WEBVTT

1
00:00:04.172 --> 00:00:08.115
What I found really interesting is to look at how people's quotes evolve over time with their incentives.

2
00:00:08.135 --> 00:00:14.919
So I was looking at all of the things they've said on the record on podcasts, in their blog posts, to see how it's evolved over time.

3
00:00:14.939 --> 00:00:25.626
And Dario, who was the former VP of Research, OpenAI, and has now moved on to Anthropic, who are taking a slightly different approach to developing AI, said back in 2017, while he was still at OpenAI...

4
00:00:26.507 --> 00:00:33.294
that, this is a quote, I think at the extreme end is the Nick Bostrom style of fear that an AGI could destroy humanity.

5
00:00:33.854 --> 00:00:37.538
I can't see any reason in principle why that couldn't happen.

6
00:00:38.059 --> 00:00:45.366
My chance that something goes really quite catastrophically wrong on the scale of human civilization might be somewhere between 10% and 25%.

7
00:00:47.594 --> 00:00:52.337
And also you mentioned Ilya, who was a co-founder of OpenAI and then left.

8
00:00:53.057 --> 00:00:56.059
I guess the first question I'd ask is, why did Ilya leave?

9
00:00:57.120 --> 00:00:58.181
That's a great question.

10
00:00:59.341 --> 00:01:03.204
So he was instrumental in trying to get Sam Altman fired.

11
00:01:03.624 --> 00:01:13.610
And he's another one of the people who over time began to feel like he was being manipulated by Altman towards contributing something that he didn't believe in.

12
00:01:13.750 --> 00:01:14.571
How did you know?

13
00:01:15.091 --> 00:01:16.212
Because I interviewed a lot of people.

14
00:01:16.693 --> 00:01:22.840
Ilya, in particular, had two pillars that he cared about deeply.

15
00:01:23.761 --> 00:01:27.105
One is making sure we get to so-called AGI.

16
00:01:28.195 --> 00:01:30.615
and the other is making sure that we get to it safely.

17
00:01:31.516 --> 00:01:35.656
And he felt that Altman was actively undermining both things.

18
00:01:36.476 --> 00:01:45.878
He felt that Altman was creating a very chaotic environment within the company where he was pitting teams against each other, where he was telling different things to different people.

19
00:01:46.458 --> 00:01:47.338
Have you ever spoken to him?

20
00:01:48.039 --> 00:01:55.860
I have, so I interviewed him in 2019 for a profile that I did of OpenAI for MIT Technology Review.

21
00:01:56.312 --> 00:02:00.514
And back in 2019, he has a quote where he says, the future is going to be good for AIs regardless.

22
00:02:00.834 --> 00:02:02.975
It would be nice if it was also good for humans as well.

23
00:02:03.936 --> 00:02:08.518
It's not that it's going to actively hate humans or want to harm them, but it's just going to be so powerful.

24
00:02:08.558 --> 00:02:11.319
And I think a good analogy would be the way that humans treat animals.

25
00:02:11.759 --> 00:02:12.720
It's not that we hate animals.

26
00:02:12.960 --> 00:02:15.763
I think humans love animals and have a lot of affection for them.

27
00:02:16.143 --> 00:02:21.107
But when the time comes to build a highway between two cities, we are not asking the animals for permission.

28
00:02:21.628 --> 00:02:23.309
We just do it because it's important to us.

29
00:02:23.749 --> 00:02:33.478
And I think by default, that's the kind of relationship that's going to be between us and AI, which are truly autonomous and operating on their own behalf.

30
00:02:33.638 --> 00:02:35.860
And that was in 2019, the year that you interviewed him.

31
00:02:36.585 --> 00:02:44.453
One of the things that I feel like we should take a step back to examine is going back to this idea of what even is artificial intelligence and what do we mean by intelligence?

32
00:02:45.254 --> 00:02:58.307
And a huge part of the views of the different people and the quotes that you're reading derives from a specific belief that they each have in this question of what is intelligence, what constitutes intelligence.

33
00:02:59.395 --> 00:03:09.846
For Ilya, he has, throughout his research career, felt that ultimately our brains are giant statistical models.

34
00:03:10.606 --> 00:03:15.391
This is not something that, you know, we actually know, but this is his own hypothesis.

