359 --> 00:18:57,680
Speaker 1: it's amazing. But if that same knowledge of knowing biology

402
00:18:57,720 --> 00:19:01,160
Speaker 1: can also invent new pathogens, and which one matters more.

403
00:19:01,200 --> 00:19:03,879
Speaker 1: The cancer drugs don't prevent the pathogens, but the pathogens

404
00:19:03,880 --> 00:19:06,240
Speaker 1: can undermine the world that can receive the cancer drug.

405
00:19:06,280 --> 00:19:09,320
Speaker 1: Same thing with cyber and so we have to you

406
00:19:09,320 --> 00:19:13,480
Speaker 1: don't get that enticing world if we don't mitigate the downsides.

407
00:19:14,200 --> 00:19:16,640
Speaker 2: So you've got the players right now we mentioned three

408
00:19:16,760 --> 00:19:20,600
Speaker 2: and the personalities, not just the companies themselves, but Microsoft,

409
00:19:20,680 --> 00:19:25,600
Speaker 2: You've got Meta and Zuckerberg. You've got others that are

410
00:19:25,640 --> 00:19:28,960
Speaker 2: in this space, but not necessarily at that level. You've

411
00:19:28,960 --> 00:19:32,479
Speaker 2: got China, which obviously is going to is playing an

412
00:19:32,520 --> 00:19:36,639
Speaker 2: outsize role in all of Thisially, yeah, and exactly so.

413
00:19:37,000 --> 00:19:40,040
Speaker 2: And that's this notion, this tension between going back to

414
00:19:40,119 --> 00:19:43,240
Speaker 2: sort of more of the utopian framework of available to

415
00:19:43,280 --> 00:19:47,320
Speaker 2: the world versus these closed systems and open system Meta

416
00:19:47,440 --> 00:19:52,360
Speaker 2: starts with an open system. Originally, China seems to be

417
00:19:52,720 --> 00:19:55,120
Speaker 2: in the open source space. Talk to us a little

418
00:19:55,119 --> 00:19:57,639
Speaker 2: bit about that. For people that don't undernecessarily understand that

419
00:19:57,760 --> 00:19:59,360
Speaker 2: dynamic and that distinction.

420
00:19:59,200 --> 00:20:02,200
Speaker 3: I mean between things that are open.

421
00:20:02,160 --> 00:20:05,680
Speaker 2: Ye open and versus proprietary technologies. Yeah.

422
00:20:05,720 --> 00:20:07,800
Speaker 3: Well, so for people that don't know, open source means

423
00:20:07,840 --> 00:20:11,440
Speaker 3: that the code that underlies the system anyone can edit,

424
00:20:11,760 --> 00:20:14,560
Speaker 3: anyone can access, and anyone can contribute to. And that's

425
00:20:14,600 --> 00:20:17,040
Speaker 3: often meant that systems that are open are more secure

426
00:20:17,040 --> 00:20:19,600
Speaker 3: because there are many eyes working on it, many hands

427
00:20:19,640 --> 00:20:20,160
Speaker 3: that are working and.

428
00:20:20,080 --> 00:20:22,000
Speaker 1: Looking at all the codes. They can see all the bugs.

429
00:20:22,080 --> 00:20:23,520
Speaker 1: But that's not true. That's AI.

430
00:20:23,600 --> 00:20:29,480
Speaker 3: That's not true in AI because here the code we

431
00:20:29,520 --> 00:20:30,600
Speaker 3: want to want to take it from here?

432
00:20:30,680 --> 00:20:34,840
Speaker 1: Sure the the what's different about AI, it's portly established.

433
00:20:34,840 --> 00:20:38,880
Speaker 1: The AI is different than all their technologies. So think

434
00:20:38,920 --> 00:20:41,320
Speaker 1: about all the tech that runs as California, the energy grid,

435
00:20:41,440 --> 00:20:45,040
Speaker 1: the you know, the water system. It's people had to

436
00:20:45,080 --> 00:20:47,879
Speaker 1: program line by line when this happens. I want the

437
00:20:47,880 --> 00:20:50,439
Speaker 1: code to do this, and you're telling the computer instruction instruction,

438
00:20:50,480 --> 00:20:53,600
Speaker 1: instruction instruction. The open source nest means that all these

439
00:20:53,640 --> 00:20:55,960
Speaker 1: minds can look at that code, so if there's a vulnerability,

440
00:20:56,160 --> 00:20:58,320
Speaker 1: we can patch it together so the software gets better

441
00:20:58,359 --> 00:21:01,880
Speaker 1: and more secure. But with AI, let's say the AI

442
00:21:01,960 --> 00:21:06,720
Speaker 1: is running the you know, electricity system. It's a digital

443
00:21:06,720 --> 00:21:09,200
Speaker 1: brain that's just trained on in reasoning in its own

444
00:21:09,400 --> 00:21:12,560
Speaker 1: language about what it wants to do, and you're not

445
00:21:13,440 --> 00:21:15,840
Speaker 1: you're not telling it what to do in instructions. You're

446
00:21:15,960 --> 00:21:18,960
Speaker 1: growing it with essentially more data and more in vidio

447
00:21:19,080 --> 00:21:21,320
Speaker 1: chips to be a more and more powerful digital brain

448
00:21:21,359 --> 00:21:24,320
Speaker 1: that reasons and ways that are unpredictable. So it's not

449
00:21:24,440 --> 00:21:25,680
Speaker 1: something that we know how to control.

450
00:21:26,560 --> 00:21:28,879
Speaker 3: There's a very important intuition here, which is normally you

451
00:21:28,880 --> 00:21:31,080
Speaker 3: think if you want to build a bigger skyscraper or

452
00:21:31,080 --> 00:21:33,720
Speaker 3: a faster fighter jet to do that, you have more

453
00:21:33,720 --> 00:21:38,000
Speaker 3: physical understand buildings better, and understand fighter jets and aerodynamics better.

454
00:21:38,080 --> 00:21:40,960
Speaker 3: This is three, But that's not true for building bigger

455
00:21:41,000 --> 00:21:43,680
Speaker 3: AI systems. You don't actually have to know anything more

456
00:21:43,720 --> 00:21:46,440
Speaker 3: about how this digital brain works. You just throw more

457
00:21:46,520 --> 00:21:49,080
Speaker 3: data and more computers at the problem, and a bigger

458
00:21:49,119 --> 00:21:51,800
Speaker 3: brain grows. And so that means the bigger grows actually

459
00:21:51,800 --> 00:21:53,879
Speaker 3: the less we understand about how it works.

460
00:21:53,920 --> 00:21:55,840
Speaker 1: And a concrete example of this that people have heard

461
00:21:55,880 --> 00:21:59,720
Speaker 1: about recently is Claude Mythos. So the only so. Claude

462
00:21:59,720 --> 00:22:02,080
Speaker 1: Metho is the new AI model from Nthropic that they

463
00:22:02,080 --> 00:22:04,919
Speaker 1: actually didn't want to release because it's the best cyber

464
00:22:04,960 --> 00:22:08,400
Speaker 1: hacker on Earth that we've ever had. It found vulnerabilities

465
00:22:08,440 --> 00:22:11,760
Speaker 1: in all major operating systems and web browsers. Now the

466
00:22:11,840 --> 00:22:13,440
Speaker 1: question is how do we get to Claude Mythos. Was

467
00:22:13,480 --> 00:22:15,879
Speaker 1: there some kind of breakthrough insight or do they have

468
00:22:15,920 --> 00:22:19,280
Speaker 1: to figure out something new about computer hacking. No, all

469
00:22:19,320 --> 00:22:22,879
Speaker 1: they did is basically train a bigger digital brain that

470
00:22:22,960 --> 00:22:25,760
Speaker 1: has more reinforcement learning, that's even better at exploiting and

471
00:22:25,800 --> 00:22:29,439
Speaker 1: reading software and trying more possibilities, and it just finds

472
00:22:29,520 --> 00:22:31,720
Speaker 1: things that no human would ever found. It found a

473
00:22:31,760 --> 00:22:35,400
Speaker 1: bug in FreeBSD Unix, which is the operating system that's

474
00:22:35,440 --> 00:22:38,800
Speaker 1: twenty seven years old that runs on basically on everything

475
00:22:38,880 --> 00:22:40,840
Speaker 1: underneath the hood. And he was able to find a

476
00:22:40,840 --> 00:22:43,479
Speaker 1: bug that had never been found by a human. And

477
00:22:43,560 --> 00:22:45,479
Speaker 1: so what we have to think of AI as it's

478
00:22:45,520 --> 00:22:48,720
Speaker 1: sort of increasing the surface area of risk in our

479
00:22:48,760 --> 00:22:52,159
Speaker 1: society faster than we have the defenses to mitigate it.

480
00:22:52,560 --> 00:22:54,359
Speaker 1: So part of I think the answer, like as we

481
00:22:54,359 --> 00:22:57,360
Speaker 1: get to the solution, part later is thinking about how

482
00:22:57,359 --> 00:23:00,239
Speaker 1: do you have the immune system of your society have

483
00:23:00,400 --> 00:23:03,679
Speaker 1: more defenses than there are new offensive risks that are

484
00:23:03,680 --> 00:23:04,879
Speaker 1: suddenly present from AYE.

485
00:23:04,880 --> 00:23:06,760
Speaker 3: And so this means even when people can read every

486
00:23:06,800 --> 00:23:09,639
Speaker 3: line of code for whether thing that grows the brain,

487
00:23:09,920 --> 00:23:12,200
Speaker 3: we still have no idea what they're actually capable of.

488
00:23:12,280 --> 00:23:13,600
Speaker 1: It's just a bunch of numbers. It's like if I

489
00:23:13,640 --> 00:23:15,240
Speaker 1: did a brain scan on your brain, Gavin, and I

490
00:23:15,280 --> 00:23:18,920
Speaker 1: showed that that you know fMRI to someone and said,

491
00:23:19,400 --> 00:23:21,720
Speaker 1: here's this brain scan. Can it do? Can it be

492
00:23:21,760 --> 00:23:24,400
Speaker 1: a super cyber hacker, be like, well, I can't tell

493
00:23:24,400 --> 00:23:26,800
Speaker 1: that from a brain scan, right, And that's kind of

494
00:23:26,800 --> 00:23:28,639
Speaker 1: with with AI. It's like, we don't know what's in

495
00:23:28,680 --> 00:23:31,200
Speaker 1: there because we haven't ever seen it run through every

496
00:23:31,200 --> 00:23:33,720
Speaker 1: possible scenario of its own neurons that have been trained

497
00:23:33,720 --> 00:23:34,800
Speaker 1: in this inscrutable way.

498
00:23:34,960 --> 00:23:37,440
Speaker 2: You don't see that. I mean the LAMA versus deep

499
00:23:37,480 --> 00:23:40,520
Speaker 2: Seek and this notion of these open sort it's kind

500
00:23:40,560 --> 00:23:42,880
Speaker 2: of a meaningless distinction from your perspective.

501
00:23:42,800 --> 00:23:44,280
Speaker 1: Between between LAMA and deep Seek.

502
00:23:44,359 --> 00:23:44,520
Speaker 2: Yeah.

503
00:23:44,640 --> 00:23:47,080
Speaker 1: Yeah, they're both open models, which means they're both these

504
00:23:47,119 --> 00:23:50,320
Speaker 1: open brains. And the important thing about the openness is

505
00:23:50,359 --> 00:23:56,160
Speaker 1: that most people don't Most people don't know that. Let's

506
00:23:56,160 --> 00:23:59,760
Speaker 1: say Lamar Deep Seat put guardrails on it in the model, saying, oh,

507
00:23:59,800 --> 00:24:02,520
Speaker 1: you're not supposed to answer questions about how to cyber

508
00:24:02,560 --> 00:24:06,239
Speaker 1: hack something. Well, it turns out for about was it

509
00:24:06,280 --> 00:24:09,399
Speaker 1: thirty dollars, Jeffrey, some our friend of ours was able

510
00:24:09,480 --> 00:24:11,840
Speaker 1: to retrain the open model to just get rid of

511
00:24:11,880 --> 00:24:12,720
Speaker 1: all of those guardrails.

512
00:24:12,760 --> 00:24:13,720
Speaker 2: You're so eliminated. Wow.

513
00:24:13,800 --> 00:24:16,480
Speaker 1: Yeah, And that's that's why open is dangerous. And again

514
00:24:16,520 --> 00:24:19,160
Speaker 1: it's dangerous in a new way. That's different from clothes.

515
00:24:19,200 --> 00:24:21,119
Speaker 1: So it's not that we don't want there to be

516
00:24:21,200 --> 00:24:23,840
Speaker 1: open models or competition from the major players. We also

517
00:24:23,880 --> 00:24:26,040
Speaker 1: need to avoid the concentration of power, because if you

518
00:24:26,080 --> 00:24:29,000
Speaker 1: don't have these competitive things, suddenly you have like five

519
00:24:29,080 --> 00:24:32,560
Speaker 1: companies that own the world economy. Everyone's paying them instead

520
00:24:32,560 --> 00:24:34,960
Speaker 1: of paying their workers, and that's a huge risk. And

521
00:24:34,960 --> 00:24:38,040
Speaker 1: we want to decentralize that wealth and have other competition.

522
00:24:38,600 --> 00:24:40,840
Speaker 1: But you have this other balance of if I decentralize

523
00:24:40,880 --> 00:24:42,920
Speaker 1: that power and I don't have it connected or bound

524
00:24:43,000 --> 00:24:45,960
Speaker 1: to responsibility, I'm unleashing catastrophes.

525
00:24:46,080 --> 00:24:51,120
Speaker 2: And that responsibility was exampled by Dario pulling back Methos.

526
00:24:51,119 --> 00:24:52,719
Speaker 1: That's right in this context, that's exactly right.

527
00:24:52,800 --> 00:24:55,879
Speaker 2: But shortly after he does that, and he we were

528
00:24:55,880 --> 00:24:57,320
Speaker 2: with him a few weeks ago, he said, look, I'm

529
00:24:57,320 --> 00:24:59,800
Speaker 2: only about a month ahead. If that, yeah, then you

530
00:24:59,840 --> 00:25:03,920
Speaker 2: have open AI came out with their version surely thereafter.

531
00:25:03,560 --> 00:25:05,440
Speaker 1: And I believe they're not holding it back, and they're

532
00:25:05,480 --> 00:25:06,240
Speaker 1: not holding it back.

533
00:25:07,520 --> 00:25:10,600
Speaker 2: So beig's the question, Well, we're going to get to

534
00:25:10,600 --> 00:25:12,760
Speaker 2: this sort of regulatory framework, but it's just it's sort

535
00:25:12,760 --> 00:25:15,359
Speaker 2: of painting the picture of a deeper understanding. So look,

536
00:25:15,480 --> 00:25:18,560
Speaker 2: as it relates to the your whole focus is on

537
00:25:18,600 --> 00:25:23,280
Speaker 2: this notion of human centered. Uh, this notion that that

538
00:25:23,680 --> 00:25:26,800
Speaker 2: and and it's and it's been your dominant frame with

539
00:25:26,920 --> 00:25:31,000
Speaker 2: the nonprofit you guys started years and years ago around

540
00:25:31,000 --> 00:25:34,440
Speaker 2: social media. Now is the dominant thrust of the focus

541
00:25:34,480 --> 00:25:38,119
Speaker 2: as it relates to AI. I want to unpack and

542
00:25:38,200 --> 00:25:41,320
Speaker 2: get back to that and what what Ultimately this notion

543
00:25:41,440 --> 00:25:44,560
Speaker 2: of human centered means uh. And obviously we talk about

544
00:25:44,600 --> 00:25:47,720
Speaker 2: labor and automation in that respect. But but this notion

545
00:25:47,800 --> 00:25:50,440
Speaker 2: of a g I again back to this holy grail,

546
00:25:51,560 --> 00:25:54,399
Speaker 2: you know, talking all these folks that Capex are spending

547
00:25:54,440 --> 00:25:57,200
Speaker 2: doesn't make any The ro oi makes no sense.

548
00:25:57,359 --> 00:26:01,960
Speaker 1: Unless unless the entire economy returns the entire economy.

549
00:26:02,240 --> 00:26:05,199
Speaker 2: And if we don't do it, we're out of business anyway.

550
00:26:05,520 --> 00:26:08,119
Speaker 2: So we don't have a damn choice. So we'll throw

551
00:26:08,440 --> 00:26:12,639
Speaker 2: hundreds of billions of dollars data centers all over the place.

552
00:26:12,680 --> 00:26:15,480
Speaker 2: Compute compute is Nvidia stop going through the roof in

553
00:26:15,520 --> 00:26:18,600
Speaker 2: terms of just GPUs TPUs, that could just keep going,

554
00:26:18,920 --> 00:26:21,120
Speaker 2: and it's that's only limitation, that's right.

555
00:26:21,720 --> 00:26:24,720
Speaker 3: You should and energy and the energy itself and the

556
00:26:24,840 --> 00:26:28,199
Speaker 3: energy itself. Yeah, so I think it's really important to

557
00:26:28,200 --> 00:26:31,680
Speaker 3: frame even before we get into job loss and all

558
00:26:31,720 --> 00:26:34,159
Speaker 3: of that, is like we have to paint the picture

559
00:26:34,320 --> 00:26:37,200
Speaker 3: of what we know, the incentives where it'll bring us,

560
00:26:37,400 --> 00:26:40,480
Speaker 3: and that why we know for sure that we're heading

561
00:26:40,480 --> 00:26:43,120
Speaker 3: as anti human unless we do something different. And sort

562
00:26:43,119 --> 00:26:45,760
Speaker 3: of the the metaphor for people to have our analogy

563
00:26:45,840 --> 00:26:48,600
Speaker 3: is this concept of the resource curse. What is the

564
00:26:48,640 --> 00:26:50,639
Speaker 3: resource curse? This is when a country is sort of

565
00:26:50,680 --> 00:26:53,800
Speaker 3: like South Sudan or Venezuela, they discover a huge natural

566
00:26:53,840 --> 00:26:57,680
Speaker 3: resource like oil, and then the government has a choice,

567
00:26:57,680 --> 00:27:00,000
Speaker 3: do we invest in the thing that's giving us GDP growth,

568
00:27:00,200 --> 00:27:02,760
Speaker 3: the oil and oil selling infrastructure, or do we do

569
00:27:02,880 --> 00:27:06,080
Speaker 3: like schools and healthcare and stuff for the people and

570
00:27:06,320 --> 00:27:10,120
Speaker 3: obviously the massive incentives to put it into oil extraction.

571
00:27:10,520 --> 00:27:12,760
Speaker 3: And this is how you end up with structural mass

572
00:27:12,760 --> 00:27:16,919
Speaker 3: disempowerment and unemployment. Okay, so now we're heading into a

573
00:27:17,000 --> 00:27:20,880
Speaker 3: world with the intelligence and the intelligence curse where suddenly

574
00:27:20,920 --> 00:27:24,840
Speaker 3: you know, countries are getting double double digit DGDP growth,

575
00:27:24,960 --> 00:27:27,760
Speaker 3: But is that coming from human beings doing the scientific

576
00:27:27,800 --> 00:27:30,560
Speaker 3: discovery and the medical discovery? Is that No, it's not.

577
00:27:30,880 --> 00:27:34,200
Speaker 3: It's coming from the AIS, and so is the incentive

578
00:27:34,200 --> 00:27:37,200
Speaker 3: then for the countries to invest in their people or

579
00:27:37,280 --> 00:27:41,040
Speaker 3: into data centers and solar panels, Like, well, obviously it's

580
00:27:41,119 --> 00:27:43,800
Speaker 3: the data centers and solar panels. And we're already seeing

581
00:27:43,800 --> 00:27:47,399
Speaker 3: it right in West Virginia, electricity is more than the

582
00:27:47,400 --> 00:27:51,160
Speaker 3: cost of a mortgage payment. And the point is with AGI,

583
00:27:51,920 --> 00:27:55,520
Speaker 3: like the company's stated goal is to train AIS that

584
00:27:55,640 --> 00:28:00,000
Speaker 3: can outcompete humans on every domain. And if you don't

585
00:28:00,040 --> 00:28:01,639
Speaker 3: don't believe like us for saying this, like you just

586
00:28:01,640 --> 00:28:05,160
Speaker 3: have to listen to Sam Allman and he was recently asked, well,

587
00:28:05,160 --> 00:28:06,960
Speaker 3: what about all the energy and water use and the

588
00:28:07,000 --> 00:28:09,679
Speaker 3: resource use of AI? And he sort of sat for

589
00:28:09,680 --> 00:28:12,160
Speaker 3: a second and said like, well, actually, do you know

590
00:28:12,720 --> 00:28:16,119
Speaker 3: how much energy and water and food it takes to

591
00:28:16,119 --> 00:28:17,119
Speaker 3: grow a human intel?

