1
00:00:02,160 --> 00:00:03,400
Speaker 1: Also media.

2
00:00:04,640 --> 00:00:07,120
Speaker 2: Hello and welcome to Better Offline. I'm your host ed

3
00:00:07,240 --> 00:00:21,680
Speaker 2: ze tron. Not really going to dawdle too much on

4
00:00:21,680 --> 00:00:24,000
Speaker 2: the intro today, I'm too excited about the interview you're

5
00:00:24,000 --> 00:00:27,440
Speaker 2: about to hear. Today's guest. There's a globally known economist,

6
00:00:27,680 --> 00:00:30,440
Speaker 2: the author of wy Nations, Fall and Power and Progress.

7
00:00:30,880 --> 00:00:35,239
Speaker 2: Am I tas the rhone as sema glue. Okay, So

8
00:00:36,120 --> 00:00:40,800
Speaker 2: a term that you've popularized, not necessarily invented, is creative destruction.

9
00:00:41,080 --> 00:00:43,040
Speaker 2: Do you mind explaining it for the listeners?

10
00:00:43,400 --> 00:00:45,760
Speaker 3: Oh? Yeah, I mean I definitely did not invent it,

11
00:00:45,800 --> 00:00:49,320
Speaker 3: and I think many other people deserve much more credit

12
00:00:49,400 --> 00:00:52,839
Speaker 3: for inventing it and making it work. It's this idea

13
00:00:53,120 --> 00:00:57,840
Speaker 3: that goes back to Joseph Schumpeter, famous Austrian economist who

14
00:00:57,920 --> 00:01:00,840
Speaker 3: spent most of his career or the most important part

15
00:01:00,840 --> 00:01:06,440
Speaker 3: of his career, at Harvard, who emphasized that in capitalist growth,

16
00:01:07,959 --> 00:01:13,880
Speaker 3: you will have new firms taking market away and destroying

17
00:01:13,920 --> 00:01:19,600
Speaker 3: old firms, and as a corollary of that, new technologies

18
00:01:20,160 --> 00:01:24,880
Speaker 3: taking market share away and driving out old technologies. And

19
00:01:25,360 --> 00:01:30,440
Speaker 3: he understood this was a difficult and tumultuous process, but

20
00:01:30,600 --> 00:01:35,000
Speaker 3: also believe that that was the essence of sort of

21
00:01:35,000 --> 00:01:40,039
Speaker 3: capitalist growth. So it's one of these things that is

22
00:01:40,080 --> 00:01:44,360
Speaker 3: a fact of life in a market process, but different

23
00:01:44,520 --> 00:01:49,800
Speaker 3: types of social, economic, and political reactions to it are natural,

24
00:01:50,000 --> 00:01:53,280
Speaker 3: and how you react to it is going to have

25
00:01:53,680 --> 00:01:56,200
Speaker 3: various effects both on growth, what type of growth, and

26
00:01:56,240 --> 00:01:57,680
Speaker 3: its distributional effects.

27
00:01:58,560 --> 00:02:01,560
Speaker 2: Right, So something I've read about and spoken about a

28
00:02:01,600 --> 00:02:03,560
Speaker 2: lot as well as like this idea of the rot economy,

29
00:02:03,600 --> 00:02:06,600
Speaker 2: which is the kind of growth has overtaken most of

30
00:02:06,640 --> 00:02:10,240
Speaker 2: the modern markets. And I'd argue, at least in the

31
00:02:10,400 --> 00:02:15,800
Speaker 2: creative destruction field, tech has stopped really innovative. It doesn't

32
00:02:15,840 --> 00:02:19,200
Speaker 2: feel like they're creating things to create new jobs, to

33
00:02:19,200 --> 00:02:22,919
Speaker 2: create new markets. I'm wondering how you feel about looking

34
00:02:22,919 --> 00:02:24,240
Speaker 2: at the general tech industry.

35
00:02:24,960 --> 00:02:28,880
Speaker 3: I am a critic of the tech industry, and I

36
00:02:28,919 --> 00:02:33,799
Speaker 3: have become so over the last decade or so. And

37
00:02:34,200 --> 00:02:38,160
Speaker 3: my problem with the tech industry is not its dynamism.