35
00:03:15.471 --> 00:03:20.497
Also the hypothesis of his mentor, Jeffrey Hinton, who also was on this podcast.

36
00:03:21.077 --> 00:03:33.842
This is why they have such a strong conviction in the idea of building AI systems that are statistical models and that this particular approach is going to lead to intelligent systems as we are intelligent.

37
00:03:34.062 --> 00:03:35.703
It's a hypothesis that they have.

38
00:03:36.163 --> 00:03:38.864
It's not one that has been proven by science.

39
00:03:39.284 --> 00:03:43.124
And some people vehemently disagree with them on this particular thing.

40
00:03:43.564 --> 00:04:01.588
But if you step into their shoes and take on that hypothesis and assume that it's true, that our brains are in fact statistical engines, and that these systems that they're building are also statistical engines that they're making bigger and bigger and bigger until they become the size of the human brain—

41
00:04:02.788 --> 00:04:15.278
That's why they say that making this comparison where the system will become equal to human intelligence and then maybe exceed human intelligence is relevant in their framework.

42
00:04:15.898 --> 00:04:19.341
And Ilia gave a talk at one point at this point.

43
00:04:19.701 --> 00:04:25.044
really prominent AI research conference that happens every year called Neural Information Processing Systems.

44
00:04:25.385 --> 00:04:25.985
It's a mouthful.

45
00:04:26.485 --> 00:04:35.631
But he gave this keynote where he shows this chart of the size of brains and the intelligence of a species.

46
00:04:36.352 --> 00:04:38.133
And it's roughly linear.

47
00:04:38.193 --> 00:04:41.175
The bigger the size of the brain, the more intelligent the species.

48
00:04:41.775 --> 00:04:44.437
And so for him, he thinks he's

49
00:04:45.217 --> 00:04:49.882
building a digital brain because he thinks brains are just statistical engines.

50
00:04:50.382 --> 00:05:01.273
So from that logic, it's like, okay, if we then build a bigger statistical engine than the human brain, then based on this chart, it will be more intelligent.

51
00:05:01.333 --> 00:05:05.017
And then we will be subjected to the same treatment that we've subjected to.

52
00:05:05.577 --> 00:05:05.937
animals.

53
00:05:06.397 --> 00:05:13.339
But it's really important to understand that these are scientific hypotheses of specific individuals within the AI research community.

54
00:05:13.859 --> 00:05:18.161
And there's a lot, a lot of debate about whether this is in fact the case.

55
00:05:18.921 --> 00:05:24.983
And some of the biggest critics say it's very reductive to think of our brains as simply just statistical engines.

56
00:05:25.383 --> 00:05:29.484
Why does it matter to know the mechanism of

57
00:05:31.168 --> 00:05:39.556
Is it not just important to know the outcome, which is that it's going to be able to do, make a video for me, or agents are going to be able to do the work that I do?

58
00:05:39.636 --> 00:05:43.400
Does it really, really matter for us to know the mechanism behind it?

59
00:05:43.920 --> 00:05:44.521
Yes and no.

60
00:05:45.121 --> 00:05:48.244
So it matters because these companies...

61
00:05:49.215 --> 00:05:54.361
They are driving their future actions based on this hypothesis.

62
00:05:55.082 --> 00:06:00.608
So they have decided, we think that this hypothesis is true.

63
00:06:00.709 --> 00:06:06.415
Like we should just continue building larger and larger statistical models in the pursuit of artificial general intelligence.

64
00:06:06.435 --> 00:06:06.615
And.

65
00:06:07.236 --> 00:06:09.998
And that's then having global consequences.

66
00:06:10.058 --> 00:06:14.020
Like in order to continue doing that, they're hoovering up more and more data.

67
00:06:14.421 --> 00:06:16.282
They're building more and more data centers.

68
00:06:16.762 --> 00:06:22.506
They are having, they're, you know, exploiting more and more labor in order to continue on this path.

69
00:06:22.966 --> 00:06:25.868
Here's a question that I think is important to ask is,

70
00:06:27.032 --> 00:06:30.154
why are we trying to build AI systems that are duplicative of humans?

71
00:06:30.734 --> 00:06:37.779
We're kind of having this conversation right now where we've just taken the premise of this industry as a good thing.

72
00:06:37.939 --> 00:06:42.422
Like they said that we should be building AGI, so we say that we should be building AGI.