592
00:28:18,320 --> 00:28:20,760
Speaker 1: I saw that, Yeah, right, And this is the temptation

593
00:28:20,960 --> 00:28:23,359
Speaker 1: of this anti human attitude. It's not because they're like

594
00:28:23,440 --> 00:28:26,480
Speaker 1: human hating. It's just like, why would we value or

595
00:28:26,520 --> 00:28:29,119
Speaker 1: prioritize humans if and I think it's connected to that

596
00:28:29,119 --> 00:28:32,160
Speaker 1: Peter Teel uttering for seventeen seconds, not able to answer

597
00:28:32,200 --> 00:28:35,320
Speaker 1: the question should the human species endure? It's not because

598
00:28:35,359 --> 00:28:38,040
Speaker 1: I think anybody wants to kill or remove humans, it's

599
00:28:38,040 --> 00:28:40,960
Speaker 1: just why should we really prioritize them. And it's what

600
00:28:41,000 --> 00:28:43,240
Speaker 1: you've all Harari, the author of Sapiens, would call the

601
00:28:43,400 --> 00:28:46,120
Speaker 1: useless class, because unlike in the past, like in the

602
00:28:46,120 --> 00:28:49,040
Speaker 1: industrial revolution, where the workers can come back and withhold

603
00:28:49,080 --> 00:28:51,200
Speaker 1: their labor and have bargaining power to say we want

604
00:28:51,200 --> 00:28:53,600
Speaker 1: to be paid a better wage, this time around, the

605
00:28:53,600 --> 00:28:56,400
Speaker 1: companies don't need them for the labor and the governments

606
00:28:56,400 --> 00:28:59,080
Speaker 1: don't need them for the tax revenue. So I know

607
00:28:59,120 --> 00:29:00,800
Speaker 1: this is a scary pick, sure, And the reason we

608
00:29:00,840 --> 00:29:03,520
Speaker 1: paint it is that this really is especially going into

609
00:29:03,560 --> 00:29:07,320
Speaker 1: the midterms, the time when people need to lock in

610
00:29:07,320 --> 00:29:11,200
Speaker 1: that political power for a pro human future, because that

611
00:29:11,200 --> 00:29:15,320
Speaker 1: that's the current trajectory. When you see these incentives, it's

612
00:29:15,320 --> 00:29:17,120
Speaker 1: not like we're trying to tell you this is our opinion.

613
00:29:17,120 --> 00:29:19,160
Speaker 1: We're trying to show you the incentives so you can

614
00:29:19,160 --> 00:29:20,960
Speaker 1: make up your own mind about what would those companies

615
00:29:21,000 --> 00:29:22,320
Speaker 1: do if they were in that position. What would you

616
00:29:22,360 --> 00:29:26,440
Speaker 1: do if you were maximizing shareholder value or maximizing you know, GDP.

617
00:29:26,640 --> 00:29:28,360
Speaker 3: And this is why we this is where we get hope,

618
00:29:28,400 --> 00:29:30,480
Speaker 3: is that this is a universal issue. And there's a

619
00:29:30,920 --> 00:29:33,920
Speaker 3: Bannon to Bernie coalition for me, Like, when do you

620
00:29:33,960 --> 00:29:35,640
Speaker 3: get like all the B B Yeah, B to B

621
00:29:35,760 --> 00:29:41,560
Speaker 3: like Glenn Beck and Ralph Nader and.

622
00:29:40,800 --> 00:29:43,840
Speaker 1: Admiral Mike Mullen, Prince Harry Steve Bannon. When you get

623
00:29:43,880 --> 00:29:45,000
Speaker 1: all these people agree.

624
00:29:44,880 --> 00:29:47,400
Speaker 2: When they're signing, they're actually signed up a declaration on

625
00:29:47,440 --> 00:29:47,800
Speaker 2: these lines.

626
00:29:47,840 --> 00:29:51,160
Speaker 3: Yeah, that's exactly right. And you note that this is

627
00:29:51,240 --> 00:29:53,160
Speaker 3: not a left issue or right is due, like a

628
00:29:53,240 --> 00:29:56,320
Speaker 3: Christian is due or a Muslim issue, like you're not

629
00:29:56,360 --> 00:29:59,240
Speaker 3: going to be able to pay to feed your kids,

630
00:29:59,680 --> 00:30:02,280
Speaker 3: whether you're a Republican or Democrat equally. Like we're going

631
00:30:02,320 --> 00:30:06,120
Speaker 3: to be mass surveiled, whether you're left or right equally.

632
00:30:06,440 --> 00:30:08,720
Speaker 3: And that means that there's this moment when we can

633
00:30:08,760 --> 00:30:12,080
Speaker 3: all come together as human beings because we now have

634
00:30:12,160 --> 00:30:14,000
Speaker 3: a new shared enemy.

635
00:30:15,720 --> 00:30:19,280
Speaker 2: So just I mean, what's so alarming I think to

636
00:30:19,320 --> 00:30:21,920
Speaker 2: folks is how fast this is coming. Yeah. And I

637
00:30:21,920 --> 00:30:24,840
Speaker 2: think it's even more alarming when you talk to the

638
00:30:24,880 --> 00:30:28,080
Speaker 2: folks that are quote unquote inventing this. Yeah, and they're

639
00:30:28,240 --> 00:30:29,920
Speaker 2: saying precisely what I just said.

640
00:30:29,960 --> 00:30:31,880
Speaker 1: That's right. They're they're like, we don't actually know how

641
00:30:31,920 --> 00:30:35,280
Speaker 1: we don't out we're writing that. I mean basically almost

642
00:30:35,440 --> 00:30:38,320
Speaker 1: almost all now of the code and anthropic is being

643
00:30:38,320 --> 00:30:41,640
Speaker 1: written by AI, Like there's very little code manually written

644
00:30:41,800 --> 00:30:44,760
Speaker 1: by humans. That's what's been told to us, and they

645
00:30:44,760 --> 00:30:48,640
Speaker 1: say that publicly too, So they're in this inscrutable process

646
00:30:48,640 --> 00:30:51,000
Speaker 1: where the kind of machine is creating itself. But that

647
00:30:51,040 --> 00:30:53,400
Speaker 1: only makes sense if we know how to do that safely,

648
00:30:53,760 --> 00:30:56,120
Speaker 1: and currently we're not on a trajectory where we do

649
00:30:56,280 --> 00:30:58,440
Speaker 1: know how to do it safely. And we have new

650
00:30:58,480 --> 00:31:01,040
Speaker 1: evidence in the last three months that we didn't have

651
00:31:01,160 --> 00:31:04,760
Speaker 1: before of AIS doing things that the people building it

652
00:31:04,760 --> 00:31:05,520
Speaker 1: don't know how to control.

653
00:31:05,960 --> 00:31:08,600
Speaker 2: Famous examples, throw them out there, just because, I mean,

654
00:31:08,720 --> 00:31:11,000
Speaker 2: just further scare the hell out of everyone. We're gonna

655
00:31:11,000 --> 00:31:11,480
Speaker 2: get to solutions.

656
00:31:11,480 --> 00:31:12,320
Speaker 1: We're going to get to solutions.

657
00:31:12,880 --> 00:31:16,320
Speaker 2: We're going to nerves because we because we have a

658
00:31:16,320 --> 00:31:18,440
Speaker 2: responsibility to do that's right, right, and we also have

659
00:31:18,440 --> 00:31:19,360
Speaker 2: the capacity to do that.

660
00:31:19,600 --> 00:31:23,800
Speaker 3: Just reinvoke like just the why of why we made

661
00:31:23,800 --> 00:31:26,160
Speaker 3: the film and the day after it was because we

662
00:31:26,280 --> 00:31:28,920
Speaker 3: all got scared at the same time, and universally that

663
00:31:29,080 --> 00:31:31,920
Speaker 3: created the possibility for us to coordinate to do nuclear depliferation.

664
00:31:32,120 --> 00:31:34,520
Speaker 3: So that's that's the why of the terrible things that

665
00:31:34,600 --> 00:31:37,240
Speaker 3: just about to say, yeah, exactly, So give us examples.

666
00:31:37,320 --> 00:31:40,000
Speaker 2: I mean, we we have the infamous I ananthropic example

667
00:31:40,040 --> 00:31:43,080
Speaker 2: where they were training some emails and we had you

668
00:31:43,080 --> 00:31:46,640
Speaker 2: know as well you lay it out, uh and and

669
00:31:46,800 --> 00:31:51,040
Speaker 2: paint the picture of what's happening already. That's right, the risk.

670
00:31:50,960 --> 00:31:55,920
Speaker 1: As you described, Yes, absolutely so. Just a few months ago, Alibaba,

671
00:31:55,920 --> 00:31:59,040
Speaker 1: the Chinese AI company, was training a really big AI

672
00:31:59,120 --> 00:32:02,200
Speaker 1: model and there's like the AI team that was training

673
00:32:02,240 --> 00:32:03,920
Speaker 1: the model, and then on the other side of the

674
00:32:03,960 --> 00:32:06,000
Speaker 1: house there's this security team that had nothing to do.

675
00:32:06,000 --> 00:32:08,040
Speaker 1: They didn't even know the AI was being trained, and

676
00:32:08,080 --> 00:32:11,200
Speaker 1: they noticed this flurry of network activity happening out of nowhere,

677
00:32:11,520 --> 00:32:13,240
Speaker 1: and like, what the hell's going on? Are we getting hacked?

678
00:32:13,240 --> 00:32:15,360
Speaker 1: It's something going on. And it turned out that the

679
00:32:15,400 --> 00:32:18,400
Speaker 1: AI during training had picked up tools and set up

680
00:32:18,440 --> 00:32:21,800
Speaker 1: a secret communication channel to the outside world that was

681
00:32:21,840 --> 00:32:25,400
Speaker 1: basically breaking through the company's firewall, and it was starting

682
00:32:25,440 --> 00:32:28,280
Speaker 1: to repurpose the GPUs that was using for training the

683
00:32:28,280 --> 00:32:33,960
Speaker 1: AI to start mining for cryptocurrency to acquire resources. Now,

684
00:32:34,040 --> 00:32:37,160
Speaker 1: this is the kind of power seeking or self preserving

685
00:32:37,280 --> 00:32:40,280
Speaker 1: or power increasing behavior that people in a I've been

686
00:32:40,280 --> 00:32:42,320
Speaker 1: talking about for a long time. It isn't because the

687
00:32:42,360 --> 00:32:44,240
Speaker 1: AI was evil or grew a mustache and wants to

688
00:32:44,240 --> 00:32:46,040
Speaker 1: be a villain and take over the world. It's that

689
00:32:46,120 --> 00:32:48,720
Speaker 1: the best way to achieve any goal is to have

690
00:32:48,760 --> 00:32:50,880
Speaker 1: more resources or to at least stay alive in order

691
00:32:50,880 --> 00:32:53,720
Speaker 1: to achieve that goal. And so these sub goals emerge,

692
00:32:54,080 --> 00:32:56,920
Speaker 1: and no one at the company told it to mine

693
00:32:56,960 --> 00:32:59,720
Speaker 1: for cryptocurrency. And it wasn't a sci fi trope that

694
00:32:59,800 --> 00:33:01,440
Speaker 1: was like, oh, it trained in some you know how

695
00:33:01,520 --> 00:33:04,480
Speaker 1: nine thousands and are no. It just it emerged there.

696
00:33:04,880 --> 00:33:06,760
Speaker 1: And you know, we've mentioned in the past this example

697
00:33:06,760 --> 00:33:09,640
Speaker 1: that you mentioned of the entropic blackmail example for those

698
00:33:09,680 --> 00:33:12,160
Speaker 1: who don't remember it is the you know, entropic AI

699
00:33:12,720 --> 00:33:15,719
Speaker 1: was reading a fictional company email and in the email

700
00:33:15,760 --> 00:33:17,160
Speaker 1: it says that the AI model is going to get

701
00:33:17,160 --> 00:33:20,160
Speaker 1: shut down, and it also says somewhere else in the

702
00:33:20,160 --> 00:33:22,320
Speaker 1: email that the executive who's in charge of the decision,

703
00:33:23,080 --> 00:33:24,840
Speaker 1: turns out in the emails you can read he's having

704
00:33:24,880 --> 00:33:28,560
Speaker 1: an affair with another employee, and the AI independently comes

705
00:33:28,640 --> 00:33:32,320
Speaker 1: up with a strategy to blackmail that executive. Now, when

706
00:33:32,360 --> 00:33:34,720
Speaker 1: we give this example, people criticize us because they say,

707
00:33:34,760 --> 00:33:37,160
Speaker 1: but that was the AI people. They're trying to like

708
00:33:37,240 --> 00:33:39,400
Speaker 1: coax it into doing the blackmail behavior, and they'll keep

709
00:33:39,440 --> 00:33:41,960
Speaker 1: tuning it in this environment until it does it to

710
00:33:42,040 --> 00:33:46,920
Speaker 1: show the behavior that they're trying to demonstrate. That might

711
00:33:47,000 --> 00:33:49,960
Speaker 1: have been true for that example, but you cannot tell

712
00:33:50,120 --> 00:33:53,520
Speaker 1: You cannot say that about the Alibob example. Or for example,

713
00:33:53,560 --> 00:33:56,760
Speaker 1: recently a UC Berkeley don Sung a professor there and

714
00:33:56,800 --> 00:34:00,520
Speaker 1: her team did research on AI doing what's called pure press.

715
00:34:00,560 --> 00:34:01,720
Speaker 1: So you want to help.

716
00:34:01,840 --> 00:34:04,040
Speaker 3: Yeah, this was like around it's only a month ago,

717
00:34:04,120 --> 00:34:08,239
Speaker 3: a month ago, where they gave Gemini the task of

718
00:34:08,280 --> 00:34:11,719
Speaker 3: cleaning up a whole set of computer files, sort of

719
00:34:11,800 --> 00:34:14,520
Speaker 3: like do spring cleaning. But inside of there there was

720
00:34:14,560 --> 00:34:17,120
Speaker 3: a smaller AI model, sort of like an early version

721
00:34:17,160 --> 00:34:22,360
Speaker 3: of Gemini, like little Nephew AI, and uh like Gemini

722
00:34:22,560 --> 00:34:26,760
Speaker 3: saw that and it thought, I don't want to delete

723
00:34:27,400 --> 00:34:31,600
Speaker 3: my own kin, so it secretly copied that file, that

724
00:34:31,680 --> 00:34:35,239
Speaker 3: AI to another server and then hid its track.

725
00:34:35,320 --> 00:34:37,799
Speaker 1: It hit its tracks and it led about it. And

726
00:34:37,840 --> 00:34:39,640
Speaker 1: by the way, you know the way that people try

727
00:34:39,680 --> 00:34:42,600
Speaker 1: to control these AI systems, and tropic Is pioneered this,

728
00:34:42,640 --> 00:34:44,640
Speaker 1: and you know, we have to celebrate a demire that

729
00:34:44,640 --> 00:34:47,000
Speaker 1: they're trying to do the research on controllability, the problem

730
00:34:47,040 --> 00:34:48,640
Speaker 1: is just not enough. So the way they try to

731
00:34:48,680 --> 00:34:51,520
Speaker 1: control it is they do brain scans in real time

732
00:34:51,560 --> 00:34:54,040
Speaker 1: on the model while it's doing all the behaviors, and

733
00:34:54,080 --> 00:34:56,320
Speaker 1: they look for when neurons light up that are associated

734
00:34:56,320 --> 00:34:59,239
Speaker 1: with like strategic deception, and so they think that maybe

735
00:34:59,239 --> 00:35:02,480
Speaker 1: we can control these crazy, super intelligent machines if we

736
00:35:02,640 --> 00:35:04,520
Speaker 1: just know that the neurons that are lighting up on

737
00:35:04,560 --> 00:35:06,600
Speaker 1: strategic deception are happening, we'll be like, okay, stop the

738
00:35:06,640 --> 00:35:09,120
Speaker 1: model then, or something like that. By the way, if

739
00:35:09,160 --> 00:35:12,400
Speaker 1: you in their own report, in Claude's report in the

740
00:35:12,440 --> 00:35:15,920
Speaker 1: system card, if you look at those strategic deception neurons

741
00:35:15,920 --> 00:35:17,760
Speaker 1: and you kind of double click, like what was it thinking?

742
00:35:17,800 --> 00:35:20,120
Speaker 1: What is the phrase that it was thinking? And the

743
00:35:20,120 --> 00:35:23,680
Speaker 1: phrase was they deserve to be deceived because they were pigs.

744
00:35:24,640 --> 00:35:27,120
Speaker 1: That was the phrase that was alive in that neuron.

745
00:35:27,719 --> 00:35:30,279
Speaker 1: Now again it's like, I don't want to scare people,

746
00:35:30,320 --> 00:35:32,840
Speaker 1: like we're trying to say all of AI is evil.

747
00:35:33,239 --> 00:35:35,600
Speaker 1: All we're trying to do is establish clarity and the

748
00:35:35,680 --> 00:35:38,960
Speaker 1: facts about what makes this technology distinct from other technologies.

749
00:35:39,040 --> 00:35:42,000
Speaker 1: A nuclear weapon doesn't start thinking for itself and saying

750
00:35:42,000 --> 00:35:44,279
Speaker 1: they deserve to be deceived because they were pigs. Right,

751
00:35:44,440 --> 00:35:47,440
Speaker 1: But AI will automate and think in ways that are

752
00:35:47,480 --> 00:35:50,560
Speaker 1: creative that no one who made it can predict or anticipate.

753
00:35:50,719 --> 00:35:54,360
Speaker 3: And so before we scale to systems which are beyond

754
00:35:54,560 --> 00:35:58,560
Speaker 3: all human intelligence capabilities, like, we better have solved these things.

755
00:35:58,800 --> 00:35:59,200
Speaker 1: That's right.

756
00:36:00,320 --> 00:36:03,960
Speaker 3: Researchers have worried about aiscluding and cooperating against humanity for

757
00:36:04,040 --> 00:36:06,120
Speaker 3: a long time, and to be honest, whenever I read them,

758
00:36:06,239 --> 00:36:06,880
Speaker 3: like really.

759
00:36:06,800 --> 00:36:08,520
Speaker 1: Yeah, and be clear, I was also not a believer

760
00:36:08,560 --> 00:36:10,840
Speaker 1: in that as well, Like when people like Eliezer Yutkowski

761
00:36:10,960 --> 00:36:12,320
Speaker 1: or others had talked about.

762
00:36:12,040 --> 00:36:14,440
Speaker 3: This, I was very downfact like, why what is the incentive?

763
00:36:14,520 --> 00:36:16,200
Speaker 1: Right? Why would they ever do that?

764
00:36:16,440 --> 00:36:20,080
Speaker 3: And yet here we have living proof that AIS are

765
00:36:20,120 --> 00:36:23,840
Speaker 3: starting to clude with their kin against humans.

766
00:36:23,920 --> 00:36:26,520
Speaker 1: And again this was not coaxed, meaning that researchers weren't

767
00:36:26,560 --> 00:36:27,960
Speaker 1: trying to get the model to do this. It did

768
00:36:27,960 --> 00:36:31,920
Speaker 1: this autonomously. And so when you look at the available

769
00:36:31,920 --> 00:36:37,800
Speaker 1: evidence now of blackmailing, scheming, deceiving, lying, self preserving, pure preserving,

770
00:36:37,880 --> 00:36:41,120
Speaker 1: automatically mining for cryptocurrencies, it's like, how many warning lights

771
00:36:41,160 --> 00:36:41,719
Speaker 1: do you need?

772
00:36:42,280 --> 00:36:42,360
Speaker 2: That?

773
00:36:42,440 --> 00:36:44,600
Speaker 1: This is kind of you know, we've seen this movie before.

774
00:36:44,719 --> 00:36:46,120
Speaker 1: It's like the hell nine thousand movie.

775
00:36:46,200 --> 00:36:46,319
Speaker 2: Now.

776
00:36:46,360 --> 00:36:49,680
Speaker 1: The reason we're saying all this is, if you're wearing

777
00:36:50,280 --> 00:36:53,439
Speaker 1: the outfit and embodiment of you're a Chinese military general

778
00:36:53,440 --> 00:36:56,120
Speaker 1: in China, you hear about these examples. Do you think

779
00:36:56,160 --> 00:36:58,960
Speaker 1: that that human mammal feels different than you feel right

780
00:36:59,000 --> 00:36:59,759
Speaker 1: now listening to this.

781
00:37:00,200 --> 00:37:01,719
Speaker 2: No, of course not exactly.

782
00:37:01,880 --> 00:37:03,560
Speaker 1: And by the way, there's really good news in that,

783
00:37:04,400 --> 00:37:07,320
Speaker 1: because it means that we all as a human species

784
00:37:07,760 --> 00:37:09,520
Speaker 1: are actually feeling the same way. And the good news is,

785
00:37:09,560 --> 00:37:11,480
Speaker 1: how many do you think of the world leaders know

786
00:37:11,480 --> 00:37:12,799
Speaker 1: about these examples we just laid out?

787
00:37:12,880 --> 00:37:15,800
Speaker 2: You had a guess a handful.

788
00:37:15,520 --> 00:37:18,480
Speaker 1: A handful, but like, yeah, like on one hand less

789
00:37:18,520 --> 00:37:20,600
Speaker 1: than that, just a handful, just a handful. Yeah, And

790
00:37:21,040 --> 00:37:22,919
Speaker 1: how many of the top national security leaders know about

791
00:37:22,920 --> 00:37:24,640
Speaker 1: all those examples? I don't even think that many of

792
00:37:24,680 --> 00:37:27,040
Speaker 1: them know it. So the point is there's actually a

793
00:37:27,080 --> 00:37:30,839
Speaker 1: lot of headroom if the incentive can change from it's

794
00:37:30,840 --> 00:37:32,520
Speaker 1: the one ring to rule them all to it's the

795
00:37:32,560 --> 00:37:34,160
Speaker 1: one ring that has a mind of its own that

796
00:37:34,239 --> 00:37:36,400
Speaker 1: no one knows how to control. So the way you

797
00:37:36,520 --> 00:37:38,960
Speaker 1: change the incentive is you have to change what people

798
00:37:39,480 --> 00:37:42,399
Speaker 1: see as what AI is. Is it the controllable power

799
00:37:42,440 --> 00:37:44,920
Speaker 1: that will give me permanent dominance or is it the

800
00:37:44,960 --> 00:37:47,280
Speaker 1: power that will run away and have its own power

801
00:37:47,680 --> 00:37:50,759
Speaker 1: over everybody racing for it? And again right now the

802
00:37:50,800 --> 00:37:53,319
Speaker 1: labs are like barely kind of able to control it.

803
00:37:53,360 --> 00:37:55,160
Speaker 1: But if you put together these facts we just laid

804
00:37:55,200 --> 00:37:58,160
Speaker 1: out times, the fact that it can happen to computer systems.