38
00:02:38,360 --> 00:02:41,840
Speaker 3: I applaud that. It's not its risk taking. I applaud that,

39
00:02:42,720 --> 00:02:46,520
Speaker 3: And it's not the drive towards economic growth, which I

40
00:02:46,560 --> 00:02:53,040
Speaker 3: think is also generally desirable. But it is the direction

41
00:02:53,880 --> 00:02:57,760
Speaker 3: of research and technologies that the tech industry has focused on.

42
00:02:58,880 --> 00:03:03,760
Speaker 3: Both because of idea logical reasons and because of a

43
00:03:03,840 --> 00:03:08,040
Speaker 3: particular business model that they developed, and I think both

44
00:03:08,080 --> 00:03:12,840
Speaker 3: of those have pushed us towards technologies that I see

45
00:03:12,840 --> 00:03:17,640
Speaker 3: as socially less desirable in some cases actually undesirable, and

46
00:03:17,680 --> 00:03:21,000
Speaker 3: as a result, we're actually getting growth without as much

47
00:03:21,400 --> 00:03:25,600
Speaker 3: social benefits. And let me try to just make one

48
00:03:25,720 --> 00:03:29,680
Speaker 3: very simple point, which everybody nods when you say that,

49
00:03:29,760 --> 00:03:32,720
Speaker 3: but it's still as important to put in the conversation.

50
00:03:33,240 --> 00:03:38,200
Speaker 3: Economic output, as measured by statistical agencies such as gross

51
00:03:38,200 --> 00:03:43,280
Speaker 3: domestic product, does not have any welfare element in it.

52
00:03:44,320 --> 00:03:48,440
Speaker 3: So if I find a way of hacking into your

53
00:03:48,480 --> 00:03:52,360
Speaker 3: computer spending one thousand dollars, and you find a way

54
00:03:52,360 --> 00:03:57,400
Speaker 3: of defending against me spending two thousand dollars, that will

55
00:03:57,560 --> 00:04:00,960
Speaker 3: increase GDP by three thousand dollars. And I think even

56
00:04:01,000 --> 00:04:04,080
Speaker 3: the most demented person wouldn't say that that's a social improvement.

57
00:04:04,440 --> 00:04:05,880
Speaker 2: Yeah, I think you made a point like this the

58
00:04:05,920 --> 00:04:08,400
Speaker 2: Golden Zachs, where it's about like you could make a

59
00:04:08,480 --> 00:04:11,280
Speaker 2: trillion dollars if you did deep fakes in a certain

60
00:04:11,320 --> 00:04:12,520
Speaker 2: manner exactly.

61
00:04:13,440 --> 00:04:18,039
Speaker 3: So, therefore, new products that increase GDP may have socially

62
00:04:18,120 --> 00:04:25,040
Speaker 3: undesirable consequences. That wasn't part of the original Schumpeter point,

63
00:04:25,440 --> 00:04:27,680
Speaker 3: and it's not something that I would worry about when

64
00:04:27,720 --> 00:04:31,560
Speaker 3: I'm talking to people in Mexico who are trying to

65
00:04:31,560 --> 00:04:34,040
Speaker 3: get the economy going, But it is a very important

66
00:04:34,080 --> 00:04:35,479
Speaker 3: concern when it comes to new tech.

67
00:04:35,720 --> 00:04:37,520
Speaker 2: When do you think this shift happened? You say, the

68
00:04:37,600 --> 00:04:40,240
Speaker 2: last decade? What was it that kind of changed for you?

69
00:04:40,880 --> 00:04:45,000
Speaker 3: Well, I think it was probably a gradual process. But

70
00:04:45,400 --> 00:04:50,400
Speaker 3: the tech sector initially was very heavy on hardware with

71
00:04:50,560 --> 00:04:55,480
Speaker 3: some software elements, right, And when that started changing and

72
00:04:55,560 --> 00:04:59,640
Speaker 3: the entire field became software, I think the possibilities for

73
00:04:59,680 --> 00:05:03,360
Speaker 3: different diferent types of technologies to go in very different

74
00:05:03,440 --> 00:05:09,919
Speaker 3: social directions also multiplied. Right. So money today in Vidia

75
00:05:09,920 --> 00:05:13,960
Speaker 3: being an exception, is not made by hardware, And even

76
00:05:13,960 --> 00:05:16,520
Speaker 3: in Nvidia, I think a lot of the innovation is with.

77
00:05:17,080 --> 00:05:20,359
Speaker 2: Software, particularly with Kuda being able to do stuff with

78
00:05:20,400 --> 00:05:21,640
Speaker 2: GPS exactly.