73
00:06:42.462 --> 00:06:44.784
But I would like to ask, like, why are we doing that?

74
00:06:45.820 --> 00:06:52.982
Why is it that we are building a technology that is ultimately designed to replace and automate people away?

75
00:06:53.843 --> 00:06:56.604
That is not the enterprise of technology.

76
00:06:57.104 --> 00:06:59.805
Like, we should be building technology.

77
00:06:59.985 --> 00:07:07.627
And the purpose of technology throughout history has been to improve human flourishing, not to replace people.

78
00:07:08.168 --> 00:07:08.628
And so...

79
00:07:09.508 --> 00:07:24.119
This is like a critical part of my critique of these companies and these scientists that have just adopted this goal and have relentlessly pursued it and have had enormous capital and enormous resources to pursue it, is, is this the right goal?

80
00:07:24.179 --> 00:07:25.480
Like, why are we doing this?

81
00:07:25.960 --> 00:07:36.788
Why can't we just build AI systems that do things like accelerate drug discovery and improve people's healthcare outcomes, which are systems that have nothing to do

82
00:07:37.308 --> 00:07:40.990
with the statistical engines that they're trying to build to duplicate the human brain.

83
00:07:41.270 --> 00:07:42.130
So why are they doing it?

84
00:07:42.450 --> 00:07:43.731
I mean, you've interviewed all these people.

85
00:07:43.751 --> 00:07:49.233
I think it's 300 people in total, 80 or 90 of them from OpenAI, the maker of ChatGPT.

86
00:07:49.713 --> 00:07:50.714
Why do you think they're doing it?

87
00:07:52.294 --> 00:07:54.535
I think it's because they're driven by an imperial agenda.

88
00:07:54.795 --> 00:07:57.076
And that is why I call these companies Empires of AI.

89
00:07:57.317 --> 00:07:58.837
What do you mean by an imperial agenda?

90
00:07:59.778 --> 00:08:00.498
What does that term mean?

91
00:08:01.001 --> 00:08:14.554
Empire is the only metaphor that I've ever found to fully encapsulate all of the dimensions of what these companies do and the scale that they operate and what motivates them to do what they do.

92
00:08:15.154 --> 00:08:20.499
And there are many parallels that you see between what I call the empires of AI and the empires of old.

93
00:08:21.220 --> 00:08:25.443
They lay claim to resources that are not their own in the pursuit of training these models.

94
00:08:25.483 --> 00:08:29.667
That's the data of individuals, the intellectual property of artists, writers, and creators.

95
00:08:30.647 --> 00:08:35.571
They're land grabbing in order to build these supercomputer facilities for training the next generation models.

96
00:08:35.991 --> 00:08:38.193
Second, they exploit an extraordinary amount of labor.

97
00:08:38.593 --> 00:08:47.560
They contract hundreds of thousands of workers all around the world, including in the U.S., to ultimately make these technologies.

98
00:08:47.740 --> 00:08:49.242
We can talk about that more.

99
00:08:49.982 --> 00:09:00.654
And they also design their tools to be labor automating so that when the technologies are deployed, it also affects labor rights because it erodes away labor rights.

100
00:09:00.714 --> 00:09:03.097
And this is a political choice that they have.

101
00:09:03.657 --> 00:09:06.420
Third, they monopolize knowledge production.

102
00:09:06.440 --> 00:09:07.842
So they project this idea that

103
00:09:07.962 --> 00:09:10.626
They're the only ones that really understand how the technology works.

104
00:09:10.766 --> 00:09:15.151
And so if the public doesn't like it, it's because they don't actually know enough about this technology.

105
00:09:15.512 --> 00:09:16.513
They do this to the public.

106
00:09:16.794 --> 00:09:18.075
They do this to policymakers.

107
00:09:18.496 --> 00:09:25.285
And they've also captured the majority of the scientists that are working on understanding the limitations and capabilities of AI.

108
00:09:25.525 --> 00:09:27.887
You think they're gaslighting the public, in a way?

109
00:09:28.168 --> 00:09:29.269
They are, yeah.

110
00:09:29.289 --> 00:09:38.438
So if most of the climate scientists in the world were bankrolled by fossil fuel companies, do you think we would get an accurate picture of the climate crisis?

111
00:09:38.839 --> 00:09:38.999
No.