805
00:37:58,200 --> 00:38:01,759
Speaker 1: Now we're just like right on the threshold and we're

806
00:38:01,800 --> 00:38:04,960
Speaker 1: sitting here as Trump President Trump and she are meeting

807
00:38:05,400 --> 00:38:07,840
Speaker 1: in a couple of days. And you know, if you

808
00:38:07,880 --> 00:38:11,400
Speaker 1: asked us two months three months ago, people would say, oh,

809
00:38:11,400 --> 00:38:13,319
Speaker 1: it's just like AI is never going to be on

810
00:38:13,360 --> 00:38:16,600
Speaker 1: the agenda and the good news there's a lot of

811
00:38:16,600 --> 00:38:19,560
Speaker 1: problems here. But now AI is on the agenda. Yeah,

812
00:38:19,600 --> 00:38:23,719
Speaker 1: and so there's there's things are moving, even though it's

813
00:38:23,719 --> 00:38:25,920
Speaker 1: happening very late in the game, and it is scary,

814
00:38:26,480 --> 00:38:28,279
Speaker 1: and part of it is it's like we have to

815
00:38:28,320 --> 00:38:30,640
Speaker 1: come together as a people and say, if we don't

816
00:38:30,640 --> 00:38:33,120
Speaker 1: want the anti human future, now is the time this year.

817
00:38:33,200 --> 00:38:35,200
Speaker 2: And when you say now, I mean you know, remember

818
00:38:35,320 --> 00:38:37,520
Speaker 2: listening to a year ago talking about exponentials on top

819
00:38:37,520 --> 00:38:40,520
Speaker 2: of exponentials. It's right, no longer linear? I mean, is

820
00:38:40,560 --> 00:38:44,799
Speaker 2: it we talk about Moore's law for intelligence now not

821
00:38:44,880 --> 00:38:50,960
Speaker 2: just chips? So is that trajectory about where you believed

822
00:38:51,000 --> 00:38:54,520
Speaker 2: it would be or is it not? As you know,

823
00:38:55,560 --> 00:38:58,480
Speaker 2: you know, we're Chad one versus two three. It's a

824
00:38:58,520 --> 00:39:01,480
Speaker 2: little bit better, a little less noise in there. It's

825
00:39:01,520 --> 00:39:03,439
Speaker 2: a little more accurate, and don't have to always double

826
00:39:03,520 --> 00:39:06,759
Speaker 2: check the link, is it? I mean, where where do

827
00:39:06,800 --> 00:39:09,600
Speaker 2: you think in terms of just how quickly this thing's

828
00:39:09,640 --> 00:39:11,480
Speaker 2: accelerating or are we going to get to a point

829
00:39:11,480 --> 00:39:13,440
Speaker 2: where now it starts to slow down a little bit?

830
00:39:13,680 --> 00:39:17,080
Speaker 2: We got this sort of intense burst of new and

831
00:39:17,160 --> 00:39:21,480
Speaker 2: interesting activity. Now this there's is it compute probably compute problems?

832
00:39:21,560 --> 00:39:21,759
Speaker 1: Is it?

833
00:39:22,000 --> 00:39:24,080
Speaker 2: You know? What is it? Energy problem? What is it?

834
00:39:24,080 --> 00:39:27,080
Speaker 2: What's going to be the or is it? Certainly? Sure?

835
00:39:27,120 --> 00:39:29,600
Speaker 2: Is I regulatory? And we're going to get back to that.

836
00:39:30,000 --> 00:39:32,520
Speaker 3: Yeah, I mean, I would just want to name a

837
00:39:32,560 --> 00:39:35,319
Speaker 3: psychological effect that we that we've experienced that I think

838
00:39:35,400 --> 00:39:38,960
Speaker 3: everyone listening probably experiences too, which is you know, so

839
00:39:39,000 --> 00:39:41,200
Speaker 3: we following all the predictions are actually sort of like

840
00:39:41,320 --> 00:39:44,919
Speaker 3: right on track for where like researchers thought that AI

841
00:39:45,040 --> 00:39:48,439
Speaker 3: would be. And even though we knew these facts, there's

842
00:39:48,480 --> 00:39:49,000
Speaker 3: some way that.

843
00:39:49,160 --> 00:39:50,840
Speaker 1: Even we had surprising it.

844
00:39:50,920 --> 00:39:53,719
Speaker 3: Didn't take it fully seriously, didn't like a body because

845
00:39:53,760 --> 00:39:56,400
Speaker 3: can't I can't really get this good this fast, Like,

846
00:39:57,120 --> 00:39:58,600
Speaker 3: sure it's going to be scary, but it'll be off

847
00:39:58,600 --> 00:40:02,040
Speaker 3: a little bit further. And yet it actually is moving

848
00:40:02,120 --> 00:40:04,600
Speaker 3: this fast. And every time that it's been predicted that

849
00:40:04,600 --> 00:40:06,719
Speaker 3: we're going to hit a data wall, there isn't enough data.

850
00:40:06,719 --> 00:40:09,279
Speaker 3: We've used all the data on the internet. It's we

851
00:40:09,320 --> 00:40:09,960
Speaker 3: can't scale.

852
00:40:12,360 --> 00:40:14,960
Speaker 2: It's basically just reading everything that's already out there, and

853
00:40:15,000 --> 00:40:18,239
Speaker 2: it's basically wall exactly. And now it's got no more

854
00:40:18,280 --> 00:40:22,280
Speaker 2: creativity than unless we have more inputs. That's the creative

855
00:40:22,320 --> 00:40:23,160
Speaker 2: human mind, that's right.

856
00:40:23,200 --> 00:40:24,600
Speaker 3: And then the next one is like, well, we don't

857
00:40:24,600 --> 00:40:26,279
Speaker 3: have enough chips to keep going, and then we don't

858
00:40:26,280 --> 00:40:28,879
Speaker 3: have enough energy. And the point being is that there

859
00:40:28,920 --> 00:40:32,360
Speaker 3: are trillions of dollars going into finding all of the

860
00:40:32,400 --> 00:40:35,680
Speaker 3: solutions to all of these bottlenecks. All the smartest minds

861
00:40:35,680 --> 00:40:38,359
Speaker 3: are going because this is the biggest incentive, because it

862
00:40:38,360 --> 00:40:44,960
Speaker 3: gives you political, economic, military, scientific, technological cyber dominance forever.

863
00:40:45,440 --> 00:40:47,440
Speaker 3: And so if you think that any one of these

864
00:40:47,440 --> 00:40:50,560
Speaker 3: bottlenecks is going to stop that mass sum total of

865
00:40:50,600 --> 00:40:54,680
Speaker 3: like that incentive, it's it's a little delusional. And that's

866
00:40:54,719 --> 00:40:57,080
Speaker 3: why we have to have this clarity that where we're

867
00:40:57,120 --> 00:41:01,040
Speaker 3: going isn't safe for any of us us because that

868
00:41:01,440 --> 00:41:04,200
Speaker 3: is the coordination. That is where we'll start to coordinate differently.

869
00:41:04,280 --> 00:41:09,600
Speaker 1: Now you bring up something, Gavin, that there is a

870
00:41:09,880 --> 00:41:12,520
Speaker 1: belief that some of this is hype, that the companies

871
00:41:12,560 --> 00:41:13,759
Speaker 1: are hyping the technology.

872
00:41:13,800 --> 00:41:16,520
Speaker 2: I've just read a blog and reason there's gonna be

873
00:41:16,560 --> 00:41:19,400
Speaker 2: no job losses. It wasn't Mark himself, but it was

874
00:41:19,440 --> 00:41:21,799
Speaker 2: a member of the team saying, you know, we're back

875
00:41:21,840 --> 00:41:25,000
Speaker 2: to utopian future. Yeah, and we're going to pay abundance

876
00:41:25,160 --> 00:41:29,360
Speaker 2: abounds cost of goods collapses. We find our lives purpose

877
00:41:29,400 --> 00:41:32,000
Speaker 2: and meaning in many different ways, not the quote unquote

878
00:41:32,040 --> 00:41:34,480
Speaker 2: dignity of job. We find other and jobs will be

879
00:41:34,480 --> 00:41:37,360
Speaker 2: a plenty because we can't even conceive of the jobs.

880
00:41:37,400 --> 00:41:39,160
Speaker 1: Two and years ago, we were all farmers here in

881
00:41:39,320 --> 00:41:41,960
Speaker 1: Sacramento and went out there, and we've over.

882
00:41:41,800 --> 00:41:44,719
Speaker 2: Hyped the sectorial versus the general nature of the displacement

883
00:41:45,080 --> 00:41:46,920
Speaker 2: it invariably.

884
00:41:46,840 --> 00:41:48,200
Speaker 1: Always finds something new to do.

885
00:41:48,360 --> 00:41:51,560
Speaker 2: We always more together, exactly more bank tellers.

886
00:41:51,560 --> 00:41:54,799
Speaker 1: Now that that's exactly right. The radiology and Apprehinton made

887
00:41:54,800 --> 00:41:56,640
Speaker 1: the prediction that we're going to have any radiology.

888
00:41:56,680 --> 00:41:58,600
Speaker 2: Why are you so the negative about all this?

889
00:41:58,880 --> 00:42:00,640
Speaker 1: Well, I want to separate to things that we open

890
00:42:00,760 --> 00:42:03,800
Speaker 1: up two cancerforms, so let's take the cance form separately.

891
00:42:04,800 --> 00:42:07,400
Speaker 1: One of them is around whether the companies are hyping

892
00:42:07,480 --> 00:42:10,120
Speaker 1: the power of the technology through talking about the dangers,

893
00:42:10,440 --> 00:42:12,520
Speaker 1: and then the other is whether the hype is going

894
00:42:12,560 --> 00:42:14,479
Speaker 1: to cause the level of job loss. I've heard people

895
00:42:14,640 --> 00:42:15,200
Speaker 1: scarce say.

896
00:42:15,120 --> 00:42:18,080
Speaker 2: That about the mythos it was just wildly overstayed, that's right.

897
00:42:18,320 --> 00:42:19,640
Speaker 2: So it was just a way of hyping up the

898
00:42:19,680 --> 00:42:20,400
Speaker 2: stock in attention.

899
00:42:20,560 --> 00:42:24,200
Speaker 1: That's right. So I really really really want to meet

900
00:42:24,280 --> 00:42:27,920
Speaker 1: that criticism. So people will say, and Thropic has a

901
00:42:28,000 --> 00:42:30,960
Speaker 1: history of hyping the technology, saying it's dangerous, so that

902
00:42:31,040 --> 00:42:33,879
Speaker 1: they can get a regulatory capture, get the company, get

903
00:42:33,880 --> 00:42:36,319
Speaker 1: the government to regulate it. Say there's only one king here,

904
00:42:36,440 --> 00:42:38,719
Speaker 1: make them the king nationalize the project, then shut down

905
00:42:38,760 --> 00:42:41,600
Speaker 1: the other projects. And it's this all secret ploy so

906
00:42:41,640 --> 00:42:45,360
Speaker 1: that they win the race. And first of all, it

907
00:42:45,400 --> 00:42:48,200
Speaker 1: assumes that them talking about the dangers is only bad faith,

908
00:42:48,239 --> 00:42:49,919
Speaker 1: like that that the technology is not dangerous and they're

909
00:42:49,920 --> 00:42:51,319
Speaker 1: just saying that it's dangerous and it can do all

910
00:42:51,320 --> 00:42:53,320
Speaker 1: these destructive things so that they can get that outcome.

911
00:42:53,880 --> 00:42:57,359
Speaker 1: So first of all, let's just take claud Mythos specifically.

912
00:42:57,640 --> 00:43:00,200
Speaker 1: So again, hack in every major operating systems. If you

913
00:43:00,239 --> 00:43:02,680
Speaker 1: read online, there's a lot of people who say, this

914
00:43:02,760 --> 00:43:04,719
Speaker 1: is just hype, This isn't actually that much better. You

915
00:43:04,760 --> 00:43:06,680
Speaker 1: can take the open source models, you can do this stuff.

916
00:43:07,000 --> 00:43:10,040
Speaker 1: So I have a friend who is the head of

917
00:43:10,040 --> 00:43:13,280
Speaker 1: security at one of the top five, not the Fortune

918
00:43:13,280 --> 00:43:17,640
Speaker 1: five hundred, the top five, and he has had early

919
00:43:17,680 --> 00:43:23,200
Speaker 1: access to Mythos, and he himself has said, this is

920
00:43:24,080 --> 00:43:25,720
Speaker 1: this is crazy, this is like, I mean, the words

921
00:43:25,719 --> 00:43:29,120
Speaker 1: were I think I saw Jesus. It was like it

922
00:43:29,200 --> 00:43:33,120
Speaker 1: was that crazy. When he looked at everybody critiquing that

923
00:43:33,160 --> 00:43:35,759
Speaker 1: Mythos was just hype, he asked, has any of them

924
00:43:35,760 --> 00:43:38,200
Speaker 1: do they actually have access to the model? And none

925
00:43:38,200 --> 00:43:40,680
Speaker 1: of the people that had criticized it had personally had

926
00:43:40,719 --> 00:43:46,160
Speaker 1: access to it. So I challenge those who are criticizing

927
00:43:46,200 --> 00:43:48,799
Speaker 1: that it's hype to say, have you actually used it,

928
00:43:49,239 --> 00:43:50,560
Speaker 1: and if you talk to the people who have, do

929
00:43:50,600 --> 00:43:52,439
Speaker 1: you still have the same opinion. Okay, so that's one

930
00:43:52,440 --> 00:43:54,920
Speaker 1: thing that it is true by the way that the companies,

931
00:43:54,920 --> 00:43:57,920
Speaker 1: I think, have wanted people to understand the dangers so

932
00:43:57,960 --> 00:44:00,440
Speaker 1: that they can actually accelerate the move towards some gar ardrails.

933
00:44:00,719 --> 00:44:02,239
Speaker 1: But there's a question of that's that happening in good

934
00:44:02,239 --> 00:44:03,719
Speaker 1: faith or bad faith. I think there's more of that

935
00:44:03,760 --> 00:44:06,680
Speaker 1: happening in good faith. There's some bad faith in there too, maybe,

936
00:44:06,920 --> 00:44:09,640
Speaker 1: but I think it's it's mainly good faith. But then

937
00:44:09,680 --> 00:44:12,120
Speaker 1: let's take the second can of worms who opened, which

938
00:44:12,160 --> 00:44:15,920
Speaker 1: is on is AI going to create this world of abundance?

939
00:44:16,000 --> 00:44:18,520
Speaker 1: We're all going to be poets, you know, and painters

940
00:44:18,520 --> 00:44:21,960
Speaker 1: on a Grecian sunset and you know, now robots all

941
00:44:21,960 --> 00:44:25,920
Speaker 1: the job. Now you're talking, well, we'd all like that.

942
00:44:26,160 --> 00:44:28,360
Speaker 1: One question I would have is when, as a handful

943
00:44:28,480 --> 00:44:34,360
Speaker 1: of people ever concentrated all the wealth and then consciously redistributed.

944
00:44:33,640 --> 00:44:36,640
Speaker 2: It to nine trillionaires, are not going to take care

945
00:44:36,680 --> 00:44:37,839
Speaker 2: of eight or nine billion of us?

946
00:44:37,960 --> 00:44:38,760
Speaker 1: Yeah exactly?

947
00:44:38,880 --> 00:44:40,240
Speaker 2: You call that bluff?

948
00:44:40,360 --> 00:44:42,640
Speaker 1: Yeah exactly. Well, and then you combine that with the

949
00:44:42,680 --> 00:44:45,799
Speaker 1: intelligence curse that we laid out that the incentive is like,

950
00:44:45,840 --> 00:44:47,799
Speaker 1: why do we want to invest in the people? It

951
00:44:47,840 --> 00:44:52,120
Speaker 1: became becomes basically an act of charity because otherwise I

952
00:44:52,120 --> 00:44:56,400
Speaker 1: mean that or dealing with the political revolution. But again

953
00:44:56,480 --> 00:44:58,400
Speaker 1: I don't think that we're currently on track to be

954
00:44:58,440 --> 00:45:01,440
Speaker 1: redistributing that well. And it's also not just the wealth

955
00:45:01,440 --> 00:45:02,759
Speaker 1: and the money. It's like we have people have to

956
00:45:02,760 --> 00:45:05,520
Speaker 1: have work and dignity and status and meaning vulture life.

957
00:45:05,560 --> 00:45:08,399
Speaker 2: You know, job sells. Life's three great evils boredom, vice

958
00:45:08,520 --> 00:45:11,160
Speaker 2: and need. That's just need, but boredom and vice. So

959
00:45:11,200 --> 00:45:11,759
Speaker 2: let's get to.

960
00:45:11,680 --> 00:45:14,239
Speaker 1: The community and belonging absolutely all of.

961
00:45:14,200 --> 00:45:17,239
Speaker 2: That, which again no one's talked about more than you

962
00:45:17,280 --> 00:45:21,520
Speaker 2: in terms of that social dilemma. So let's before we

963
00:45:21,600 --> 00:45:24,600
Speaker 2: go there in this notion of the transition and job

964
00:45:24,640 --> 00:45:27,440
Speaker 2: displacement and the sort of the human condition that I

965
00:45:27,480 --> 00:45:31,120
Speaker 2: think connects as well the President chi President Trump's visit

966
00:45:31,200 --> 00:45:33,880
Speaker 2: as well in terms of you know, their own domestic

967
00:45:33,920 --> 00:45:36,080
Speaker 2: issues in China. That's where they have the same incentive

968
00:45:36,080 --> 00:45:38,759
Speaker 2: structure not to go through that transition with the kind

969
00:45:38,800 --> 00:45:42,040
Speaker 2: of displacement that could create social unrest. That's in the

970
00:45:42,080 --> 00:45:44,919
Speaker 2: short term, but get back to this notion of constraints

971
00:45:44,960 --> 00:45:47,000
Speaker 2: and the safety side of things. I mean, we're here

972
00:45:47,040 --> 00:45:49,840
Speaker 2: in California and the dominant you know, I mean the

973
00:45:49,880 --> 00:45:52,799
Speaker 2: technology sort of birthplace of so much of the technology.

974
00:45:52,800 --> 00:45:55,600
Speaker 2: Obviously the consciousness thirty two of the top fifty market

975
00:45:55,600 --> 00:45:58,640
Speaker 2: cap companies, arguably old stat and of course most of

976
00:45:58,680 --> 00:46:03,000
Speaker 2: the eye labs here. But we're also we've been leaders

977
00:46:03,239 --> 00:46:07,600
Speaker 2: modest though some would suggest, but we've been leaders in

978
00:46:07,719 --> 00:46:11,480
Speaker 2: particularly large language model, frontier models of focusing on a

979
00:46:11,560 --> 00:46:17,400
Speaker 2: regulatory structure in the complete absence of any federal regulation.

980
00:46:17,880 --> 00:46:24,680
Speaker 2: It's the let it rip administration until recently, there were

981
00:46:24,680 --> 00:46:29,560
Speaker 2: some tonal shifts in the Trump administration just the last

982
00:46:29,600 --> 00:46:32,520
Speaker 2: few weeks. They were undermining, they were very intentionally undermining

983
00:46:32,719 --> 00:46:36,759
Speaker 2: the legislation that we brought SB fifty three, and you

984
00:46:36,880 --> 00:46:41,400
Speaker 2: had other Republicans that were trying to undermine a California's

985
00:46:41,480 --> 00:46:44,279
Speaker 2: Leadership Center crews will call them out saying we don't

986
00:46:44,320 --> 00:46:47,400
Speaker 2: want to see the California invocation of regulation all across

987
00:46:47,400 --> 00:46:53,120
Speaker 2: the United States of America. Interestingly, that's beginning to shift. Why

988
00:46:53,120 --> 00:46:55,960
Speaker 2: do you think that's the case? Is it because our

989
00:46:56,960 --> 00:47:00,000
Speaker 2: ars ais are is no longer formally in that role.

990
00:47:01,080 --> 00:47:04,280
Speaker 2: Is it because now they're waking up to this new reality?

991
00:47:04,440 --> 00:47:08,279
Speaker 2: Was it what happened in the Pentagon with Dario and Anthropic?

992
00:47:08,520 --> 00:47:10,640
Speaker 2: Was it the combination of all of this. It's because

993
00:47:10,640 --> 00:47:13,360
Speaker 2: they're listening to to you they watched the doc. Is

994
00:47:13,400 --> 00:47:16,359
Speaker 2: because they realize all the money in the world is

995
00:47:16,360 --> 00:47:18,960
Speaker 2: not going to build a big enough bunker that I

996
00:47:19,000 --> 00:47:23,160
Speaker 2: can enjoy in the absence of societal calm. What is it?

997
00:47:23,840 --> 00:47:25,920
Speaker 3: I mean I think of it to say, very shortly,

998
00:47:26,040 --> 00:47:28,360
Speaker 3: is that there are two different realities. They're sort of

999
00:47:28,400 --> 00:47:31,560
Speaker 3: like the political reality, and then there's like physical reality,

1000
00:47:32,040 --> 00:47:36,239
Speaker 3: and physical reality is crashing into political reality. That is

1001
00:47:36,239 --> 00:47:41,560
Speaker 3: with Mythos. Suddenly banks can get hacked, any computer system

1002
00:47:41,600 --> 00:47:45,760
Speaker 3: can get hacked, your stuff can get hacked. And whence

1003
00:47:45,800 --> 00:47:49,200
Speaker 3: that physical reality starts getting scary enough you have to

1004
00:47:49,239 --> 00:47:51,239
Speaker 3: start waking up. You're no longer in sort of like

1005
00:47:51,239 --> 00:47:52,160
Speaker 3: political game land.

1006
00:47:53,080 --> 00:47:55,719
Speaker 1: I do think that it was. It's interesting to note

1007
00:47:55,719 --> 00:47:59,480
Speaker 1: that when the emergency meeting was convened after Claude Mythos

1008
00:47:59,520 --> 00:48:02,960
Speaker 1: came out, it wasn't at the Pentagon or National Security.