112
00:09:39.882 --> 00:09:47.270
And in the same way, they employ and bankroll, the AI industry employs and bankrolls most of the AI researchers in the world.

113
00:09:47.971 --> 00:09:57.281
So they set the agenda on AI research in soft ways, simply by funneling money to their priorities so that only certain types of AI research are produced.

114
00:09:57.922 --> 00:10:04.506
But they also will censor researchers when they do not like what the researcher has found.

115
00:10:04.667 --> 00:10:18.736
And so I talk about the case of Dr. Timnit Gebru in my book, who was the ethical AI team co-lead at Google, when she was literally hired to critique the types of AI systems that Google was building specifically.

116
00:10:18.856 --> 00:10:28.683
She then co-wrote a critical research paper that was showing how large language models specifically were leading to certain types of harmful outcomes.

117
00:10:29.344 --> 00:10:39.171
And in an attempt to try and stop this research from being published, Google ended up firing Gebru and then fired her other co-lead, Margaret Mitchell.

118
00:10:40.511 --> 00:10:49.481
And so they control and quash the research that is inconvenient to the empire's agenda.

119
00:10:49.921 --> 00:10:54.706
Did you have an example where this is happening to journalists as well that are asking questions of...

120
00:10:55.493 --> 00:10:56.253
They're team members.

121
00:10:56.534 --> 00:11:06.098
I think I was watching a video of yours where there was a young man that was saying he had someone show up at his door, knocked on his door and asked for information, emails, text messages.

122
00:11:06.138 --> 00:11:07.999
And this person was from one of the big AI companies.

123
00:11:08.840 --> 00:11:09.720
This was opening.

124
00:11:09.740 --> 00:11:11.641
I started subpoenaing some of its critics.

125
00:11:11.721 --> 00:11:12.502
Yeah.

126
00:11:12.722 --> 00:11:15.023
As, as part of a.

127
00:11:17.242 --> 00:11:26.788
what appears to be a campaign of intimidation, but also what appeared to be a campaign of phishing for more information to figure out, to map out the network of critics further.

128
00:11:27.348 --> 00:11:41.937
But this was a man who runs a small watchdog nonprofit, and they had been doing a lot of work during that time to try and ask questions about OpenAI's attempt to convert from a nonprofit to a for-profit.

129
00:11:42.697 --> 00:11:57.485
Ultimately, OpenAI was successful in that conversion, but during the period where it was sort of existential for OpenAI to complete this conversion, there were a lot of civil society groups and watchdog groups like Midas who were trying to...

130
00:11:58.679 --> 00:12:01.740
prevent the process from happening in the dead of night.

131
00:12:02.000 --> 00:12:04.241
They were trying to get more transparency.

132
00:12:04.542 --> 00:12:08.303
They were trying to have more public debate about this because it's unprecedented.

133
00:12:09.144 --> 00:12:14.446
And it was then that there was a knock on his door and he was served papers.

134
00:12:15.867 --> 00:12:16.627
What did the papers say?

135
00:12:17.387 --> 00:12:24.394
The papers asked him to reproduce every single piece of communication that he had had that might have involved Musk.

136
00:12:24.474 --> 00:12:31.180
So this was like the strange paranoia that opening I had that Musk was somehow funding these people to block the conversion.

137
00:12:31.200 --> 00:12:33.582
None of them were actually funded by Musk.

138
00:12:33.642 --> 00:12:40.228
So in this particular case, the request, he simply was just answered, you know, I don't have any documents because this doesn't exist.

139
00:12:41.149 --> 00:12:45.875
So going back to this point of empires, you were saying that one of the factors of an empire is a land grab.

140
00:12:46.235 --> 00:12:47.336
And then the next one was...

141
00:12:48.157 --> 00:12:49.318
was labor exploitation.

142
00:12:49.338 --> 00:12:50.258
Labor exploitation.

143
00:12:50.598 --> 00:12:54.221
The third one, controlling knowledge, production.

144
00:12:55.041 --> 00:13:14.433
And one of the other ones that's really important to understand about the AI empires in particular is empires always have this narrative that they say to the public, like, we're the good empire, and we need to be an empire in the first place because there are also bad empires in the world.

145
00:13:15.033 --> 00:13:25.518
And if you allow us to take all the resources and use all of the labor, then we promise we will bring you progress and modernity for everyone.