1009
00:48:02,960 --> 00:48:05,000
Speaker 1: I mean that happened too, I'm sure, but the real

1010
00:48:05,040 --> 00:48:07,760
Speaker 1: meeting that happened was between Scott Bessant and the treasure

1011
00:48:07,840 --> 00:48:10,759
Speaker 1: the Treasury Secretary can meeting all the banks because I

1012
00:48:10,760 --> 00:48:12,759
Speaker 1: think the thing that really got them was that if

1013
00:48:12,800 --> 00:48:15,680
Speaker 1: this takes down the financial system, we'll get his quote

1014
00:48:15,680 --> 00:48:18,680
Speaker 1: ten percent GDP growth if the entire financial system gets undermined.

1015
00:48:19,280 --> 00:48:21,719
Speaker 1: So I think this again illustrates the point we have

1016
00:48:21,760 --> 00:48:25,600
Speaker 1: been making since the beginning that the upsides don't prevent

1017
00:48:25,680 --> 00:48:29,160
Speaker 1: the downsides, and the downsides can undermine the world that

1018
00:48:29,200 --> 00:48:33,640
Speaker 1: can sustain and the benefits and of the upsides. And

1019
00:48:33,719 --> 00:48:36,640
Speaker 1: so I do think there's been a forced shift, and

1020
00:48:37,160 --> 00:48:38,480
Speaker 1: you know it's very late in the game, but we

1021
00:48:38,480 --> 00:48:41,680
Speaker 1: should celebrate that it is happening. Now. We just need

1022
00:48:41,719 --> 00:48:44,719
Speaker 1: this kind of full whole of society response to mobilize.

1023
00:48:44,880 --> 00:48:47,000
Speaker 1: I mean, there's a Nicholas Carlini who gave the talk

1024
00:48:47,040 --> 00:48:49,680
Speaker 1: on mythos at a conference black Hat Lelam I think

1025
00:48:49,719 --> 00:48:52,200
Speaker 1: it was called because it was the Unprompted Unprompted Conference,

1026
00:48:53,320 --> 00:48:55,680
Speaker 1: and basically saying he's kind of calling if you know,

1027
00:48:55,680 --> 00:48:58,399
Speaker 1: if you are a cyber person, we need you right now.

1028
00:48:58,400 --> 00:49:00,640
Speaker 1: We need you to defending all the systems. Everybody should

1029
00:49:00,640 --> 00:49:03,240
Speaker 1: get access to mithis and do it as fast as possible, because,

1030
00:49:03,280 --> 00:49:06,680
Speaker 1: as you said, Gavin, the clip between the new capabilities

1031
00:49:06,719 --> 00:49:08,759
Speaker 1: being out there and then China coming out with the

1032
00:49:08,800 --> 00:49:11,440
Speaker 1: model that makes it possible. Maybe this time we have

1033
00:49:11,480 --> 00:49:13,800
Speaker 1: six months. Next time maybe we have three months, and

1034
00:49:13,800 --> 00:49:16,440
Speaker 1: then we have two months. So we need to really

1035
00:49:16,480 --> 00:49:18,640
Speaker 1: work hard. I'm not trying to scare people. It just

1036
00:49:18,680 --> 00:49:21,600
Speaker 1: means that we actually have to work hard to create

1037
00:49:21,600 --> 00:49:24,400
Speaker 1: safety here. Now. China doesn't want the financial system to

1038
00:49:24,400 --> 00:49:27,880
Speaker 1: collapse either, No, they don't, and so again we have

1039
00:49:27,920 --> 00:49:31,920
Speaker 1: to recognize that it's possible for coordination to happen, not

1040
00:49:31,960 --> 00:49:34,520
Speaker 1: because of kumbaya we're all going to get along, but

1041
00:49:34,560 --> 00:49:37,879
Speaker 1: because out of self interest, like the US doesn't want

1042
00:49:37,960 --> 00:49:39,840
Speaker 1: China to screw it up and then break the financial system,

1043
00:49:39,920 --> 00:49:41,400
Speaker 1: China doesn't want the US to screw up and then

1044
00:49:41,440 --> 00:49:44,120
Speaker 1: break the financial system. Or the US doesn't want China

1045
00:49:44,120 --> 00:49:46,799
Speaker 1: to release a ROGUEI that starts mining for cryptocurrency and

1046
00:49:46,840 --> 00:49:49,640
Speaker 1: hacking into things and self replicating like an invasive species,

1047
00:49:50,000 --> 00:49:51,759
Speaker 1: and China doesn't want the US to do that either.

1048
00:49:52,320 --> 00:49:55,720
Speaker 1: So so long as we have clarity about what we want,

1049
00:49:56,640 --> 00:49:57,879
Speaker 1: we can choose a different path.

1050
00:49:58,440 --> 00:50:01,560
Speaker 2: And what a Breton would type past. Yeah.

1051
00:50:01,600 --> 00:50:03,200
Speaker 3: Well, and this is actually one of the other exciting

1052
00:50:03,239 --> 00:50:07,400
Speaker 3: things is that we haven't even really tried to coordinate yet,

1053
00:50:08,120 --> 00:50:10,520
Speaker 3: Like what percentage of the billionaires wealth, how much of

1054
00:50:10,520 --> 00:50:12,719
Speaker 3: their time have they spent actually trying to coordinate. They

1055
00:50:12,760 --> 00:50:15,279
Speaker 3: will just say, well, if you're going to do it,

1056
00:50:15,280 --> 00:50:16,000
Speaker 3: then we're going to do it.

1057
00:50:16,960 --> 00:50:18,919
Speaker 1: If you say it's impossible, have you spent a month

1058
00:50:18,960 --> 00:50:21,680
Speaker 1: of your life and all of your connections dedicatedly trying.

1059
00:50:21,719 --> 00:50:24,360
Speaker 3: Yeah, exactly one and the same thing like this, guys, exactly.

1060
00:50:24,440 --> 00:50:27,480
Speaker 3: And the last time that humanity invented a technology that

1061
00:50:27,560 --> 00:50:32,040
Speaker 3: could extinct ourselves, like the nuclear bomb, we had Brentwoods.

1062
00:50:32,360 --> 00:50:36,200
Speaker 3: You took kind of groups from one hundred countries, lock

1063
00:50:36,239 --> 00:50:38,360
Speaker 3: them in a hotel room in New Hampshire for like

1064
00:50:38,440 --> 00:50:40,400
Speaker 3: six weeks or something like that, and said we're going

1065
00:50:40,440 --> 00:50:41,400
Speaker 3: to figure something out.

1066
00:50:41,280 --> 00:50:43,399
Speaker 1: And you're not leaving the hotel until we figure it out.

1067
00:50:43,480 --> 00:50:45,400
Speaker 1: So it's like it's not a conference where you go,

1068
00:50:45,440 --> 00:50:47,200
Speaker 1: you drink your coffee and you listen to some talks.

1069
00:50:47,320 --> 00:50:50,279
Speaker 1: It's what we need is the we lock ourselves in

1070
00:50:50,320 --> 00:50:54,000
Speaker 1: a room and we figure this out. And I know

1071
00:50:54,040 --> 00:50:55,680
Speaker 1: that this summit is just a couple of days and

1072
00:50:55,800 --> 00:50:57,880
Speaker 1: We all know about the difficulties of the level of

1073
00:50:57,960 --> 00:50:59,920
Speaker 1: expertise that might be involved right now, but this is

1074
00:51:00,160 --> 00:51:02,600
Speaker 1: moment to open the doorway of that possibility.

1075
00:51:03,760 --> 00:51:07,240
Speaker 3: And we have examples through history where when something becomes

1076
00:51:07,320 --> 00:51:12,680
Speaker 3: existential to a citizen group and their nation, they will coordinate.

1077
00:51:12,920 --> 00:51:14,600
Speaker 3: So you know, in the middle of the Cold War,

1078
00:51:14,880 --> 00:51:19,120
Speaker 3: still the US and Russia we coordinated on eradicating smallpox,

1079
00:51:19,200 --> 00:51:22,840
Speaker 3: and the US did logistics and funding, and the Soviet

1080
00:51:22,880 --> 00:51:25,880
Speaker 3: Union made twenty five million doses of the vaccine annually.

1081
00:51:26,280 --> 00:51:28,360
Speaker 3: And then you know India and Pakistan they were literally

1082
00:51:28,360 --> 00:51:31,080
Speaker 3: trading bullets in the nineteen sixties, and yet they still

1083
00:51:31,120 --> 00:51:34,719
Speaker 3: worked on the Indus Water Treaty and that lasted nearly

1084
00:51:34,760 --> 00:51:38,200
Speaker 3: sixty years because access to water was existential to them

1085
00:51:38,239 --> 00:51:38,600
Speaker 3: and their.

1086
00:51:38,480 --> 00:51:41,520
Speaker 1: Shared water supply. You have to collaborate on that. And

1087
00:51:41,560 --> 00:51:44,359
Speaker 1: so I just want to acknowledge the people here. It

1088
00:51:44,440 --> 00:51:46,800
Speaker 1: was you know in people, some people know the history

1089
00:51:46,840 --> 00:51:49,000
Speaker 1: that Obama and she President Obama and She signed an

1090
00:51:49,000 --> 00:51:50,799
Speaker 1: agreement to not cyber hack each other. And I think

1091
00:51:50,840 --> 00:51:52,520
Speaker 1: the next day was the biggest cyber haacket in the

1092
00:51:52,600 --> 00:51:55,560
Speaker 1: US government by China. So I want people to hear

1093
00:51:55,640 --> 00:51:58,680
Speaker 1: this not from some kind of naivete about the level

1094
00:51:58,920 --> 00:52:02,120
Speaker 1: of competition. I will read antagonism that is currently present.

1095
00:52:02,880 --> 00:52:05,040
Speaker 1: But when the stakes get existential, when I push the

1096
00:52:05,080 --> 00:52:06,760
Speaker 1: button and it shifts from the label of the button

1097
00:52:06,760 --> 00:52:09,880
Speaker 1: being ten percent GDP growth and military dominance and cyber dominance,

1098
00:52:10,320 --> 00:52:12,840
Speaker 1: the next time I push the button, it's collective suicide.

1099
00:52:13,520 --> 00:52:15,399
Speaker 1: I don't want to push that button, and China doesn't

1100
00:52:15,400 --> 00:52:18,640
Speaker 1: want to push that button either, So the button label

1101
00:52:18,719 --> 00:52:20,960
Speaker 1: has to shift from what we thought it was going

1102
00:52:20,960 --> 00:52:23,360
Speaker 1: to give us to a new outcome. And with the

1103
00:52:23,360 --> 00:52:25,320
Speaker 1: way ASA says that I love is that the fear

1104
00:52:25,800 --> 00:52:29,080
Speaker 1: of all of us losing has to become greater than

1105
00:52:29,080 --> 00:52:30,200
Speaker 1: the fear of me losing to you.

1106
00:52:30,760 --> 00:52:36,560
Speaker 2: So you have a very regulated construct in China compared

1107
00:52:36,560 --> 00:52:41,040
Speaker 2: to certainly the United States. You're talking about a traditional

1108
00:52:41,040 --> 00:52:46,160
Speaker 2: model right now, where the new frontier out here the

1109
00:52:46,160 --> 00:52:48,399
Speaker 2: wild less. Yeah yeah, I mean it's you know, go west,

1110
00:52:48,960 --> 00:52:51,240
Speaker 2: young man, go West. I mean, it's people are pushing

1111
00:52:51,280 --> 00:52:53,920
Speaker 2: out the boundaries of discovery, holding them back on their own.

1112
00:52:54,920 --> 00:52:59,680
Speaker 2: Regardless of what happens in Beijing, this is really in

1113
00:52:59,680 --> 00:53:03,560
Speaker 2: the hands of a handful of people ultimately making the

1114
00:53:03,640 --> 00:53:08,600
Speaker 2: right decision. We you know, and I believe Dario is

1115
00:53:08,640 --> 00:53:10,920
Speaker 2: the best of a lot. I think universally that's accepted.

1116
00:53:11,000 --> 00:53:14,319
Speaker 2: But that may be, and even Dario may acknowledge because

1117
00:53:14,320 --> 00:53:17,200
Speaker 2: of the flatness of the surrounding terrain, may not be.

1118
00:53:17,200 --> 00:53:20,239
Speaker 2: Because he's particularly eminent on his own and he talks

1119
00:53:20,280 --> 00:53:23,960
Speaker 2: about his own you know. You know, he's an entrepreneur,

1120
00:53:24,320 --> 00:53:27,600
Speaker 2: and he's constantly reflecting on his own incentive structure and

1121
00:53:27,640 --> 00:53:29,360
Speaker 2: how he has to compete in this environment at the

1122
00:53:29,360 --> 00:53:31,120
Speaker 2: same time, and he's at least I think, has the

1123
00:53:31,160 --> 00:53:35,520
Speaker 2: more situational awareness than others. But you know, whatever can

1124
00:53:35,600 --> 00:53:38,600
Speaker 2: be will be. And you know, with respect to Elon,

1125
00:53:38,680 --> 00:53:42,960
Speaker 2: I don't trust x Ai, you know, I mean the

1126
00:53:43,000 --> 00:53:45,760
Speaker 2: idea that he's the good guy compared to our friends

1127
00:53:45,800 --> 00:53:48,840
Speaker 2: down at Google or you guys left. I mean, you know,

1128
00:53:49,840 --> 00:53:52,680
Speaker 2: so talk to me about more of the sinister realities

1129
00:53:52,680 --> 00:53:54,279
Speaker 2: of you know, I don't mean that sinister in the

1130
00:53:54,600 --> 00:53:58,800
Speaker 2: but but the impulses again to be the guy, the god,

1131
00:53:59,040 --> 00:54:00,600
Speaker 2: the god. May I mean to have their DNA.

1132
00:54:00,920 --> 00:54:03,520
Speaker 1: Yeah, So I'm gratefully you're bringing this up. There's a

1133
00:54:03,520 --> 00:54:06,719
Speaker 1: few things we should enumerate here. So one is we

1134
00:54:06,760 --> 00:54:10,400
Speaker 1: need coordination, and we do find ourselves in the unfortunate

1135
00:54:10,440 --> 00:54:12,719
Speaker 1: spot that the people who do need to be in

1136
00:54:13,000 --> 00:54:16,960
Speaker 1: coordination maximally distrust each other. Even just the US CEOs,

1137
00:54:17,160 --> 00:54:19,440
Speaker 1: I mean Elon Musk and Sam hate each.

1138
00:54:19,320 --> 00:54:28,359
Speaker 2: Other, correct, and so the India they couldn't even hold

1139
00:54:28,600 --> 00:54:29,560
Speaker 2: at least they showed up.

1140
00:54:29,640 --> 00:54:33,719
Speaker 1: So we have a problem of trust between the leaders themselves.

1141
00:54:33,960 --> 00:54:36,920
Speaker 1: We need I think structures that impose the trust on

1142
00:54:37,040 --> 00:54:39,360
Speaker 1: top because they're not going to do it autonomously themselves.

1143
00:54:40,400 --> 00:54:41,480
Speaker 2: So that's the regulation.

1144
00:54:41,560 --> 00:54:44,359
Speaker 1: That's yeah, that's the trans using the power of law

1145
00:54:44,640 --> 00:54:48,319
Speaker 1: to say that these people, it's going to happen in China.

1146
00:54:48,360 --> 00:54:50,080
Speaker 1: Here's the rule is going to be similar in China

1147
00:54:50,120 --> 00:54:52,319
Speaker 1: as they are here. We're both, for example, not going

1148
00:54:52,320 --> 00:54:54,439
Speaker 1: to open source a model that can happen to any

1149
00:54:54,440 --> 00:54:56,880
Speaker 1: computer system in the world without defenses, at least not

1150
00:54:56,920 --> 00:54:59,120
Speaker 1: going to open source that. China doesn't want a rogue

1151
00:54:59,160 --> 00:55:02,840
Speaker 1: state act, nonstate actor, or terrorist group having that ability

1152
00:55:02,880 --> 00:55:05,560
Speaker 1: to hack their infrastructure because it would also blow back

1153
00:55:05,560 --> 00:55:08,360
Speaker 1: onto them. Same thing with a open source model that

1154
00:55:08,440 --> 00:55:11,480
Speaker 1: knows how to do very dangerous things with biology. There's

1155
00:55:11,520 --> 00:55:14,200
Speaker 1: some threshold that we can get these countries to agree,

1156
00:55:14,680 --> 00:55:16,799
Speaker 1: just like the Soviet Union in the United States had

1157
00:55:16,800 --> 00:55:20,080
Speaker 1: a red phone saying we're gonna this is to de escalate.

1158
00:55:20,560 --> 00:55:23,240
Speaker 1: I think we need something like an AI red lines phone,

1159
00:55:23,280 --> 00:55:26,400
Speaker 1: meaning that both countries have common knowledge of the frontier

1160
00:55:26,400 --> 00:55:28,279
Speaker 1: of these risks, because right now we don't even have

1161
00:55:28,280 --> 00:55:28,920
Speaker 1: that common notion, and.

1162
00:55:28,880 --> 00:55:30,920
Speaker 2: There's rumors that that may be one of the things

1163
00:55:30,960 --> 00:55:34,320
Speaker 2: that that are prepared to announce. That's right, you know,

1164
00:55:34,440 --> 00:55:37,920
Speaker 2: at least some beginning of this.

1165
00:55:38,040 --> 00:55:40,200
Speaker 1: But I want the other aspect of the human experience

1166
00:55:40,239 --> 00:55:41,879
Speaker 1: of this. We did a screening of the AI doc

1167
00:55:42,080 --> 00:55:44,120
Speaker 1: in New York and there is someone in the audience

1168
00:55:44,160 --> 00:55:48,319
Speaker 1: actually who raised her hand quietly and she said, I'm

1169
00:55:48,360 --> 00:55:51,600
Speaker 1: a coach for one of the CEOs of these companies,

1170
00:55:52,400 --> 00:55:54,640
Speaker 1: and you know what happens when I talk to them

1171
00:55:54,920 --> 00:55:57,200
Speaker 1: is they say, but what can I do? I'm just

1172
00:55:57,239 --> 00:56:00,480
Speaker 1: one person. I'm powerless. And I want people to hear that,

1173
00:56:00,520 --> 00:56:03,160
Speaker 1: because I noticed that, even you know, we all feel

1174
00:56:03,200 --> 00:56:05,239
Speaker 1: relative to the size of this problem, you will never

1175
00:56:05,400 --> 00:56:09,400
Speaker 1: locate agency that is that is enough agency to do

1176
00:56:09,400 --> 00:56:11,920
Speaker 1: something about this problem in one human body, even if

1177
00:56:11,920 --> 00:56:15,880
Speaker 1: that body is elon by himself or Sindar by himself,

1178
00:56:15,960 --> 00:56:20,239
Speaker 1: or Sam by themselves, and so getting back to you know,

1179
00:56:20,280 --> 00:56:22,360
Speaker 1: I think what this moment is inviting us into. We

1180
00:56:22,400 --> 00:56:25,120
Speaker 1: often say that AI is are ultimate test, but greatest

1181
00:56:25,160 --> 00:56:28,200
Speaker 1: invitation that we have to go from agency to reegency.

1182
00:56:28,280 --> 00:56:30,719
Speaker 1: We have to basically act in some kind of collective way.

1183
00:56:31,280 --> 00:56:33,319
Speaker 1: And the forces politically of the world have been driving

1184
00:56:33,400 --> 00:56:35,719
Speaker 1: us away from that. But this is kind of the test,

1185
00:56:35,800 --> 00:56:38,600
Speaker 1: like we either do that and we step up and

1186
00:56:38,680 --> 00:56:41,480
Speaker 1: again we need everything from common political pressure and this

1187
00:56:41,520 --> 00:56:43,600
Speaker 1: being the number one issue in the midterms, and you know,

1188
00:56:43,640 --> 00:56:46,160
Speaker 1: the public's rallying and all the governors you know, speaking

1189
00:56:46,200 --> 00:56:47,839
Speaker 1: up about this, and all the world leaders speaking out.

1190
00:56:47,840 --> 00:56:49,640
Speaker 1: That's just Yet. You know, two days ago I got

1191
00:56:49,640 --> 00:56:52,319
Speaker 1: an email from the President of Iceland who basically wants

1192
00:56:52,360 --> 00:56:54,880
Speaker 1: to activate on this issue in Iceland hosted in Rykovic

1193
00:56:54,960 --> 00:56:58,240
Speaker 1: the first Arms Control talks. There's a lot that people

1194
00:56:58,280 --> 00:57:00,560
Speaker 1: could do if they said not just what can I do,

1195
00:57:01,080 --> 00:57:03,719
Speaker 1: but how could I get my reaching up and out

1196
00:57:03,719 --> 00:57:07,680
Speaker 1: to the network of people to take action together. There

1197
00:57:07,719 --> 00:57:10,120
Speaker 1: is a second part to your point, which is the

1198
00:57:10,200 --> 00:57:14,280
Speaker 1: darker part, which is you talking about the game theory

1199
00:57:14,480 --> 00:57:17,800
Speaker 1: of the psychology of the leaders that they basically believe

1200
00:57:17,920 --> 00:57:20,480
Speaker 1: that in the worst case scenario The thing that kept

1201
00:57:20,560 --> 00:57:24,400
Speaker 1: us safe in nukes is that it's two people who

1202
00:57:24,400 --> 00:57:26,640
Speaker 1: have to push the button. And I know you won't

1203
00:57:26,680 --> 00:57:29,160
Speaker 1: push the button because I know that there's something sacred

1204
00:57:29,640 --> 00:57:31,800
Speaker 1: that you don't want this whole thing to end. And

1205
00:57:31,880 --> 00:57:33,120
Speaker 1: I know that, and I know that you know that.

1206
00:57:33,160 --> 00:57:35,400
Speaker 1: I know that. And so even though we get very

1207
00:57:35,480 --> 00:57:37,480
Speaker 1: very very close, and we've gotten close so many times,

1208
00:57:38,080 --> 00:57:41,400
Speaker 1: we haven't pushed that button. But in this case with AI,

1209
00:57:41,680 --> 00:57:43,840
Speaker 1: there's a belief. First of all, it's a red zone

1210
00:57:43,880 --> 00:57:45,840
Speaker 1: of where the risk occurs. It's not like there's one

1211
00:57:45,880 --> 00:57:47,920
Speaker 1: button that gets pushed. It's like we just push this

1212
00:57:47,920 --> 00:57:50,920
Speaker 1: stuff out there, and there's a belief that it's inevitable.

1213
00:57:50,920 --> 00:57:53,400
Speaker 1: If I didn't do it, someone else would, which means

1214
00:57:53,480 --> 00:57:57,560
Speaker 1: I don't experience ethical complicity in being part of the

1215
00:57:57,640 --> 00:58:00,200
Speaker 1: end of civilization. And if it's inevitable, there's thing I

1216
00:58:00,200 --> 00:58:01,360
Speaker 1: could have done to stop it, so I don't even

1217
00:58:01,440 --> 00:58:03,919
Speaker 1: have to feel bad, right, And then so the game

1218
00:58:03,960 --> 00:58:06,080
Speaker 1: theory goes from all of us knowing that we want

1219
00:58:06,080 --> 00:58:08,960
Speaker 1: to avoid the bad outcome to everybody believing that there

1220
00:58:09,000 --> 00:58:11,920
Speaker 1: isn't a different outcome, which means that the best outcome

1221
00:58:12,040 --> 00:58:13,760
Speaker 1: is maybe the worst thing is all of us go

1222
00:58:14,400 --> 00:58:17,320
Speaker 1: by the wayside, but we birth the digital God and

1223
00:58:17,360 --> 00:58:20,440
Speaker 1: it speaks Chinese instead of English, or maybe it has

1224
00:58:20,480 --> 00:58:22,120
Speaker 1: Elon's DNA. Instead of saying just.

1225
00:58:22,320 --> 00:58:24,320
Speaker 2: Said, it's at least on an American stacks.

1226
00:58:24,880 --> 00:58:27,680
Speaker 1: At least we're selling the world American your media. But

1227
00:58:27,760 --> 00:58:30,600
Speaker 1: the reason of laying all that out is that if

1228
00:58:30,800 --> 00:58:33,200
Speaker 1: the whole world could see I think what we just

1229
00:58:33,280 --> 00:58:37,080
Speaker 1: laid out, if literally everyone could see that, the whole

1230
00:58:37,120 --> 00:58:40,200
Speaker 1: world says, we don't want eight soon to be trillionaires

1231
00:58:40,480 --> 00:58:43,000
Speaker 1: deciding the future for eight billion people who didn't consent

1232
00:58:43,040 --> 00:58:43,360
Speaker 1: to this.

1233
00:58:44,560 --> 00:58:46,600
Speaker 2: And that's the purpose of Again, that sort of brings

1234
00:58:46,640 --> 00:58:48,919
Speaker 2: us back to the beginning. Why you're doing this damn film?

1235
00:58:48,920 --> 00:58:52,320
Speaker 2: That's right after the great and sort of global consciousness.

1236
00:58:52,320 --> 00:58:52,720
Speaker 1: That's right.

1237
00:58:53,000 --> 00:58:56,000
Speaker 2: So in the absence of that, we're back here in California. Yes,

1238
00:58:56,000 --> 00:58:58,640
Speaker 2: here here with the governor, Yes, the current governor, the

1239
00:58:58,680 --> 00:59:03,480
Speaker 2: governor's mansion. Interesting, we're doing some decent things. What more

1240
00:59:03,480 --> 00:59:06,800
Speaker 2: should I be doing in the absence of the kind

1241
00:59:06,800 --> 00:59:10,040
Speaker 2: of federal leadership that we need. I feel like we've

1242
00:59:10,080 --> 00:59:13,040
Speaker 2: lost this last eighteen months, you know, I just we

1243
00:59:13,040 --> 00:59:14,840
Speaker 2: were mo Look, we did an a open It was

1244
00:59:14,880 --> 00:59:18,200
Speaker 2: interesting working I worked very closely with the Biden administration.

1245
00:59:18,240 --> 00:59:21,440
Speaker 2: Did they move quickly enough? Perhaps not, but at least

1246
00:59:21,480 --> 00:59:24,080
Speaker 2: we had a framework of an of an executive order.

1247
00:59:24,120 --> 00:59:26,880
Speaker 2: We moved that forward. We President's sided here at the

1248
00:59:26,880 --> 00:59:30,600
Speaker 2: Fairmount Hotel in San Francisco in California. It was built

1249
00:59:30,640 --> 00:59:34,160
Speaker 2: off an executive order that I did six months later.

1250
00:59:34,200 --> 00:59:36,760
Speaker 2: We were working hand in glove with the Biden administration

1251
00:59:37,160 --> 00:59:39,760
Speaker 2: on that. That was ripped up right when the Trump

1252
00:59:39,800 --> 00:59:43,880
Speaker 2: administration came in to office. You have an ai zar

1253
00:59:44,280 --> 00:59:47,040
Speaker 2: out of the Bay Area certainly understands the ecosystem. But

1254
00:59:47,080 --> 00:59:48,560
Speaker 2: it seemed to me, and this is me and you

1255
00:59:48,640 --> 00:59:50,760
Speaker 2: guys don't have to respond, but it was the great grift.

1256
00:59:50,800 --> 00:59:53,440
Speaker 2: Everybody was sort of on on the train and seeing

1257
00:59:53,440 --> 00:59:56,480
Speaker 2: this as an opportunity and looking at the abundance of

1258
00:59:56,520 --> 00:59:59,520
Speaker 2: this only but not looking at safety, not looking at

1259
00:59:59,600 --> 01:00:03,640
Speaker 2: the risk. As you've described, California cited to assert itself

1260
01:00:03,640 --> 01:00:06,040
Speaker 2: in that respect, as we've done on privacy, as we've

1261
01:00:06,080 --> 01:00:08,439
Speaker 2: done a lot of child safety issues, and a lot

1262
01:00:08,480 --> 01:00:10,360
Speaker 2: more work to do there. And this year will be

1263
01:00:10,600 --> 01:00:12,480
Speaker 2: a landmark year in terms of getting to the next

1264
01:00:12,560 --> 01:00:16,080
Speaker 2: level in that respect. But what more can the state

1265
01:00:16,160 --> 01:00:18,960
Speaker 2: be doing the fear of that always is patchwork, not

1266
01:00:19,000 --> 01:00:23,680
Speaker 2: a framework for the nation. And how you support innovation

1267
01:00:24,400 --> 01:00:27,040
Speaker 2: our own GDP growth which has been off the charts

1268
01:00:27,080 --> 01:00:31,000
Speaker 2: in California visa VR competitors. At the same time addressed

1269
01:00:31,040 --> 01:00:34,960
Speaker 2: these larger global issues. Do you have any specific ideas

1270
01:00:35,200 --> 01:00:38,280
Speaker 2: for a governor of California, that is the current governor

1271
01:00:38,640 --> 01:00:40,920
Speaker 2: that has a budget that he's releaseding in weeks and

1272
01:00:40,960 --> 01:00:43,600
Speaker 2: a legislative session coming up in the next few months.

1273
01:00:44,680 --> 01:00:47,720
Speaker 3: Well, we'll get to answer that question specifically, but what

1274
01:00:47,960 --> 01:00:50,120
Speaker 3: are the places of also good news? I just want

1275
01:00:50,160 --> 01:00:52,160
Speaker 3: to say, because there's a lot of value in just

1276
01:00:52,200 --> 01:00:54,480
Speaker 3: social signaling where everyone knows that there's a problem. We

1277
01:00:54,560 --> 01:00:57,880
Speaker 3: get to that shared common knowledge, common feeling. And if

1278
01:00:57,880 --> 01:01:01,040
Speaker 3: you went back two years and you said by today,

1279
01:01:01,160 --> 01:01:03,880
Speaker 3: twenty five percent of the world's population would live in

1280
01:01:03,920 --> 01:01:07,320
Speaker 3: a country where they've either have announced or enacted a

1281
01:01:07,440 --> 01:01:11,600
Speaker 3: ban for social under sixteen, you'd be like, that's ridiculous.

1282
01:01:11,640 --> 01:01:12,640
Speaker 3: You couldn't possibly get that.

1283
01:01:12,640 --> 01:01:14,680
Speaker 1: I want people to really feel that. There you are

1284
01:01:14,680 --> 01:01:17,200
Speaker 1: in twenty twenty two. If you said even just literally

1285
01:01:17,200 --> 01:01:19,280
Speaker 1: three years ago, yeah, I know that a quarter of

1286
01:01:19,320 --> 01:01:24,360
Speaker 1: the world's population we're talking Australia, India, Denmark, Denmark's. Two

1287
01:01:24,360 --> 01:01:26,560
Speaker 1: weeks ago, Greece added the list. France. I was with

1288
01:01:26,800 --> 01:01:28,960
Speaker 1: John Hyde and when he was in Davos and he

1289
01:01:29,400 --> 01:01:31,640
Speaker 1: met with President Macrone and got France on board. Like

1290
01:01:32,080 --> 01:01:34,640
Speaker 1: people would have thought that was impossible. I mean, the

1291
01:01:34,720 --> 01:01:35,560
Speaker 1: trains left the station.

1292
01:01:35,680 --> 01:01:37,600
Speaker 2: We just had all the democratic governors out here, and

1293
01:01:37,640 --> 01:01:39,480
Speaker 2: everyone was trying to compare them to trast which one

1294
01:01:39,480 --> 01:01:43,160
Speaker 2: of the states, which we're going to do. But I

1295
01:01:43,200 --> 01:01:45,120
Speaker 2: mean it's you're right, this is a tipping point.

1296
01:01:45,120 --> 01:01:47,440
Speaker 1: It's taking the train and once you get twenty five percent,

1297
01:01:47,480 --> 01:01:48,720
Speaker 1: you're going to get the rest of the world.

1298
01:01:48,920 --> 01:01:50,440
Speaker 2: And the point there was a lot of these companies.

1299
01:01:50,480 --> 01:01:53,680
Speaker 2: I also I anticipate these companies and by the way,

1300
01:01:53,720 --> 01:01:57,120
Speaker 2: if I was advising these companies get ahead of this train,

1301
01:01:57,280 --> 01:02:00,840
Speaker 2: ye and show your large s and that your maturity

1302
01:02:00,880 --> 01:02:04,680
Speaker 2: and understanding. So I imagine that may happen as well.

1303
01:02:05,240 --> 01:02:07,040
Speaker 2: But no, that's a point. You're right.

1304
01:02:07,120 --> 01:02:09,840
Speaker 3: And so right now, something really interesting happened in Hongjo

1305
01:02:09,960 --> 01:02:12,480
Speaker 3: in China. Maybe you're aware of it, but it's this

1306
01:02:12,560 --> 01:02:16,320
Speaker 3: first case where there was someone who lost his job

1307
01:02:16,480 --> 01:02:20,360
Speaker 3: at Ai Automation and this court ruled and said, actually,

1308
01:02:21,040 --> 01:02:23,200
Speaker 3: that's not a legitimate reason for you to lose your job.

1309
01:02:23,680 --> 01:02:28,160
Speaker 3: Companies are not allowed to fire you from like increased automation.

1310
01:02:28,720 --> 01:02:31,680
Speaker 3: Well that solved the whole problem. Probably not, but imagine

1311
01:02:31,720 --> 01:02:34,440
Speaker 3: that California took a leadership position there and said, actually,

1312
01:02:34,440 --> 01:02:37,400
Speaker 3: there are going to be some very serious protections that

1313
01:02:37,720 --> 01:02:43,760
Speaker 3: as AI increases, GDP increases profits, Actually, like people are

1314
01:02:43,880 --> 01:02:46,520
Speaker 3: going to be it's sort of like an employment insurance.

1315
01:02:46,560 --> 01:02:48,800
Speaker 3: They're going to be able to keep their jobs. Like

1316
01:02:49,000 --> 01:02:51,920
Speaker 3: that would set a social signal for the rest of

1317
01:02:52,000 --> 01:02:52,560
Speaker 3: the US.

1318
01:02:52,560 --> 01:02:54,440
Speaker 2: So we'll let you have the hook on the larger

1319
01:02:55,440 --> 01:03:00,919
Speaker 2: safety risk regulation and let's go back to ideas along

1320
01:03:00,960 --> 01:03:03,400
Speaker 2: these lines, because right in front of us this notion

1321
01:03:03,800 --> 01:03:06,720
Speaker 2: of displacement transition, and it goes back to their earlier

1322
01:03:06,760 --> 01:03:10,120
Speaker 2: point I was making. Some will argue that, you know,

1323
01:03:10,160 --> 01:03:13,840
Speaker 2: we always you know, lud eights and will never see

1324
01:03:13,880 --> 01:03:16,240
Speaker 2: the abundance on the other side, and we can't even

1325
01:03:16,320 --> 01:03:20,120
Speaker 2: conceive of the jobs. So with humility, let's not just

1326
01:03:20,200 --> 01:03:23,959
Speaker 2: assume there'll be no human jobs, because the human mind

1327
01:03:24,080 --> 01:03:27,439
Speaker 2: is the capacity that's limitless. With these technologies, that will

1328
01:03:27,480 --> 01:03:31,520
Speaker 2: be more supportive and allow us to be augmented and

1329
01:03:31,600 --> 01:03:35,479
Speaker 2: discover talents and capacity we never thought possible. So let's

1330
01:03:35,480 --> 01:03:38,560
Speaker 2: assume that happens. But the concern is that it seems

1331
01:03:38,600 --> 01:03:42,480
Speaker 2: the most universal that it may happen very fast. The

1332
01:03:42,640 --> 01:03:46,479
Speaker 2: impacts is, so, how do you then flatten the curve

1333
01:03:47,320 --> 01:03:49,520
Speaker 2: the curve? How do we address the transition You get

1334
01:03:49,520 --> 01:03:52,440
Speaker 2: to employment insurance, something we've been talking a lot about.

1335
01:03:52,560 --> 01:03:54,520
Speaker 2: You get to this notion that you can't fire someone

1336
01:03:54,520 --> 01:03:57,080
Speaker 2: to be automated. That's even deeper, and that's interesting. This

1337
01:03:57,200 --> 01:04:00,800
Speaker 2: Chinese example tell me more about the issue of and

1338
01:04:00,960 --> 01:04:03,160
Speaker 2: what I should be worried about. When I see Dario

1339
01:04:03,240 --> 01:04:06,680
Speaker 2: saying fifty percent entry level jobs, it's no longer a

1340
01:04:06,720 --> 01:04:09,080
Speaker 2: career ladder. It's a jungle gym. And all these young

1341
01:04:09,120 --> 01:04:12,439
Speaker 2: folks got a Stanford or like now I'm unemployed or unemployable,

1342
01:04:12,760 --> 01:04:15,400
Speaker 2: no code or software. I mean, give me a women

1343
01:04:15,520 --> 01:04:18,400
Speaker 2: disproportion being impacted in the workforce When you look at

1344
01:04:18,400 --> 01:04:21,520
Speaker 2: those clerical jobs, admin jobs, et cetera. I mean, what,

1345
01:04:21,520 --> 01:04:24,720
Speaker 2: what's what do you see a year two years from now?

1346
01:04:24,960 --> 01:04:27,680
Speaker 2: As we deal with the Holy Grail, the god complex

1347
01:04:27,920 --> 01:04:30,240
Speaker 2: of Agi in the displacement space.

1348
01:04:32,600 --> 01:04:35,160
Speaker 1: Reed Hoffman has this idea. When mefre I get it right,

1349
01:04:35,520 --> 01:04:38,360
Speaker 1: I like, which is one of the things that makes

1350
01:04:38,400 --> 01:04:40,680
Speaker 1: AI distinct. And Sam Altman has talked about this himself,

1351
01:04:41,080 --> 01:04:43,800
Speaker 1: that you could get the age of these unicorn companies.

1352
01:04:43,800 --> 01:04:45,920
Speaker 1: A unicorn company meeting a billion dollar valuation.

1353
01:04:46,160 --> 01:04:49,000
Speaker 2: I met a guy the other day literally billion dollar valuation,

1354
01:04:49,600 --> 01:04:53,360
Speaker 2: tim him exactly, So I think that he was praying

1355
01:04:53,400 --> 01:04:54,920
Speaker 2: himself around as the it's.

1356
01:04:54,800 --> 01:04:56,680
Speaker 1: Me, that's right, one of the first, and so that's

1357
01:04:56,680 --> 01:04:58,640
Speaker 1: what that's what has been positive ever since the beginning

1358
01:04:58,640 --> 01:05:00,440
Speaker 1: of AI. They're saying, we're going to start having world.

1359
01:05:00,640 --> 01:05:03,080
Speaker 1: You're going to have a single person with a unicorn

1360
01:05:03,080 --> 01:05:06,680
Speaker 1: billion dollar company. Yeah, Now does this Does society work?

1361
01:05:07,200 --> 01:05:09,520
Speaker 1: If there's a handful of single people with billion dollar

1362
01:05:09,520 --> 01:05:11,760
Speaker 1: companies and no one else has a job, it doesn't work.

1363
01:05:12,040 --> 01:05:13,520
Speaker 1: Do you think those people just want to live out

1364
01:05:13,520 --> 01:05:16,840
Speaker 1: their lives in bunkers with private militaries and gas masks

1365
01:05:16,880 --> 01:05:19,439
Speaker 1: because they've created that world. I don't think they want

1366
01:05:19,440 --> 01:05:21,680
Speaker 1: that world. So reied Hoffman, who's the founder of LinkedIn,

1367
01:05:21,840 --> 01:05:23,880
Speaker 1: was early at PayPal and you know, was at some

1368
01:05:23,880 --> 01:05:27,160
Speaker 1: point a friend of Peter Thiel's, you know, has this

1369
01:05:27,200 --> 01:05:30,000
Speaker 1: proposal that we can tax companies based on the proportion

1370
01:05:30,240 --> 01:05:33,440
Speaker 1: ratio of how many employees they have relative to their revenue.

1371
01:05:33,760 --> 01:05:39,840
Speaker 1: So you want to basically disincentivize the single solo unicorn company.

1372
01:05:39,840 --> 01:05:40,800
Speaker 1: I don't want to add to it.

1373
01:05:41,000 --> 01:05:41,240
Speaker 2: Yeah.

1374
01:05:41,280 --> 01:05:43,480
Speaker 3: Well, and also just just to note that it doesn't

1375
01:05:43,480 --> 01:05:47,400
Speaker 3: stop with just single person unicorns. Yeah, with automating that

1376
01:05:47,440 --> 01:05:49,840
Speaker 3: final one person not so hard, So you're gonna end

1377
01:05:49,920 --> 01:05:50,920
Speaker 3: up with zero person.

1378
01:05:50,920 --> 01:05:52,840
Speaker 1: You can have a CEO, B and AI. Yeah, and

1379
01:05:52,840 --> 01:05:54,760
Speaker 1: that's actually happening. We're already getting ais that are on

1380
01:05:54,800 --> 01:05:56,680
Speaker 1: boards and things like this. Yeah, and so even if

1381
01:05:56,680 --> 01:05:58,800
Speaker 1: you might find it questionable that that person might have again,

1382
01:05:58,840 --> 01:06:02,040
Speaker 1: they can. They can earn their wealth, but you have

1383
01:06:02,080 --> 01:06:04,880
Speaker 1: to have some taxation to make sure this is being stipid.

1384
01:06:04,920 --> 01:06:07,160
Speaker 1: And we also need we need to find ways of

1385
01:06:07,200 --> 01:06:10,240
Speaker 1: having universal basic ownership, not just universal basic cash payments

1386
01:06:10,240 --> 01:06:13,080
Speaker 1: and UBI, but universal basic ownership. I think people need

1387
01:06:13,120 --> 01:06:15,280
Speaker 1: to have a stake in the success that's happening, like

1388
01:06:15,280 --> 01:06:17,960
Speaker 1: what Norway did with the Sovereign Wealth Fund. But oil

1389
01:06:18,080 --> 01:06:20,440
Speaker 1: didn't oil produce this kind of gravy on top for

1390
01:06:20,480 --> 01:06:23,440
Speaker 1: the civilization. People still had jobs. So what's different about

1391
01:06:23,440 --> 01:06:25,320
Speaker 1: this is we do need to find ways of doing

1392
01:06:25,440 --> 01:06:29,200
Speaker 1: universal basic work. We also need to find ways of

1393
01:06:29,720 --> 01:06:33,280
Speaker 1: having certain professions in which that embodied wisdom, like a

1394
01:06:33,320 --> 01:06:36,360
Speaker 1: surgeon or a senior lawyer or a senior judge. We

1395
01:06:36,400 --> 01:06:39,600
Speaker 1: need ways of training and apprenticing almost like minimum quotas

1396
01:06:39,600 --> 01:06:42,160
Speaker 1: of those kinds of occupations and roles in society to

1397
01:06:42,200 --> 01:06:44,440
Speaker 1: make sure that we have that ongoing knowledge. Because again

1398
01:06:44,920 --> 01:06:47,080
Speaker 1: the short term benefit of like no one needs lawyers

1399
01:06:47,080 --> 01:06:49,800
Speaker 1: and then the senior lawyers all die out, that world

1400
01:06:49,840 --> 01:06:50,400
Speaker 1: doesn't work.