146
00:13:26.258 --> 00:13:30.120
We will bring you to this utopic state akin to an AI heaven.

147
00:13:30.840 --> 00:13:34.722
But if the evil empire does it first, we will descend into a hell.

148
00:13:35.442 --> 00:13:37.003
And the evil empire being in this case.

149
00:13:37.323 --> 00:13:40.024
In this case, most often it's China.

150
00:13:40.544 --> 00:13:44.965
But actually in the early days, OpenAI evoked Google as the evil empire.

151
00:13:45.605 --> 00:14:01.608
So all of their decisions were about, we need to do it first, because otherwise Google, this evil corporation that's driven by profit, us as a benevolent nonprofit, like this is a critical contest of who wins.

152
00:14:01.988 --> 00:14:04.589
Do you think the people building these AI companies

153
00:14:05.573 --> 00:14:08.974
believe that the outcome is going to be all good now?

154
00:14:09.474 --> 00:14:12.395
Do you think they think that it's going to be, it's going to serve everyone?

155
00:14:12.415 --> 00:14:13.636
It's going to be the age of abundance.

156
00:14:13.956 --> 00:14:14.716
Everything's going to go well.

157
00:14:14.896 --> 00:14:16.097
What do you think they believe?

158
00:14:16.297 --> 00:14:17.357
What do you think Sam believes?

159
00:14:18.538 --> 00:14:24.940
So this is so funny, is such a core part of the mythology that they create around the AI industry,

160
00:14:25.760 --> 00:14:29.122
includes the belief that it could go very badly.

161
00:14:29.642 --> 00:14:30.923
It goes hand in hand.

162
00:14:31.483 --> 00:14:40.848
Like, they need that part of the myth in order to then say, and that's why we need to be in control of the technology because that's the only way that it's going to go really, really well.

163
00:14:41.569 --> 00:14:43.130
And Altman has said publicly, you know,

164
00:14:44.327 --> 00:14:46.428
The worst case lights out for everyone.

165
00:14:47.228 --> 00:14:51.210
But best case, we cure cancer, we solve climate change, and there's abundance.

166
00:14:51.850 --> 00:14:54.531
And Dario Amede, same kind of rhetoric.

167
00:14:54.951 --> 00:15:00.073
He's like, worst case, catastrophic or existential harm for humanity.

168
00:15:00.773 --> 00:15:02.994
Best case, mass human flourishing.

169
00:15:03.554 --> 00:15:05.835
So this is like two sides of the same coin.

170
00:15:05.855 --> 00:15:19.862
They have to use both of these narratives in order to continue justifying an extremely anti-democratic approach to AI development where there should not be broad participation in developing this technology.

171
00:15:20.002 --> 00:15:22.943
They must be the ones controlling it at every step of the way.

172
00:15:24.253 --> 00:15:28.135
Sam Altman did a tweet saying, there are some books coming out about OpenAI and me.

173
00:15:28.636 --> 00:15:30.097
We only participated in two of them.

174
00:15:30.957 --> 00:15:32.938
One by Kesh Hagee.

175
00:15:33.058 --> 00:15:33.559
Keech Hagee.

176
00:15:33.799 --> 00:15:35.000
Keech Hagee focused on me.

177
00:15:35.360 --> 00:15:37.641
And one by Ashley Vance on OpenAI.

178
00:15:39.702 --> 00:15:44.785
He went on to say, no book will get everything right, especially when some people are so intent on twisting things.

179
00:15:45.546 --> 00:15:47.207
But these two authors are trying to.

180
00:15:49.138 --> 00:15:56.608
you quote retweeted that tweet from Sam Altman and you said, the unnamed book Empire of AI is mine.

181
00:15:58.611 --> 00:16:02.416
Do you believe that tweet from Sam Altman was in reference to your book?

182
00:16:02.436 --> 00:16:05.020
100% because there's only three books coming out about him.

183
00:16:05.985 --> 00:16:07.906
And he'd caught wind that your book was coming out.

184
00:16:08.086 --> 00:16:13.949
He knew my book was coming out because I had contacted OpenAI from the very beginning of my process and said, I'm working on a book now.

185
00:16:14.449 --> 00:16:15.450
Will you participate in it?

186
00:16:16.030 --> 00:16:18.612
And actually, initially, they said yes.