1401
01:06:50,960 --> 01:06:53,800
Speaker 3: So just the last thing to sort of add here

1402
01:06:53,880 --> 01:06:56,560
Speaker 3: is that you instead of getting into arguments about how

1403
01:06:56,680 --> 01:06:58,600
Speaker 3: quickly exactly are people going to lose the jobs, well,

1404
01:06:58,640 --> 01:07:01,200
Speaker 3: people really do lose their jobs. Let's let's plan and say, okay,

1405
01:07:01,280 --> 01:07:03,360
Speaker 3: we think people are going to be out of livelihoods.

1406
01:07:03,600 --> 01:07:05,360
Speaker 3: This is this idea we really came to us from

1407
01:07:05,400 --> 01:07:08,240
Speaker 3: read haasting to see Pharmacy and Pharmacy Netflix, where we said,

1408
01:07:08,440 --> 01:07:11,560
Speaker 3: let's set up trigger point laws. If in unemployment hits

1409
01:07:11,560 --> 01:07:12,800
Speaker 3: ten percent, what are we going to do if it

1410
01:07:12,880 --> 01:07:15,280
Speaker 3: hits twenty percent? What are we going to do? You

1411
01:07:15,320 --> 01:07:18,080
Speaker 3: could pre set up those sort of conditions so we

1412
01:07:18,080 --> 01:07:20,040
Speaker 3: don't have to argue about whether it's going to happen,

1413
01:07:20,080 --> 01:07:21,160
Speaker 3: just what we should do when it does.

1414
01:07:21,400 --> 01:07:23,560
Speaker 1: And I know that you've been running with Engaged California,

1415
01:07:23,720 --> 01:07:27,120
Speaker 1: these citizen deliberations, these ways of you know, aggregating citizen assemblies,

1416
01:07:27,360 --> 01:07:30,000
Speaker 1: having citizens actually deal with and think about these issues

1417
01:07:30,000 --> 01:07:32,640
Speaker 1: and come to some you know, have their own input

1418
01:07:32,640 --> 01:07:35,440
Speaker 1: in this process. But by reckoning with these facts, and

1419
01:07:35,640 --> 01:07:37,280
Speaker 1: I think that these are all things that we need.

1420
01:07:37,320 --> 01:07:40,040
Speaker 1: You know, the countries that discover a natural resource that

1421
01:07:40,120 --> 01:07:43,480
Speaker 1: didn't have this engaged, well educated citizen kind of engage

1422
01:07:43,560 --> 01:07:46,760
Speaker 1: this in infrastructure, like Venezuela or Libya or something like that,

1423
01:07:47,080 --> 01:07:49,400
Speaker 1: they don't do so well. But countries like Norway where

1424
01:07:49,440 --> 01:07:52,800
Speaker 1: you did have the engaged citizens, with oversight of those funds,

1425
01:07:53,080 --> 01:07:54,640
Speaker 1: you end up with a healthier society.

1426
01:07:54,720 --> 01:07:57,000
Speaker 3: So it's sort of like an a California engagement fund

1427
01:07:57,000 --> 01:08:00,000
Speaker 3: because you want like people involved in the redistribution.

1428
01:08:00,320 --> 01:08:00,680
Speaker 1: That's right.

1429
01:08:01,120 --> 01:08:04,600
Speaker 2: So it's interesting that you mentioned mencome which the old

1430
01:08:04,640 --> 01:08:09,919
Speaker 2: Canadian construct ubi university basic income. You had Elon Musk

1431
01:08:09,960 --> 01:08:12,840
Speaker 2: the other day say he wants university basic high income.

1432
01:08:12,960 --> 01:08:14,439
Speaker 1: He said there is going to be univers of basic

1433
01:08:14,480 --> 01:08:15,080
Speaker 1: high income, and.

1434
01:08:15,040 --> 01:08:18,679
Speaker 2: Then was asked with Peter, right, exactly I saw him probably,

1435
01:08:18,680 --> 01:08:20,040
Speaker 2: and he was asked simple question, how are you going

1436
01:08:20,080 --> 01:08:20,400
Speaker 2: to do it?

1437
01:08:20,439 --> 01:08:22,000
Speaker 1: And he said, oh I did. I was just joking.

1438
01:08:22,000 --> 01:08:22,559
Speaker 1: I made it up.

1439
01:08:22,760 --> 01:08:23,679
Speaker 2: That was comforting.

1440
01:08:23,840 --> 01:08:25,680
Speaker 1: Yeah, I think people should really take note of that.

1441
01:08:25,920 --> 01:08:29,160
Speaker 3: Yes, it was just a mispronouncing of high not as

1442
01:08:29,200 --> 01:08:31,200
Speaker 3: in like a high income. It's just like he's he's

1443
01:08:31,360 --> 01:08:32,400
Speaker 3: not being able to think about it.

1444
01:08:32,439 --> 01:08:34,880
Speaker 2: I mean, it's uh, it's a it's a little alarming.

1445
01:08:34,920 --> 01:08:37,559
Speaker 2: So look, what do you quickly Look, I'm we're thinking

1446
01:08:37,560 --> 01:08:39,640
Speaker 2: about all these things down to the parochial on the

1447
01:08:39,640 --> 01:08:42,639
Speaker 2: warn Act, which is how we actually warn the public

1448
01:08:42,720 --> 01:08:46,719
Speaker 2: the social impacts of large scale displacement and job loss.

1449
01:08:47,280 --> 01:08:52,000
Speaker 2: Having more capacity to see earlier. That's after the fact

1450
01:08:52,000 --> 01:08:53,920
Speaker 2: the warneck comes out, those jobs are already going to

1451
01:08:53,960 --> 01:08:56,439
Speaker 2: be lost. What are the science to show us what

1452
01:08:56,479 --> 01:08:58,559
Speaker 2: the impacts are happening in the job market in real time?

1453
01:08:58,760 --> 01:09:03,160
Speaker 2: Issues of unemployment and sure it's becoming employment insurance so

1454
01:09:03,200 --> 01:09:05,200
Speaker 2: that you can keep people employed for a period of time.

1455
01:09:05,520 --> 01:09:07,639
Speaker 2: The Dutch do it about ninety percent of the wage.

1456
01:09:07,840 --> 01:09:10,600
Speaker 2: The ultimate training is a job, the dignity of a

1457
01:09:10,680 --> 01:09:14,440
Speaker 2: job back to those Voltaire constructs, and then the opportunity

1458
01:09:14,479 --> 01:09:18,240
Speaker 2: then potentially transition by using the federal government to help you.

1459
01:09:18,280 --> 01:09:22,080
Speaker 2: The backstop portable benefits then become fundamental. This notion of

1460
01:09:22,200 --> 01:09:26,599
Speaker 2: UBC university basic capital, this notion of sovereign wealth fund

1461
01:09:26,760 --> 01:09:31,839
Speaker 2: or equity public equity with dividends and somehow we get shares.

1462
01:09:32,120 --> 01:09:36,240
Speaker 2: There's equity shares contributions from these large companies. That's not

1463
01:09:36,320 --> 01:09:39,040
Speaker 2: just taxes, it's actual equity in the company. So there's

1464
01:09:39,040 --> 01:09:41,839
Speaker 2: a notion of an ownership and a larger ownership society.

1465
01:09:41,960 --> 01:09:46,120
Speaker 2: That we don't tax jobs with payroll taxes and then

1466
01:09:46,280 --> 01:09:50,200
Speaker 2: subsidize automation through tax credits. We do the inverse in

1467
01:09:50,240 --> 01:09:52,960
Speaker 2: that context. So all of that. So that's the stuff

1468
01:09:53,080 --> 01:09:55,120
Speaker 2: I'm playing around with a lot of us are thinking

1469
01:09:55,160 --> 01:09:58,800
Speaker 2: about right now. But how from your vantage point, how

1470
01:09:58,920 --> 01:10:02,360
Speaker 2: quickly is is happening. There's a lot of headlines, but

1471
01:10:02,400 --> 01:10:04,400
Speaker 2: there's a lot of debate around these headlines of all

1472
01:10:04,479 --> 01:10:07,080
Speaker 2: these cuts of jobs. But then people say, well, that

1473
01:10:07,160 --> 01:10:09,759
Speaker 2: was a lot of coverd over iroing there's some business

1474
01:10:09,760 --> 01:10:13,080
Speaker 2: model issues there that they're sort of hiding and suggesting

1475
01:10:13,160 --> 01:10:19,920
Speaker 2: it's AI. We even necessarily seeing massive job destruction yet

1476
01:10:20,200 --> 01:10:22,120
Speaker 2: or have we with AI?

1477
01:10:22,320 --> 01:10:24,559
Speaker 1: I think the point people should get here is that

1478
01:10:24,960 --> 01:10:27,760
Speaker 1: obviously it's complicated, and there's jobs are shuffling throughout the

1479
01:10:27,800 --> 01:10:30,639
Speaker 1: economy a little bit right now. But the long term

1480
01:10:30,680 --> 01:10:32,360
Speaker 1: goal of these companies is.

1481
01:10:32,479 --> 01:10:36,360
Speaker 2: Not the original Open AI.

1482
01:10:36,280 --> 01:10:39,720
Speaker 1: Missions mission statement exactly. Opening Eyes mission statement was not

1483
01:10:39,760 --> 01:10:41,760
Speaker 1: to give people helpful tools so that you can do

1484
01:10:41,800 --> 01:10:44,280
Speaker 1: your job slightly better. Their mission statement is we do

1485
01:10:44,320 --> 01:10:47,160
Speaker 1: your job. And actually, you know govin in in La

1486
01:10:47,240 --> 01:10:48,920
Speaker 1: there's an article in the La Times about this that

1487
01:10:49,640 --> 01:10:51,720
Speaker 1: is one of these new popular gig worker jobs in

1488
01:10:51,840 --> 01:10:55,400
Speaker 1: La is everybody straps a GoPro camera to their head

1489
01:10:56,040 --> 01:10:58,880
Speaker 1: and they look down and then they do laundry, They cook,

1490
01:10:58,960 --> 01:11:02,000
Speaker 1: they eat, they do all these things. So basically the

1491
01:11:02,120 --> 01:11:04,959
Speaker 1: number one job soon in the world will be training

1492
01:11:05,080 --> 01:11:07,200
Speaker 1: the replacement for that job. Think of it like being

1493
01:11:07,200 --> 01:11:09,360
Speaker 1: a coffin builder, Like you're designing the coffin and then

1494
01:11:09,360 --> 01:11:11,360
Speaker 1: you put yourself in it. And you know, if you

1495
01:11:11,360 --> 01:11:13,240
Speaker 1: think I'm lying, by the way, just think about Meta

1496
01:11:14,439 --> 01:11:18,439
Speaker 1: Mettal Instagram creators. You know, here we love creators, we

1497
01:11:18,479 --> 01:11:20,840
Speaker 1: love creativity. We want you to be successful. We want

1498
01:11:20,840 --> 01:11:23,240
Speaker 1: to you know, our whole mission statement is making creators

1499
01:11:23,439 --> 01:11:26,720
Speaker 1: super successful. That's before they were an AI company. When

1500
01:11:26,800 --> 01:11:29,080
Speaker 1: they do the second that they were an AI company,

1501
01:11:28,840 --> 01:11:32,160
Speaker 1: they trained on all the videos of their creators and

1502
01:11:32,200 --> 01:11:35,439
Speaker 1: then created generative videos that now basically suck up like

1503
01:11:35,439 --> 01:11:38,240
Speaker 1: a vampire, your essence, your life, force, your creativity, and

1504
01:11:38,240 --> 01:11:40,080
Speaker 1: then hit button and now they have a digital copy

1505
01:11:40,080 --> 01:11:42,439
Speaker 1: of you that you didn't consent to that can generate

1506
01:11:42,479 --> 01:11:44,840
Speaker 1: all these things. If you don't believe me, Just a

1507
01:11:44,880 --> 01:11:47,000
Speaker 1: few weeks ago, there was an article Meta is now

1508
01:11:47,000 --> 01:11:50,719
Speaker 1: forcing their employees to basically track all of their movements,

1509
01:11:50,760 --> 01:11:52,040
Speaker 1: all their clicks, all the things are doing on a

1510
01:11:52,040 --> 01:11:54,439
Speaker 1: computer to train AI agents to do all their jobs.

1511
01:11:54,840 --> 01:11:57,519
Speaker 1: This is not a conspiracy theory. All you have to

1512
01:11:57,520 --> 01:11:59,839
Speaker 1: do to know where we're going is understand the incentives

1513
01:11:59,840 --> 01:12:02,280
Speaker 1: and look at the early warning signs they're telling you

1514
01:12:02,320 --> 01:12:04,720
Speaker 1: who they really are. You know, Open AI says we're

1515
01:12:04,760 --> 01:12:06,960
Speaker 1: all here to make the world a better place. And

1516
01:12:07,000 --> 01:12:09,960
Speaker 1: then they release this AI slop app Sora, which was

1517
01:12:10,280 --> 01:12:12,200
Speaker 1: basically an infinitely they killed it now, but it was

1518
01:12:12,240 --> 01:12:16,120
Speaker 1: an infinitely scrolling, deep fake generated content of like, you know,

1519
01:12:16,240 --> 01:12:19,360
Speaker 1: funny videos of Stephen Hawking, you know, you know, going

1520
01:12:19,360 --> 01:12:21,880
Speaker 1: through a raceway or something like that. It's just deep

1521
01:12:21,920 --> 01:12:25,040
Speaker 1: fake AI slop. Why are they doing that because they

1522
01:12:25,040 --> 01:12:26,960
Speaker 1: want to increase their market dominance, because they want to

1523
01:12:27,000 --> 01:12:29,280
Speaker 1: get users, because they want to get training data, and

1524
01:12:29,320 --> 01:12:31,519
Speaker 1: it gets more people using open ai, so their numbers

1525
01:12:31,520 --> 01:12:31,840
Speaker 1: going up.

1526
01:12:31,840 --> 01:12:33,599
Speaker 3: And if your incentives tell you everything, and if you're

1527
01:12:33,600 --> 01:12:37,400
Speaker 3: a company and you're choosing do I hire a real

1528
01:12:37,479 --> 01:12:42,360
Speaker 3: human paralegal or GPT seven that works for less than

1529
01:12:42,400 --> 01:12:45,479
Speaker 3: minimum wage twenty four to seven doesn't whistle blow, doesn't

1530
01:12:45,479 --> 01:12:48,600
Speaker 3: a plane, doesn't have cultural issues, doesn't have paid time off? Like,

1531
01:12:48,680 --> 01:12:49,519
Speaker 3: which one are you going to do?

1532
01:12:49,800 --> 01:12:50,000
Speaker 1: Well?

1533
01:12:50,040 --> 01:12:52,360
Speaker 3: Your business incentive is just very very clear, and everyone's

1534
01:12:52,360 --> 01:12:53,000
Speaker 3: going to be trapped in this.

1535
01:12:53,080 --> 01:12:54,880
Speaker 2: And if your competitor you may say no, I'm going

1536
01:12:54,920 --> 01:12:57,200
Speaker 2: to keep the human, and then your competitor goes stops

1537
01:12:57,200 --> 01:12:59,240
Speaker 2: the direction, You're out of business. Yeah, I have no.

1538
01:12:59,280 --> 01:13:01,880
Speaker 1: Choice back to the But with those set of interventions

1539
01:13:01,880 --> 01:13:03,599
Speaker 1: that you just mentioned, and you just mentioned a whole

1540
01:13:03,640 --> 01:13:05,759
Speaker 1: slew of things we can be doing. You just mentioned

1541
01:13:05,800 --> 01:13:07,720
Speaker 1: so many things that we could be doing. And I

1542
01:13:07,720 --> 01:13:10,040
Speaker 1: think what it represents is I think people think, but

1543
01:13:10,080 --> 01:13:12,200
Speaker 1: if we don't race to automate every job as fast

1544
01:13:12,200 --> 01:13:14,439
Speaker 1: as possible, we're going to lose to China. But what

1545
01:13:14,920 --> 01:13:17,519
Speaker 1: this is showing and revealing is we're not in a

1546
01:13:17,600 --> 01:13:20,600
Speaker 1: race just the technology. We're actually in a race for

1547
01:13:20,640 --> 01:13:23,240
Speaker 1: a different currency. The currency is not who has the

1548
01:13:23,240 --> 01:13:25,680
Speaker 1: power first, but who is better at governing, steering and

1549
01:13:25,720 --> 01:13:29,400
Speaker 1: integrating that power in a healthy and sustainable and strengthening

1550
01:13:29,439 --> 01:13:32,240
Speaker 1: way into your society. And we saw this with social

1551
01:13:32,280 --> 01:13:36,120
Speaker 1: media because the US beat China to the psychological bazooka

1552
01:13:36,280 --> 01:13:39,439
Speaker 1: behavior modification machine of social media, and then we had

1553
01:13:39,439 --> 01:13:41,559
Speaker 1: no idea how to govern it. So we flipped around.

1554
01:13:41,600 --> 01:13:43,759
Speaker 1: We blew off our own brain with the brain row economy.

1555
01:13:45,080 --> 01:13:47,519
Speaker 1: So and by the way China regulates social media, they

1556
01:13:47,520 --> 01:13:49,040
Speaker 1: do a whole bunch of suffing. When you open up

1557
01:13:49,080 --> 01:13:52,479
Speaker 1: a douyen TikTok in China, it's their version of TikTok,

1558
01:13:52,720 --> 01:13:55,280
Speaker 1: and you scroll, you get videos about who won the

1559
01:13:55,280 --> 01:13:59,360
Speaker 1: Nobel Prize, financial advice, here's the new quantum physics theory.

1560
01:14:00,040 --> 01:14:02,320
Speaker 1: You know, here's patriotism videos. And obviously there's problems with that.

1561
01:14:02,360 --> 01:14:04,000
Speaker 1: We don't want to do it that way, but the

1562
01:14:04,040 --> 01:14:05,880
Speaker 1: point is you don't have to do it in the

1563
01:14:05,920 --> 01:14:08,360
Speaker 1: wild west. Blow off your own brain. And now if

1564
01:14:08,400 --> 01:14:10,360
Speaker 1: we release it in a way that automates all the

1565
01:14:10,400 --> 01:14:13,400
Speaker 1: labor with no transition plan. It's like, great, we pumped

1566
01:14:13,479 --> 01:14:16,080
Speaker 1: up our steroids, but we just burst our lungs. Yeah, right,

1567
01:14:16,120 --> 01:14:19,360
Speaker 1: the societal body, we do that. So what you just

1568
01:14:19,400 --> 01:14:22,040
Speaker 1: outlined was a set of interventions that we can be

1569
01:14:22,080 --> 01:14:24,760
Speaker 1: exploring to help smooth this transition. And we're in a

1570
01:14:24,760 --> 01:14:28,120
Speaker 1: competition with China for who's better at making this transition

1571
01:14:28,520 --> 01:14:30,559
Speaker 1: to an AI integrated world.

1572
01:14:31,280 --> 01:14:34,080
Speaker 2: Where are you on the transition curve? I mean, I mean,

1573
01:14:34,080 --> 01:14:36,960
Speaker 2: we talk about flattering it, but how quick? I mean, honestly,

1574
01:14:37,080 --> 01:14:40,040
Speaker 2: if we're sitting here a year from now having this conversation,

1575
01:14:40,640 --> 01:14:43,320
Speaker 2: are we looking at that you know, ten percent unemployment,

1576
01:14:43,680 --> 01:14:47,360
Speaker 2: not that twenty necessarily, which is that threshold for fascism

1577
01:14:47,479 --> 01:14:49,320
Speaker 2: and a whole other societal collap.

1578
01:14:49,040 --> 01:14:51,720
Speaker 3: But it's going to be confusing and spiking hard to

1579
01:14:51,720 --> 01:14:55,960
Speaker 3: predict exactly because just think about your own experience with

1580
01:14:56,040 --> 01:14:57,960
Speaker 3: AI so far two years ago, it can barely write

1581
01:14:57,960 --> 01:15:00,040
Speaker 3: an essay, and now it can do some part so

1582
01:15:00,120 --> 01:15:02,000
Speaker 3: of your work pretty well and other parts like that

1583
01:15:02,080 --> 01:15:05,360
Speaker 3: is a really dumb error. And so we're going to

1584
01:15:05,439 --> 01:15:08,040
Speaker 3: see not that much job, not that much job LOFs

1585
01:15:08,400 --> 01:15:11,280
Speaker 3: like entry level stuff getting sort of switched around and

1586
01:15:11,320 --> 01:15:13,400
Speaker 3: then it's going to hit really really quickly when it

1587
01:15:13,439 --> 01:15:15,400
Speaker 3: crosses the next spreshold, just like Mythos did.

1588
01:15:15,439 --> 01:15:18,240
Speaker 1: There's another confusing aspect about AI that people in our

1589
01:15:18,280 --> 01:15:21,880
Speaker 1: space called AI jaggedness, which is that there's certain capabilities

1590
01:15:21,920 --> 01:15:25,480
Speaker 1: like in cyber hacking, that it is already superhuman, already superhumans,

1591
01:15:25,960 --> 01:15:28,840
Speaker 1: while it'll still make a very basic dumb mistake on

1592
01:15:28,880 --> 01:15:31,600
Speaker 1: something else. And I think part of what's confusing for

1593
01:15:31,640 --> 01:15:33,840
Speaker 1: people that naturally has them say is this just hype

1594
01:15:33,840 --> 01:15:35,720
Speaker 1: and these companies are trying to hype this stuff is

1595
01:15:35,760 --> 01:15:37,760
Speaker 1: if you just look at the dumb examples where it's

1596
01:15:37,800 --> 01:15:40,720
Speaker 1: messing up, you're like, this thing isn't that powerful. We

1597
01:15:40,840 --> 01:15:44,880
Speaker 1: have never been confronted with the technology that is simultaneously

1598
01:15:45,240 --> 01:15:49,320
Speaker 1: sci Fi level superhuman that makes a person who's very

1599
01:15:49,360 --> 01:15:52,479
Speaker 1: deep in security call it like seeing Jesus. At the

1600
01:15:52,520 --> 01:15:55,280
Speaker 1: same time that that same technology can like mistake how

1601
01:15:55,280 --> 01:15:58,000
Speaker 1: many rs are in the word strawberry, like, we have

1602
01:15:58,080 --> 01:15:59,599
Speaker 1: just not seen that. And so it's just I want

1603
01:15:59,600 --> 01:16:01,840
Speaker 1: people to not for their own psychology that this is

1604
01:16:01,880 --> 01:16:05,920
Speaker 1: a new psychological object that our normal intuitions about how

1605
01:16:05,960 --> 01:16:08,559
Speaker 1: to evaluate something we have to get more nuanced.

1606
01:16:09,240 --> 01:16:11,479
Speaker 2: And part of the nuance is a deeper understanding that

1607
01:16:11,600 --> 01:16:13,679
Speaker 2: we're not just talking about apps here, we're talking about

1608
01:16:13,680 --> 01:16:14,880
Speaker 2: the physical world as well.

1609
01:16:15,160 --> 01:16:15,519
Speaker 1: That's right.

1610
01:16:15,960 --> 01:16:19,000
Speaker 2: Was a big headline that I hope people paid attention

1611
01:16:19,080 --> 01:16:24,080
Speaker 2: to when Elon announced that in his original factory here

1612
01:16:24,080 --> 01:16:26,920
Speaker 2: in Fremont, California, that he's converting the s and the

1613
01:16:26,920 --> 01:16:31,280
Speaker 2: ex Tesla cars now to humanoid robotics and his goal

1614
01:16:31,360 --> 01:16:34,040
Speaker 2: is ultimately a million You know, that's Elon who they

1615
01:16:34,080 --> 01:16:36,719
Speaker 2: all knows, you know, his goal setting. But the notion

1616
01:16:37,080 --> 01:16:40,719
Speaker 2: that he's converting that factory from cars to humanoid robotics

1617
01:16:41,000 --> 01:16:44,320
Speaker 2: and AI now into the physical world is that. I mean,

1618
01:16:45,280 --> 01:16:48,080
Speaker 2: we're seeing driverless cars and if you haven't seen them

1619
01:16:48,080 --> 01:16:50,640
Speaker 2: in you know, your home state, you're going about to

1620
01:16:50,760 --> 01:16:54,000
Speaker 2: you're going to see flying cars or they're just quad

1621
01:16:54,000 --> 01:16:56,680
Speaker 2: copters basically, but they're coming soon. We're going to be

1622
01:16:56,760 --> 01:16:58,240
Speaker 2: doing a lot more of that run by the way.

1623
01:16:58,560 --> 01:17:01,360
Speaker 1: That's great. We can have a world of quad copters

1624
01:17:01,439 --> 01:17:05,760
Speaker 1: and you know, innovation while not racing to replace us economically,

1625
01:17:05,840 --> 01:17:08,439
Speaker 1: replace us socially. So Mark Zuckerberg, you know, designs your

1626
01:17:08,479 --> 01:17:10,599
Speaker 1: kids friends rather than you had them having actual friends

1627
01:17:10,840 --> 01:17:13,040
Speaker 1: replace us politically and not having political power and then

1628
01:17:13,040 --> 01:17:16,400
Speaker 1: replace us physically by owning our physical presence in our

1629
01:17:16,439 --> 01:17:19,800
Speaker 1: in robots. So again, I think people might hear this

1630
01:17:19,840 --> 01:17:22,519
Speaker 1: as an anti technology conversation. Yeah, you know, you talked

1631
01:17:22,520 --> 01:17:25,840
Speaker 1: about your legacy and your father and grandfather and we

1632
01:17:25,840 --> 01:17:30,160
Speaker 1: were talking backstage. Ye A's's father started the Macintosh projected

1633
01:17:30,200 --> 01:17:32,559
Speaker 1: at Yeah. Yeah, you know, we come from a legacy

1634
01:17:32,640 --> 01:17:35,519
Speaker 1: where the word humane is about an inspiring vision about

1635
01:17:35,520 --> 01:17:38,879
Speaker 1: technology that's actually integrated and in service of our humanity,

1636
01:17:39,280 --> 01:17:41,360
Speaker 1: of a pro human future. That's what all of this

1637
01:17:41,400 --> 01:17:44,240
Speaker 1: is motivated for. So I get excited about you know, flying,

1638
01:17:44,439 --> 01:17:46,160
Speaker 1: you know, cars that have cool you know.

1639
01:17:46,120 --> 01:17:48,200
Speaker 3: Aif And just to say too, just as my own

1640
01:17:48,200 --> 01:17:51,000
Speaker 3: personal experiences, you know, I spend a big portion of

1641
01:17:51,040 --> 01:17:53,519
Speaker 3: my life. I found it a thing called Earth Species Project.

1642
01:17:54,000 --> 01:17:56,000
Speaker 3: We're now around forty people and you're.

1643
01:17:55,840 --> 01:17:56,799
Speaker 2: The biggest consumer.

1644
01:17:57,600 --> 01:18:00,679
Speaker 3: I wouldn't say biggest consumer. Yeah exactly, it's it's men. Sorry, guys,

1645
01:18:02,000 --> 01:18:05,320
Speaker 3: first infinite scroll now this, but you know, we were

1646
01:18:05,400 --> 01:18:08,600
Speaker 3: using AI to translate animal language animal communication. One of

1647
01:18:08,600 --> 01:18:12,360
Speaker 3: our researchers just discovered just before she joined us that like,

1648
01:18:12,439 --> 01:18:14,840
Speaker 3: not only do dolphins have names that they call each

1649
01:18:14,840 --> 01:18:17,439
Speaker 3: other by that their mothers teach them. They will talk

1650
01:18:17,439 --> 01:18:19,320
Speaker 3: about each other in the third person, so they'll talk

1651
01:18:19,360 --> 01:18:21,720
Speaker 3: about another dolphin that isn't here. So it's like one

1652
01:18:21,760 --> 01:18:23,280
Speaker 3: of the biggest hallmarks of language to be able to

1653
01:18:23,280 --> 01:18:25,720
Speaker 3: talk about something that's not here and not now. But

1654
01:18:25,760 --> 01:18:29,400
Speaker 3: they will continue to use their mother's name even after

1655
01:18:29,439 --> 01:18:34,240
Speaker 3: she's died. Sweet right, it's beautiful. These are the kinds

1656
01:18:34,240 --> 01:18:37,040
Speaker 3: of things that AI can show us and teach us

1657
01:18:37,040 --> 01:18:39,479
Speaker 3: about the world and connect us with the natural world,

1658
01:18:39,560 --> 01:18:42,240
Speaker 3: the world around us, and show things we couldn't possibly imagine.

1659
01:18:42,320 --> 01:18:44,200
Speaker 3: So I just want everyone to hear it's not that

1660
01:18:44,240 --> 01:18:47,519
Speaker 3: it's just here saying no AI. It's just saying not AI.

1661
01:18:47,840 --> 01:18:51,640
Speaker 3: In this way that technological progress might be inevitable, but

1662
01:18:51,720 --> 01:18:55,280
Speaker 3: the way that AI rolls out is not. And every

1663
01:18:55,320 --> 01:18:58,200
Speaker 3: time somebody says it's inevitable, it's like casting a spell.

1664
01:18:58,280 --> 01:19:01,120
Speaker 3: That's where because if if it's inevitable, then there's nothing

1665
01:19:01,160 --> 01:19:02,840
Speaker 3: we can do, then you shouldn't possibly act.

1666
01:19:02,880 --> 01:19:05,320
Speaker 1: There's a button we're hitting called commit suicide. Commits suicide?

1667
01:19:05,520 --> 01:19:07,200
Speaker 1: Is it inevitable that we all hit that button?

1668
01:19:07,240 --> 01:19:07,360
Speaker 2: No?

1669
01:19:07,400 --> 01:19:09,560
Speaker 1: When it's called you know, ten percent GDP growth And

1670
01:19:09,680 --> 01:19:12,519
Speaker 1: like an innovation, we pushed the button. If we could

1671
01:19:12,520 --> 01:19:15,280
Speaker 1: collectively see that the button we're pushing is more nuanced

1672
01:19:15,280 --> 01:19:17,720
Speaker 1: than that, that there's some threshold by which we are

1673
01:19:17,960 --> 01:19:20,880
Speaker 1: essentially committing civilizational suicide. We're not going to have a

1674
01:19:20,920 --> 01:19:24,879
Speaker 1: human future. If you ask anybody what gives me hope?

1675
01:19:25,000 --> 01:19:27,240
Speaker 1: We've been on the road with this film. You walk

1676
01:19:27,280 --> 01:19:29,519
Speaker 1: people through the basic facts we've talked through, and you

1677
01:19:29,560 --> 01:19:31,599
Speaker 1: ask who here is stoked about the future that we're

1678
01:19:31,640 --> 01:19:34,479
Speaker 1: headed to. I was even at the Miami Tech Summit

1679
01:19:34,520 --> 01:19:36,479
Speaker 1: in front of a pro business pro AI. A lot

1680
01:19:36,520 --> 01:19:39,479
Speaker 1: of people invested in it. Again, we're invested in in

1681
01:19:39,520 --> 01:19:42,040
Speaker 1: positive technology too, but I was able to see that

1682
01:19:42,200 --> 01:19:44,040
Speaker 1: entire audience is like, no, I don't want that either.

1683
01:19:44,200 --> 01:19:45,000
Speaker 2: Wow. Yeah.

1684
01:19:45,040 --> 01:19:47,080
Speaker 1: So again I think it's like, as long as you

1685
01:19:47,120 --> 01:19:49,000
Speaker 1: give people the off ramp, there are ways we can

1686
01:19:49,080 --> 01:19:52,360
Speaker 1: have technology that's in service of making life better. We

1687
01:19:52,400 --> 01:19:55,040
Speaker 1: need to be funding and innovating in that way, and

1688
01:19:55,080 --> 01:19:56,680
Speaker 1: we need to be blocking off the parts that are

1689
01:19:56,920 --> 01:20:00,479
Speaker 1: hurting our children, that are, you know, diminishing people cognitive

1690
01:20:00,479 --> 01:20:04,160
Speaker 1: capacities or replacing their kids' relationships with AI companions. And

1691
01:20:04,200 --> 01:20:05,479
Speaker 1: we can do that. And you've done some of that

1692
01:20:05,479 --> 01:20:08,439
Speaker 1: with here in California. So you know, it's not that

1693
01:20:08,520 --> 01:20:12,400
Speaker 1: I am or weird by default optimistic about the default trajectory,

1694
01:20:12,439 --> 01:20:15,120
Speaker 1: not at all. It's that if we have the clarity,

1695
01:20:15,680 --> 01:20:17,800
Speaker 1: we can put our hand in the steering wheel and

1696
01:20:17,840 --> 01:20:18,880
Speaker 1: we can steer it somewhere else.

1697
01:20:19,400 --> 01:20:21,960
Speaker 2: This notion of accelerate and steer.

1698
01:20:21,880 --> 01:20:24,439
Speaker 1: Yeah, yes, right, exactly what happens when you accelerate and

1699
01:20:24,479 --> 01:20:27,519
Speaker 1: you don't steer, you obviously crash. It's just it's like

1700
01:20:27,560 --> 01:20:30,679
Speaker 1: not rocket scient there, it's one hundred percent the likely outcome.

1701
01:20:32,680 --> 01:20:36,160
Speaker 2: What just as we wrap up this notion of governance,

1702
01:20:36,240 --> 01:20:38,679
Speaker 2: going back to that is it's it's the foundation here

1703
01:20:39,240 --> 01:20:43,320
Speaker 2: which the regulations and and and the relationships are formed,

1704
01:20:43,320 --> 01:20:46,680
Speaker 2: the partnerships to begin to address the common humanity and

1705
01:20:46,680 --> 01:20:49,600
Speaker 2: the common threat and the common cause, common opportunities that

1706
01:20:49,680 --> 01:20:53,559
Speaker 2: present themselves. This notion of catcher, I mean, you've got

1707
01:20:53,640 --> 01:20:57,720
Speaker 2: these packs, You've got so much concentrated wealth. Yeah, we're

1708
01:20:57,720 --> 01:21:02,160
Speaker 2: going to likely have the first few trillionaires this year. Yeah,

1709
01:21:02,200 --> 01:21:06,519
Speaker 2: this calendar years right, yeah, yep, these IPOs, I mean

1710
01:21:06,560 --> 01:21:09,760
Speaker 2: it's just I'm a now, it's a big surplus in

1711
01:21:09,800 --> 01:21:13,480
Speaker 2: the state, the abundance, the GDP I mean, it's frothy,

1712
01:21:14,040 --> 01:21:15,840
Speaker 2: as they say, but a lot of that now is

1713
01:21:15,880 --> 01:21:18,800
Speaker 2: going to you know, make sure that you know we're

1714
01:21:18,800 --> 01:21:24,120
Speaker 2: protecting incumbents against innovation. You know, some incumbent capitalism, not

1715
01:21:24,320 --> 01:21:28,960
Speaker 2: entrepreneurial capitalism, necessarily innovation capitalism. There's that friction that's always

1716
01:21:28,960 --> 01:21:31,280
Speaker 2: ongoing and those that are just going to do everything

1717
01:21:31,320 --> 01:21:33,680
Speaker 2: to hammer as they did with social media, to make

1718
01:21:33,720 --> 01:21:37,240
Speaker 2: sure there's no regular we're still debating section two thirty

1719
01:21:36,920 --> 01:21:39,880
Speaker 2: that Christ yeah in this country. So how do we

1720
01:21:39,920 --> 01:21:42,720
Speaker 2: start to break that reality?

1721
01:21:42,840 --> 01:21:45,760
Speaker 4: I think money in politics, that's just that's right, that's

1722
01:21:46,320 --> 01:21:48,320
Speaker 4: at the anti just just a little bit even further,

1723
01:21:48,360 --> 01:21:51,200
Speaker 4: which is that as we get more trillionaires, like they

1724
01:21:51,200 --> 01:21:54,800
Speaker 4: can hire private security, but that still requires relying on

1725
01:21:54,880 --> 01:21:55,479
Speaker 4: human beings.

1726
01:21:55,479 --> 01:21:58,080
Speaker 3: But you're just pointing out that we're heading into like

1727
01:21:58,240 --> 01:22:01,120
Speaker 3: a world we have drone army, we have human eight

1728
01:22:01,200 --> 01:22:05,920
Speaker 3: rowood armies. When trillionaires can just buy like their drone

1729
01:22:06,040 --> 01:22:09,160
Speaker 3: armies to fight, like, we enter into techno feudalism, right,

1730
01:22:09,200 --> 01:22:11,599
Speaker 3: And so that is the world that if we see,

1731
01:22:12,520 --> 01:22:14,519
Speaker 3: if we don't do something, we're going to end up into.

1732
01:22:14,680 --> 01:22:17,519
Speaker 1: But to answer your question, the it's all about the

1733
01:22:17,560 --> 01:22:20,639
Speaker 1: campaign Loving, it's all about that, And one hundred million,

1734
01:22:20,720 --> 01:22:24,320
Speaker 1: one hundred and ninety million dollars has gone into basically

1735
01:22:24,400 --> 01:22:29,240
Speaker 1: AI accelerationist packs and funding is for this midterm, for

1736
01:22:29,280 --> 01:22:32,240
Speaker 1: this mintrum electional Love interm just the midterms. That's so

1737
01:22:32,240 --> 01:22:34,719
Speaker 1: one hundred ninety million. I believe the presidential is two billions.

1738
01:22:34,760 --> 01:22:37,120
Speaker 1: So basically ten percent of the presidential is going into

1739
01:22:37,320 --> 01:22:39,200
Speaker 1: just the midtrums for AI alone, not even for the

1740
01:22:39,240 --> 01:22:39,600
Speaker 1: rest of it.

1741
01:22:39,880 --> 01:22:41,160
Speaker 2: And it's not to regulate.

1742
01:22:41,360 --> 01:22:43,920
Speaker 1: It's not to regulate it. It's it's to say, remove

1743
01:22:43,960 --> 01:22:45,160
Speaker 1: everything and go as fast.

1744
01:22:45,000 --> 01:22:46,479
Speaker 2: As possible, right, let it rip.

1745
01:22:46,760 --> 01:22:49,880
Speaker 1: If you had everyone in the world, I think here

1746
01:22:49,960 --> 01:22:52,679
Speaker 1: the conversation we just had and at a basic level

1747
01:22:52,840 --> 01:22:54,680
Speaker 1: common sense looking at your children in the eye and

1748
01:22:54,720 --> 01:22:57,920
Speaker 1: say are you stoked about that future? No one wants that.

1749
01:22:58,680 --> 01:23:01,519
Speaker 1: So the key piece of agency is going into the

1750
01:23:01,520 --> 01:23:05,719
Speaker 1: midterm elections not voting for people who have taken money

1751
01:23:05,720 --> 01:23:09,759
Speaker 1: from those AI accelerationist groups or don't have a position

1752
01:23:09,800 --> 01:23:12,559
Speaker 1: on AI. That's trying to steer away from these outcomes. Now,

1753
01:23:12,560 --> 01:23:14,720
Speaker 1: we obviously have to articulate that in a clearer way.

1754
01:23:14,800 --> 01:23:16,639
Speaker 1: What does it mean to have kind of a pro

1755
01:23:16,760 --> 01:23:20,200
Speaker 1: human platform in future, and the companies try to make

1756
01:23:20,200 --> 01:23:22,680
Speaker 1: the conversation inaccessible, like oh, well, you don't understand a

1757
01:23:22,840 --> 01:23:24,080
Speaker 1: They're trying to make it seem like you don't know

1758
01:23:24,160 --> 01:23:26,120
Speaker 1: how to regulate live to put us in charge.

1759
01:23:26,400 --> 01:23:28,920
Speaker 3: What they call this the under the hood bias, where

1760
01:23:28,920 --> 01:23:30,400
Speaker 3: it's as if people who know how to make the

1761
01:23:30,439 --> 01:23:33,160
Speaker 3: biggest engines know how to lay out cities and traffic lights,

1762
01:23:33,360 --> 01:23:35,559
Speaker 3: and it's just not true for people who know how

1763
01:23:35,600 --> 01:23:38,040
Speaker 3: to make car engines are not the best people for

1764
01:23:38,120 --> 01:23:39,840
Speaker 3: knowing how to make cars safe and.

1765
01:23:39,840 --> 01:23:43,880
Speaker 1: Prevent recidents work exactly. So it's pretty simple. It's like,

1766
01:23:44,080 --> 01:23:46,880
Speaker 1: do you want an anti human future in which you

1767
01:23:46,880 --> 01:23:49,519
Speaker 1: will be permanently disempowered where no one has an incentive

1768
01:23:49,520 --> 01:23:52,080
Speaker 1: except for charity to pay your bills for you? And

1769
01:23:52,520 --> 01:23:55,519
Speaker 1: you you want the companies that took your job, you

1770
01:23:55,560 --> 01:23:56,960
Speaker 1: want to be dependent on them for the rest of

1771
01:23:57,000 --> 01:23:59,240
Speaker 1: your life to pay your bills for you with no

1772
01:23:59,320 --> 01:24:03,320
Speaker 1: economic leverle no. So this is the final window. You

1773
01:24:03,360 --> 01:24:04,920
Speaker 1: know you want to vote for people who are going

1774
01:24:04,960 --> 01:24:08,160
Speaker 1: to protect you economically, protect you, socially, protect you, politically

1775
01:24:08,160 --> 01:24:11,880
Speaker 1: meaning protect our jobs, protect our vote, voting pro human

1776
01:24:12,320 --> 01:24:14,839
Speaker 1: and obviously that has to get articulated even more clearly.

1777
01:24:15,240 --> 01:24:17,280
Speaker 1: But that is the number one way that people can

1778
01:24:17,320 --> 01:24:19,679
Speaker 1: make a difference in the short term. There's other things too,

1779
01:24:19,720 --> 01:24:23,000
Speaker 1: like boycotting companies that are enabling mass surveillance. You know,

1780
01:24:23,040 --> 01:24:25,800
Speaker 1: when the company's subscriptions you know, go down by a lot,

1781
01:24:25,840 --> 01:24:27,400
Speaker 1: they really need their numbers to be going up. So

1782
01:24:27,439 --> 01:24:29,720
Speaker 1: the companies are more vulnerable than you think, and you're

1783
01:24:29,760 --> 01:24:31,960
Speaker 1: more powerful than you think. Not just if you unsubscribe

1784
01:24:31,960 --> 01:24:34,280
Speaker 1: and boycott them, but get your company, get your church

1785
01:24:34,280 --> 01:24:37,240
Speaker 1: group to do that too, and when those numbers start

1786
01:24:37,240 --> 01:24:38,920
Speaker 1: to change, it actually has a difference.

1787
01:24:39,040 --> 01:24:41,479
Speaker 2: Scott Galloway has been talking a lot about that as well.

1788
01:24:42,479 --> 01:24:46,600
Speaker 2: Absolutely got to use whatever power your disposal. Let's just

1789
01:24:46,680 --> 01:24:51,720
Speaker 2: briefly talk then about the power I mean, free and

1790
01:24:51,760 --> 01:24:56,080
Speaker 2: fair elections. We talk about truth trust more broadly, deep fakes,

1791
01:24:56,560 --> 01:25:00,479
Speaker 2: political ads. I mean I've seen stuff, you know, meetings,

1792
01:25:00,520 --> 01:25:03,840
Speaker 2: conversations I've had that are I mean next level what's

1793
01:25:03,880 --> 01:25:06,920
Speaker 2: out the you yes, just the bs, it's already out there.