187
00:16:18.892 --> 00:16:23.554
So my history with OpenAI, I profiled the company for MIT Technology Review.

188
00:16:23.694 --> 00:16:26.916
I embedded within the office for three days in 2019.

189
00:16:27.036 --> 00:16:28.116
My profile comes out in 2020.

190
00:16:29.457 --> 00:16:31.881
the leadership are very unhappy.

191
00:16:32.582 --> 00:16:39.874
And in my book, I actually quote an email that I received that Sam Altman sent to the company about my profile saying,

192
00:16:40.475 --> 00:16:41.555
Yeah, this is not great.

193
00:16:43.616 --> 00:16:54.680
And from then on, the company's stance to me was, we are not going to participate in anything that you do.

194
00:16:54.900 --> 00:16:57.721
We are not going to respond to anything, any questions that you receive.

195
00:16:57.841 --> 00:16:59.662
And this was, you know...

196
00:17:01.068 --> 00:17:02.909
This was things that they explicitly articulated.

197
00:17:02.929 --> 00:17:04.289
It wasn't like me inferring.

198
00:17:05.250 --> 00:17:09.472
So I had a colleague at MIT Technology Review that also covered AI.

199
00:17:10.032 --> 00:17:14.194
And at one point opening, I sent him this press release being like, we would love for you to cover this story.

200
00:17:14.214 --> 00:17:15.634
And he was like, I'm really busy.

201
00:17:16.135 --> 00:17:18.035
Will you send it to Karen?

202
00:17:18.356 --> 00:17:21.637
And they were like, oh, no, we have a history you understand.

203
00:17:22.657 --> 00:17:27.299
And so for three years, they refused to talk to me.

204
00:17:27.459 --> 00:17:27.960
But then...

205
00:17:28.840 --> 00:17:36.888
I ended up at the Wall Street Journal where if they felt a bit compelled because it was the journal to reopen the lines of communication.

206
00:17:37.909 --> 00:17:42.093
And so I started having, you know, more dialogue with them.

207
00:17:42.694 --> 00:17:45.757
Every time I wrote a piece, I would always send them, here's my request for comment.

208
00:17:46.197 --> 00:17:48.139
I would always ask them, like, will you sit for interviews?

209
00:17:48.679 --> 00:17:51.782
And we did get to a more productive relationship.

210
00:17:52.543 --> 00:17:53.864
And then I embarked on the book.

211
00:17:53.924 --> 00:17:56.427
So I left the journal to focus on the book full time.

212
00:17:57.227 --> 00:18:00.290
And I told them right away, I'm working on this book.

213
00:18:00.530 --> 00:18:09.098
I want to continue this productive conversation where I make sure I reflect OpenAI's perspective in the book.

214
00:18:09.598 --> 00:18:12.721
And so they were like, we can arrange interviews for you.

215
00:18:12.761 --> 00:18:14.082
You can come back to the office.

216
00:18:14.102 --> 00:18:14.503
Yeah.

217
00:18:14.903 --> 00:18:17.205
We'll set up some conversations.

218
00:18:18.346 --> 00:18:22.869
And then as we were going back and forth on this, the board fires Sam Altman.

219
00:18:24.731 --> 00:18:31.055
And that's when things started going kind of south because the company started becoming very sensitive to scrutiny.

220
00:18:31.896 --> 00:18:35.759
And so then they started pushing, kicking the can down the road, down the road, down the road.

221
00:18:35.799 --> 00:18:37.720
And I kept saying, hey, when are we rescheduling this?

222
00:18:37.760 --> 00:18:38.481
What's going on?

223
00:18:39.242 --> 00:18:40.542
And then I got an email saying,

224
00:18:41.383 --> 00:18:42.884
we are not going to participate at all.

225
00:18:42.924 --> 00:18:44.145
You are not coming to the office.

226
00:18:44.165 --> 00:18:44.925
You're not doing interviews.

227
00:18:44.945 --> 00:18:47.447
And I had actually already booked my tickets.

228
00:18:48.047 --> 00:18:53.090
So I was already going to fly to San Francisco to have the interviews.

229
00:18:53.890 --> 00:18:57.433
And so then I told them, I was like, that's fine.

230
00:18:58.393 --> 00:19:02.594
I will still engage in the process while I'll give you extensive requests for comment.