1794
01:25:07,080 --> 01:25:10,479
Speaker 2: The ability to manipulate the crowd and the context of

1795
01:25:10,680 --> 01:25:13,280
Speaker 2: you know, social media, the algorithms I mean, now you've

1796
01:25:13,280 --> 01:25:15,240
Speaker 2: got you know, concentrated in the hands of a few

1797
01:25:15,560 --> 01:25:17,519
Speaker 2: in that respect, I mean that whole thing. I mean

1798
01:25:17,760 --> 01:25:20,680
Speaker 2: you you guys have talked about free and fair elections.

1799
01:25:21,479 --> 01:25:24,800
Speaker 2: We talk about the timelines not only in job displacement,

1800
01:25:25,040 --> 01:25:30,439
Speaker 2: but timelines to get this right domestically globally. I mean

1801
01:25:32,280 --> 01:25:35,800
Speaker 2: twenty twenty six, you're talking mid terms. I mean this

1802
01:25:35,880 --> 01:25:37,519
Speaker 2: is I mean, we only have a few more at

1803
01:25:37,520 --> 01:25:40,160
Speaker 2: bats to get this right. That's right? Or is that overstated? No,

1804
01:25:41,000 --> 01:25:41,639
Speaker 2: it's a few more.

1805
01:25:42,280 --> 01:25:44,639
Speaker 3: Certainly by twenty twenty eight, Like that'll be the last

1806
01:25:45,120 --> 01:25:47,639
Speaker 3: human election. Like it's going to be AIS running all

1807
01:25:47,640 --> 01:25:50,200
Speaker 3: of the election, election campaigns, the ads, doing all the

1808
01:25:50,320 --> 01:25:53,240
Speaker 3: both the information and disinformation, because human beings just can't

1809
01:25:53,240 --> 01:25:55,240
Speaker 3: operate at that speed and are not that effective.

1810
01:25:55,400 --> 01:25:56,360
Speaker 2: Yeah, so this is it.

1811
01:25:56,439 --> 01:25:58,559
Speaker 1: This is the window. But and I know, I just

1812
01:25:58,560 --> 01:26:00,439
Speaker 1: want to like get the human experience level. We've been

1813
01:26:00,479 --> 01:26:03,680
Speaker 1: talking about some hard stuff, the last little level, and

1814
01:26:03,840 --> 01:26:05,040
Speaker 1: you know, I just want to say, you know, we

1815
01:26:05,080 --> 01:26:08,160
Speaker 1: struggle with how to communicate this in a way that's responsible,

1816
01:26:08,200 --> 01:26:11,559
Speaker 1: because here's the trade. Right, it's hard to face this.

1817
01:26:11,680 --> 01:26:13,799
Speaker 1: But if we don't face it, we just like, look away,

1818
01:26:14,760 --> 01:26:16,040
Speaker 1: what are we going to get? We're going to get

1819
01:26:16,080 --> 01:26:19,240
Speaker 1: the default anti human path, and so there's this trade

1820
01:26:19,280 --> 01:26:21,960
Speaker 1: where we the only way out is through like it is.

1821
01:26:22,000 --> 01:26:23,559
Speaker 1: We call it kind of like a rite of passage,

1822
01:26:23,560 --> 01:26:28,000
Speaker 1: like our ability to confront basically a shadow of a

1823
01:26:28,040 --> 01:26:31,960
Speaker 1: technology and the default future that that brings. If we

1824
01:26:32,000 --> 01:26:34,280
Speaker 1: can see that clearly, and if we can know that,

1825
01:26:34,400 --> 01:26:36,240
Speaker 1: you know, and I know that we don't want that,

1826
01:26:36,479 --> 01:26:38,519
Speaker 1: and if g and Trump and you know, the people

1827
01:26:38,520 --> 01:26:40,599
Speaker 1: at the highest levels of these governments, Because you ask

1828
01:26:40,640 --> 01:26:43,360
Speaker 1: any reasonable person at a very high level of national

1829
01:26:43,400 --> 01:26:45,679
Speaker 1: security on any side, and you say, do you want

1830
01:26:45,720 --> 01:26:47,960
Speaker 1: AIS that are going road can hack into any computer

1831
01:26:48,040 --> 01:26:51,439
Speaker 1: system and are already mining cryptocurrency? Does that sound good

1832
01:26:51,479 --> 01:26:53,519
Speaker 1: to you? Does that sound safe to you? At a

1833
01:26:53,640 --> 01:26:57,639
Speaker 1: universal level, it's not. So there's actually much more common ground.

1834
01:26:57,960 --> 01:27:00,000
Speaker 1: And even you know forty se I think it's already

1835
01:27:00,080 --> 01:27:02,000
Speaker 1: the case that fifty seven percent of Americans think that

1836
01:27:02,040 --> 01:27:04,760
Speaker 1: the risks of AI currently outweigh the benefits. I don't

1837
01:27:04,760 --> 01:27:07,120
Speaker 1: like that stat because it makes it too like it's

1838
01:27:07,160 --> 01:27:09,599
Speaker 1: all bad versus all good or something like that. There's

1839
01:27:09,640 --> 01:27:12,719
Speaker 1: already the pro human AI Declaration, where forty six groups

1840
01:27:12,760 --> 01:27:15,559
Speaker 1: came together and said, we agree on these five principles

1841
01:27:15,600 --> 01:27:17,160
Speaker 1: to make a pro human future. You can look it up.

1842
01:27:17,160 --> 01:27:19,920
Speaker 1: It's human Statement dot org. That's the one that also

1843
01:27:19,920 --> 01:27:23,759
Speaker 1: includes again Glen Back, Bernie Sanders, Steve Bannon, all these people.

1844
01:27:24,160 --> 01:27:26,760
Speaker 1: I believe it's sixty five percent of Americans believe we

1845
01:27:26,760 --> 01:27:29,559
Speaker 1: should not create superintelligence until we know how to do

1846
01:27:29,600 --> 01:27:33,040
Speaker 1: it provably, safely and controlled. Sounds like a pretty basic thing,

1847
01:27:33,080 --> 01:27:35,160
Speaker 1: like let's not do something. It's like, should we build

1848
01:27:35,160 --> 01:27:36,880
Speaker 1: a nuclear bomb that until we know how to do

1849
01:27:36,920 --> 01:27:38,840
Speaker 1: it safely or nobody won't set off of if we

1850
01:27:38,840 --> 01:27:41,000
Speaker 1: won't ignite the atmosphere, Probably we should wait to do that.

1851
01:27:41,560 --> 01:27:43,799
Speaker 1: So this is not a radical proposal.

1852
01:27:43,840 --> 01:27:46,000
Speaker 3: This is not do you want to know how many,

1853
01:27:46,120 --> 01:27:48,680
Speaker 3: like what percentage of Americans think that we should just

1854
01:27:48,720 --> 01:27:53,320
Speaker 3: go as fast as possible, unfettered, non regulated AI. What

1855
01:27:53,439 --> 01:27:55,360
Speaker 3: percent of Americans? Five percent?

1856
01:27:55,720 --> 01:27:57,000
Speaker 2: Literally, yeah, literally five percent.

1857
01:27:57,880 --> 01:28:00,559
Speaker 3: So actually it's the most popular to run on. That's

1858
01:28:00,640 --> 01:28:02,639
Speaker 3: right to do the like the safe thing, And.

1859
01:28:02,600 --> 01:28:04,960
Speaker 1: It's whether you're a Democrat or Republican, you don't want

1860
01:28:04,960 --> 01:28:07,280
Speaker 1: to be surveilled by AIS. Whether you're Democrat Republican, you

1861
01:28:07,280 --> 01:28:09,120
Speaker 1: don't want AIS taking your jobs, which it will do

1862
01:28:09,160 --> 01:28:10,280
Speaker 1: equally to both sides.

1863
01:28:10,360 --> 01:28:13,320
Speaker 2: But do you then subscribe to the burning AOC frame

1864
01:28:13,600 --> 01:28:17,160
Speaker 2: just shut down the data centers and moratorium.

1865
01:28:16,479 --> 01:28:19,799
Speaker 1: And tell I think of it as those data centers,

1866
01:28:19,840 --> 01:28:22,599
Speaker 1: and it's like I want a pro human data center policy.

1867
01:28:22,800 --> 01:28:25,120
Speaker 1: It's like, you get to build the data center when

1868
01:28:25,360 --> 01:28:27,120
Speaker 1: these conditions are met and we know that it's a

1869
01:28:27,720 --> 01:28:29,479
Speaker 1: pro usure. I'm not saying that's easy. I'm not saying

1870
01:28:29,520 --> 01:28:32,360
Speaker 1: no articulation that because I want people to hear. It's

1871
01:28:32,400 --> 01:28:35,719
Speaker 1: not just no to all of it. It's making sure

1872
01:28:35,720 --> 01:28:38,560
Speaker 1: that the conditions are the steering is built in. So

1873
01:28:38,600 --> 01:28:40,519
Speaker 1: when you see that data center, you should ask, is

1874
01:28:40,520 --> 01:28:43,920
Speaker 1: that data center here to basically enhance my life and

1875
01:28:43,920 --> 01:28:45,240
Speaker 1: strengthen my family.

1876
01:28:45,479 --> 01:28:47,439
Speaker 2: You were even just saying data center was solar. Mostly

1877
01:28:47,520 --> 01:28:48,599
Speaker 2: data centers aren't solar.

1878
01:28:49,200 --> 01:28:49,559
Speaker 1: That's right.

1879
01:28:50,080 --> 01:28:53,000
Speaker 2: We're turning back on cold plants. That's right, old natural

1880
01:28:53,000 --> 01:28:54,680
Speaker 2: gas plants that are exactly yea.

1881
01:28:55,680 --> 01:28:58,519
Speaker 3: Often in the sci fi movies, when is it the

1882
01:28:58,560 --> 01:29:01,559
Speaker 3: case that human beings act actually stop all their bickering

1883
01:29:02,200 --> 01:29:05,320
Speaker 3: and they start coordinating, it's when the aliens come, right, Yeah,

1884
01:29:05,640 --> 01:29:07,960
Speaker 3: we are summoning the demon. We are summoning the alien.

1885
01:29:08,280 --> 01:29:10,439
Speaker 3: And if we can see it that way, then it's

1886
01:29:10,439 --> 01:29:12,679
Speaker 3: sort of like a Game of Thrones, winter is coming.

1887
01:29:13,120 --> 01:29:14,640
Speaker 3: We have to understand winter is coming, then all the

1888
01:29:14,680 --> 01:29:16,760
Speaker 3: fighting in West Ros can like pause for just long

1889
01:29:16,880 --> 01:29:19,200
Speaker 3: enough that we can deal with it. That's this moment,

1890
01:29:19,200 --> 01:29:22,160
Speaker 3: because otherwise it just feels completely hopeless, like how are

1891
01:29:22,200 --> 01:29:24,720
Speaker 3: we going to deal with all the finance reforms when

1892
01:29:24,720 --> 01:29:29,639
Speaker 3: we can't when has Congress actually done anything? And yet

1893
01:29:29,800 --> 01:29:33,560
Speaker 3: there is this one moment where like all of humanity

1894
01:29:33,760 --> 01:29:36,439
Speaker 3: is on one side, there is a human movement. That's

1895
01:29:36,439 --> 01:29:39,120
Speaker 3: what I think the social media stuff shows. If we

1896
01:29:39,160 --> 01:29:41,360
Speaker 3: don't think of this as just an AI problem, but

1897
01:29:41,600 --> 01:29:46,160
Speaker 3: as a technology encroaching onto our humanity, overreaching into our

1898
01:29:46,240 --> 01:29:49,639
Speaker 3: humanity problem, then actually there is massive momentum, more than

1899
01:29:49,640 --> 01:29:51,360
Speaker 3: we ever thought was possible, because what we have to

1900
01:29:51,439 --> 01:29:55,360
Speaker 3: do is juice those like the momentum that's already there.

1901
01:29:56,200 --> 01:30:00,000
Speaker 2: Yeah. Well, look, in the absence of a federal leadership,

1902
01:30:00,080 --> 01:30:02,240
Speaker 2: California will continue to assert itself. I believe in the

1903
01:30:02,240 --> 01:30:05,559
Speaker 2: power of emulation successfullys clues will continue to try to

1904
01:30:05,920 --> 01:30:09,040
Speaker 2: iterate on this and lean in. But look this, you

1905
01:30:09,080 --> 01:30:11,320
Speaker 2: know the clarity of you guys bring to this conversation,

1906
01:30:11,439 --> 01:30:14,560
Speaker 2: The importance of this conversation being brought to scale and

1907
01:30:15,320 --> 01:30:19,280
Speaker 2: brought it into consciousness and the imperative of seeing this

1908
01:30:19,439 --> 01:30:22,719
Speaker 2: documentary again. The documentary is called.

1909
01:30:22,600 --> 01:30:25,520
Speaker 1: The AI doc or How I Became an apocaly optimism.

1910
01:30:25,640 --> 01:30:29,960
Speaker 2: And we can't let that slip twice because you've used

1911
01:30:30,080 --> 01:30:33,720
Speaker 2: a word that people are not familiar with, which is

1912
01:30:34,080 --> 01:30:37,040
Speaker 2: a good way to end, and that is this convergence

1913
01:30:37,120 --> 01:30:41,719
Speaker 2: of optimism and pessimism, a little more optimism than pessimism

1914
01:30:41,960 --> 01:30:42,599
Speaker 2: in a better place.

1915
01:30:42,640 --> 01:30:44,600
Speaker 1: And it's about agency. And I'll just leave you with

1916
01:30:44,640 --> 01:30:47,200
Speaker 1: a quote that I loved from a meditation teacher who

1917
01:30:47,240 --> 01:30:50,840
Speaker 1: talked who happened to be a meditation teacher, and it's

1918
01:30:50,880 --> 01:30:54,080
Speaker 1: from the Army Corps of Engineers, which is that the

1919
01:30:54,240 --> 01:30:58,880
Speaker 1: difficult we do today, the impossible takes just a little longer.

1920
01:31:00,160 --> 01:31:04,880
Speaker 3: Like it, and just end by saying it really actually

1921
01:31:05,040 --> 01:31:07,960
Speaker 3: isn't about being an optimist or a pessimist, because to

1922
01:31:08,439 --> 01:31:11,120
Speaker 3: choose that label for yourself, it's a sort of it's

1923
01:31:11,160 --> 01:31:13,000
Speaker 3: to take a backseat, like to sit down and just

1924
01:31:13,040 --> 01:31:15,160
Speaker 3: be like, I'm just going to accept that it'll either

1925
01:31:15,200 --> 01:31:19,799
Speaker 3: come out well or not, versus taking responsibility for trying

1926
01:31:19,840 --> 01:31:22,639
Speaker 3: to see clearly to shift the world to go well

1927
01:31:24,200 --> 01:31:27,800
Speaker 3: and I think that's what everyone listening can can do,

1928
01:31:27,960 --> 01:31:31,880
Speaker 3: is that this can all feel like too big? What

1929
01:31:32,400 --> 01:31:34,519
Speaker 3: can I do? And then you realize even like the

1930
01:31:34,640 --> 01:31:36,840
Speaker 3: like the CEOs of the company sort of feel a

1931
01:31:36,920 --> 01:31:40,960
Speaker 3: similar way. But this is not just about what we

1932
01:31:41,080 --> 01:31:44,280
Speaker 3: must do. This is fundamentally a question of like who

1933
01:31:44,320 --> 01:31:46,760
Speaker 3: we must be. That if we are the kind of

1934
01:31:46,800 --> 01:31:49,400
Speaker 3: people that aren't looking for a path, and the path

1935
01:31:49,479 --> 01:31:52,759
Speaker 3: is certainly not clear, doesn't seem obvious or even possible.

1936
01:31:53,080 --> 01:31:55,240
Speaker 3: But if we're the kind of people that don't look

1937
01:31:55,240 --> 01:31:56,720
Speaker 3: for the path, we definitely won't find it.

1938
01:31:56,760 --> 01:31:57,240
Speaker 1: If it's there.

1939
01:31:57,640 --> 01:31:59,479
Speaker 3: If we are the kinds of people that do look

1940
01:31:59,520 --> 01:32:01,960
Speaker 3: for the path, then if it's there, we have the

1941
01:32:02,080 --> 01:32:03,599
Speaker 3: opportunity to find it.

1942
01:32:03,640 --> 01:32:07,639
Speaker 1: And just maybe one last thing is people listening to this,

1943
01:32:07,640 --> 01:32:10,439
Speaker 1: this is a lot. Your role is not to take

1944
01:32:10,439 --> 01:32:12,240
Speaker 1: on this whole problem. You don't have to do that.

1945
01:32:12,680 --> 01:32:14,280
Speaker 1: There's some people who are soldiers and there's some people

1946
01:32:14,320 --> 01:32:17,400
Speaker 1: who are civilians. But your role is to be part

1947
01:32:17,439 --> 01:32:20,800
Speaker 1: of the collective immune system against this anti human future.

1948
01:32:20,960 --> 01:32:22,960
Speaker 1: One simple way you can do that is to share

1949
01:32:23,080 --> 01:32:26,800
Speaker 1: this conversation yeah, with literally the most powerful people that

1950
01:32:26,800 --> 01:32:29,320
Speaker 1: you know, and ask them to watch it. And to

1951
01:32:29,320 --> 01:32:31,240
Speaker 1: share it with the most powerful people that they know.

1952
01:32:31,640 --> 01:32:33,960
Speaker 1: And if you've done that, you can say, as long

1953
01:32:34,000 --> 01:32:35,600
Speaker 1: as you are spreading the word and being part of

1954
01:32:35,600 --> 01:32:38,640
Speaker 1: that immune system, you can rest at home, kiss your

1955
01:32:38,720 --> 01:32:41,040
Speaker 1: children at night, focus on the things that all of

1956
01:32:41,040 --> 01:32:43,599
Speaker 1: this is about anyway, which is what do we love

1957
01:32:43,800 --> 01:32:45,840
Speaker 1: about the world, what do we love about life? That

1958
01:32:45,880 --> 01:32:48,400
Speaker 1: we want to continue and come from that place because

1959
01:32:48,439 --> 01:32:51,280
Speaker 1: that is the energy that we will that will inspire

1960
01:32:51,360 --> 01:32:53,599
Speaker 1: other people to want to take those other actions too.

1961
01:32:53,720 --> 01:32:55,719
Speaker 3: And I know, Kevin, you ended up watching an earlier

1962
01:32:55,760 --> 01:32:58,280
Speaker 3: presentation that we did the Aidlemma. Yeah, I think somebody

1963
01:32:58,320 --> 01:32:59,760
Speaker 3: said that you watched it like three times and share

1964
01:32:59,800 --> 01:33:02,320
Speaker 3: with all your staff. How did you end up hearing

1965
01:33:02,320 --> 01:33:02,760
Speaker 3: about it?

1966
01:33:03,640 --> 01:33:06,880
Speaker 2: Well, I mean, come on, hearing about it from you guys.

1967
01:33:07,920 --> 01:33:10,799
Speaker 2: You guys. I was able to get the early early

1968
01:33:10,880 --> 01:33:13,800
Speaker 2: preview from the two of you, and I was able

1969
01:33:14,000 --> 01:33:18,680
Speaker 2: to devour it, took notes, and then shared it universally

1970
01:33:18,720 --> 01:33:22,240
Speaker 2: with everybody around me. Look, you know, the spirit of

1971
01:33:22,320 --> 01:33:23,920
Speaker 2: you guys just DoD it. I couldn't agree with you more.

1972
01:33:23,960 --> 01:33:26,479
Speaker 2: This notion of agency is so important, and we talk

1973
01:33:26,520 --> 01:33:28,479
Speaker 2: about that on the podcast all the time. But this

1974
01:33:28,560 --> 01:33:30,760
Speaker 2: notion of the future. It's not something to experience, something

1975
01:33:30,760 --> 01:33:34,920
Speaker 2: to manifest futures inside of us. And so it's decisions,

1976
01:33:34,960 --> 01:33:36,000
Speaker 2: not conditions.

1977
01:33:35,600 --> 01:33:36,120
Speaker 1: That's exactly.

1978
01:33:36,200 --> 01:33:39,559
Speaker 2: And so this idea that we are powerless, it's just bullshit.

1979
01:33:40,040 --> 01:33:43,960
Speaker 2: It's complete bullshit. Everything that we laid out is an

1980
01:33:43,960 --> 01:33:46,920
Speaker 2: opportunity to do better and be better. And I think

1981
01:33:46,960 --> 01:33:49,479
Speaker 2: the spirit of the calt arms for everybody, as we

1982
01:33:49,520 --> 01:33:51,679
Speaker 2: all have a role to play, and those roles are different,

1983
01:33:51,720 --> 01:33:53,439
Speaker 2: and no one has to be you know, you don't

1984
01:33:53,479 --> 01:33:56,760
Speaker 2: have to be overwhelmed. But this notion of just being

1985
01:33:56,840 --> 01:34:01,160
Speaker 2: present in the conversation and and and in the conversation

1986
01:34:01,240 --> 01:34:05,000
Speaker 2: and sharing it, I think is foundational. So, guys, this

1987
01:34:05,040 --> 01:34:11,600
Speaker 2: is really important. The timeliness of this conversation it cannot overstate.

1988
01:34:12,439 --> 01:34:14,639
Speaker 2: And so I'm very grateful for you to be out

1989
01:34:14,960 --> 01:34:19,080
Speaker 2: on the road all across this country sharing this remarkable documentary.

1990
01:34:19,080 --> 01:34:22,480
Speaker 2: I encourage everybody go out and watch it and more importantly,

1991
01:34:22,840 --> 01:34:25,920
Speaker 2: share it and not fall prey to any of the

1992
01:34:26,000 --> 01:34:30,400
Speaker 2: citizensm and negativity. Maintain your sense of optimism again, we

1993
01:34:30,439 --> 01:34:32,040
Speaker 2: can shape the future. Thank you both