231
00:19:02.694 --> 00:19:09.215
I'll, through my reporting, I'll keep you updated on all the things that I'm finding so that you can choose to still comment.

232
00:19:09.955 --> 00:19:12.935
I gave them 40 pages of requests for comment.

233
00:19:13.776 --> 00:19:16.916
And I gave them over a month to respond to all of that.

234
00:19:17.276 --> 00:19:25.478
So this was when the tweet came out, was we were doing all this back and forth, trying to... And that's when Altman tweeted this.

235
00:19:27.658 --> 00:19:27.798
Hmm.

236
00:19:29.437 --> 00:19:32.581
And they never responded to a single one of the 40 pages.

237
00:19:32.941 --> 00:19:34.163
Sam Wattman does a lot of interviews.

238
00:19:34.964 --> 00:19:35.124
Yeah.

239
00:19:35.585 --> 00:19:37.227
You know, he's doing a lot of interviews all the time.

240
00:19:37.267 --> 00:19:39.149
He's done every podcast.

241
00:19:39.169 --> 00:19:43.294
I've seen him on everything from Tucker Carlson to, I think he's done Theo Von, Joe Rogan.

242
00:19:44.536 --> 00:19:45.437
Podcasts all over the world.

243
00:19:46.478 --> 00:19:47.760
I wonder why he won't do mine.

244
00:19:52.601 --> 00:19:53.041
Well, maybe.

245
00:19:53.061 --> 00:19:54.402
I don't know why.

246
00:19:54.482 --> 00:19:55.123
I don't know.

247
00:19:55.163 --> 00:19:56.544
I think I'm fair with everyone.

248
00:19:56.684 --> 00:19:58.305
I just ask questions I genuinely care about.

249
00:19:58.466 --> 00:20:02.069
I don't come in with huge preconceptions or at least meet people for the first time.

250
00:20:02.649 --> 00:20:07.453
But I've heard through the grapevine that he doesn't want to do mine.

251
00:20:07.913 --> 00:20:10.035
I mean, going back to what you were saying earlier, that...

252
00:20:11.933 --> 00:20:18.435
With this, the way that OpenAI and these companies control research, you asked, do they also do this with journalists?

253
00:20:19.555 --> 00:20:21.496
I mean, yes, the answer is yes.

254
00:20:21.616 --> 00:20:26.678
And apparently they also do it with anyone who has, you know, a broad mass communications platform.

255
00:20:28.189 --> 00:20:31.770
It's not just about the conversation that you're going to have with them.

256
00:20:32.291 --> 00:20:34.831
It's about who you also choose to platform.

257
00:20:36.012 --> 00:20:46.376
And there's this huge problem in technology journalism where companies know that a really big carrot that they can give to technology journalists is access.

258
00:20:46.556 --> 00:20:46.996
Yeah, yeah, yeah.

259
00:20:47.576 --> 00:20:56.662
And they will withhold that access at the drop of a hat if they catch wind that you're speaking to someone that they didn't want you to speak to.

260
00:20:57.083 --> 00:20:57.783
This is so true.

261
00:20:57.803 --> 00:21:00.985
And I don't think the average person really truly understands this.

262
00:21:01.266 --> 00:21:01.486
Yeah.

263
00:21:01.726 --> 00:21:07.490
So this kind of sounds like theory as you say it, but I'm not going to name names here because I don't think it's important.

264
00:21:08.010 --> 00:21:10.652
But there is a particular person in AI who...

265
00:21:12.722 --> 00:21:17.883
whose team have basically dangled the carrot of them coming here for like 18 months.

266
00:21:17.984 --> 00:21:19.964
And I'm like, you don't have to dangle the carrot.

267
00:21:19.984 --> 00:21:22.465
I'm going to speak to whoever I want to, regardless of the carrot or not.

268
00:21:22.865 --> 00:21:26.106
And when this person comes, if they want to come, I'll give them a fair shot.

269
00:21:26.126 --> 00:21:29.907
I'll ask them all genuinely curious questions about what they're doing, their incentives.

270
00:21:30.007 --> 00:21:31.327
I won't gotcha them.

271
00:21:31.527 --> 00:21:33.328
I don't have a history of ever gotchering anybody.

272
00:21:34.348 --> 00:21:36.529
Even if I have a different of opinion, I'll ask the question.

273
00:21:36.629 --> 00:21:36.809
Yeah.

274
00:21:37.289 --> 00:21:41.712
But they dangle carrots and they say, well, if, you know, he's thinking about it, let's think about a date.

275
00:21:41.732 --> 00:21:51.737
And what the strategy is, and I don't think they think most people don't understand, is if we just dangle it for long enough, then they will perform in the way that we want them to do.

276
00:21:51.937 --> 00:21:52.598
And they'll be...

277
00:21:54.313 --> 00:21:55.955
They'll be pleasant about us.

278
00:21:56.095 --> 00:21:57.036
They won't be critical.

279
00:21:57.056 --> 00:21:58.637
They won't give a voice.

280
00:21:58.657 --> 00:22:00.058
They won't platform our critics.

281
00:22:00.119 --> 00:22:00.599
Our critics.

282
00:22:01.260 --> 00:22:03.982
And I think a lot of their game is just dangle the carrot forever.

283
00:22:04.383 --> 00:22:04.603
Yes.

284
00:22:04.863 --> 00:22:05.043
Yeah.

285
00:22:05.063 --> 00:22:09.487
That's like the optimal outcome is if we just dangle it, if we just tell them, yeah, no, we're just trying to look at the schedule.

286
00:22:10.548 --> 00:22:11.269
It just doesn't work.

287
00:22:11.289 --> 00:22:16.054
I think in the modern world, you just have to go there and give your opinion and allow the clash of ideas in the public forum.

288
00:22:16.074 --> 00:22:19.397
Let the viewers decide for themselves what they think.

289
00:22:19.932 --> 00:22:35.661
Yeah, but this is such a huge part of their machinery is the way that they use these tactics to massage the public image of these companies and make sure that information that they don't want out and even opinions that they don't want out there go out there.

290
00:22:36.502 --> 00:22:44.686
And so this is, you know, I feel very lucky now that opening eyes shut the door early on me.

291
00:22:44.706 --> 00:22:45.327
Mm-hmm.

292
00:22:45.767 --> 00:22:47.288
At the time, I didn't feel lucky.

293
00:22:47.348 --> 00:22:49.250
I felt like I had screwed myself over.

294
00:22:49.270 --> 00:22:53.493
I was like, should I have been nicer to them in the profile so that I could maintain access?

295
00:22:53.874 --> 00:22:56.176
Which is a horrible question to ask as a journalist, right?

296
00:22:56.196 --> 00:23:01.920
Like, you're supposed to report the truth, and you're always supposed to report in the interest of the public.

297
00:23:01.940 --> 00:23:03.722
Like, that is the point of journalism.

298
00:23:04.362 --> 00:23:08.085
And in that moment, I was, like, relatively junior in my career.

299
00:23:08.165 --> 00:23:08.506
I was like—

300
00:23:09.126 --> 00:23:12.652
did I misunderstand what journalism is about?

301
00:23:12.712 --> 00:23:16.298
Like, should I have actually been playing the access game?

302
00:23:16.319 --> 00:23:16.759
Mm-hmm.

303
00:23:17.358 --> 00:23:18.098
But it was too late.

304
00:23:18.319 --> 00:23:19.439
I had the door shut to me.

305
00:23:20.079 --> 00:23:25.422
And so I had to build my career understanding that the front door was never gonna be open.

306
00:23:26.543 --> 00:23:33.647
And that actually really strengthened my own ability to just tell it like it is.

307
00:23:33.747 --> 00:23:34.327
The objective.

308
00:23:34.507 --> 00:23:41.231
Yeah, and just report what I see are the facts being presented to me, irrespective of whether the company likes it or not.

309
00:23:41.631 --> 00:23:44.793
And most often the company really does not like it, but,

310
00:23:46.032 --> 00:23:47.414
I can continue to do the work.

311
00:23:47.534 --> 00:23:48.975
They don't need to open the front door for me.

312
00:23:49.236 --> 00:23:52.099
I was still able to do more than 300 interviews.

313
00:23:54.021 --> 00:23:58.446
What you just listened to was a most replayed moment from a previous episode.

314
00:23:58.566 --> 00:24:01.730
If you want to listen to that full episode, I've linked it down below.

315
00:24:01.930 --> 00:24:02.751
Check the description.

316
00:24:03.031 --> 00:24:03.372
Thank you.