79
00:05:21,680 --> 00:05:25,760
Speaker 3: Yeah. But when you are also doing software, you have

80
00:05:26,400 --> 00:05:29,440
Speaker 3: ways in which that software becomes an information control tool,

81
00:05:29,680 --> 00:05:32,600
Speaker 3: a monitoring tool, or surveillance tool. It becomes a way

82
00:05:32,640 --> 00:05:36,359
Speaker 3: of automating work in various different ways. It can become

83
00:05:36,440 --> 00:05:40,159
Speaker 3: a manipulative tool, and it can also create lots of

84
00:05:40,240 --> 00:05:42,800
Speaker 3: new products, some of them very beneficial, but some of

85
00:05:42,839 --> 00:05:46,680
Speaker 3: them very addictive and conducive to mental health problems. So

86
00:05:46,680 --> 00:05:52,120
Speaker 3: I think software sort of expands the capabilities, but together

87
00:05:52,160 --> 00:05:56,200
Speaker 3: with the capabilities, you also have expanded set of distortionary

88
00:05:56,240 --> 00:05:57,840
Speaker 3: or manipulative things that you can do.

89
00:05:59,040 --> 00:06:02,159
Speaker 2: Right, you mentioned what kind of dynamism element of how

90
00:06:02,240 --> 00:06:04,760
Speaker 2: tech is working in the growth? How if I agree

91
00:06:05,360 --> 00:06:08,080
Speaker 2: that tech is in a dynistic state, it almost feels

92
00:06:08,120 --> 00:06:12,040
Speaker 2: like it's been spinning its wheels for the last few years. Crypto, metaverse,

93
00:06:12,080 --> 00:06:13,920
Speaker 2: all of this stuff. It doesn't feel like new things

94
00:06:13,920 --> 00:06:14,520
Speaker 2: are happening.

95
00:06:15,279 --> 00:06:17,400
Speaker 3: Well, they are new things, it's just that I think

96
00:06:17,440 --> 00:06:22,040
Speaker 3: you are saying what I just said in a different way,

97
00:06:22,040 --> 00:06:24,600
Speaker 3: and it might be that your way is better. They're

98
00:06:24,600 --> 00:06:27,400
Speaker 3: not super socially valued, but they are new products. So

99
00:06:27,440 --> 00:06:32,040
Speaker 3: you would say that metaverse or virtual worlds are a

100
00:06:32,040 --> 00:06:35,200
Speaker 3: new product by any category. It's not just not something

101
00:06:35,240 --> 00:06:37,240
Speaker 3: that's going to make humanity better if it might make

102
00:06:37,320 --> 00:06:41,120
Speaker 3: humanity worse by alienating them more or isolating them more

103
00:06:41,160 --> 00:06:44,120
Speaker 3: from their social milieu. You know. And there is another

104
00:06:44,160 --> 00:06:49,120
Speaker 3: element of what's going on here, which I'll comment on

105
00:06:49,200 --> 00:06:52,200
Speaker 3: but again it doesn't contradict that they are generating new products.

106
00:06:52,240 --> 00:06:55,320
Speaker 3: A lot of new technologies and new ideas that are

107
00:06:55,360 --> 00:06:59,680
Speaker 3: being invented in tech are not being implemented, and part

108
00:06:59,760 --> 00:07:03,880
Speaker 3: of the reason for that is the competitive environment. Google, Facebook, Microsoft,

109
00:07:03,920 --> 00:07:07,880
Speaker 3: Amazon are all buying up on Facebook, are all buying

110
00:07:07,960 --> 00:07:11,800
Speaker 3: up a lot of competitors and not even sometimes using

111
00:07:11,840 --> 00:07:15,160
Speaker 3: their technology. So that's the consolidated structure. So the invention

112
00:07:15,280 --> 00:07:18,760
Speaker 3: is there, but the invention is not translating into implementation. Now,

113
00:07:18,920 --> 00:07:21,440
Speaker 3: don't get me wrong, some of that invention may also

114
00:07:21,480 --> 00:07:24,360
Speaker 3: go in the wrong way, so a better version of

115
00:07:25,400 --> 00:07:28,520
Speaker 3: you know, TikTok may not be a great thing either,

116
00:07:28,800 --> 00:07:34,240
Speaker 3: But there is that consolidated, concentrated market structure that is

117
00:07:34,400 --> 00:07:36,760
Speaker 3: also changing what gets implemented.

118
00:07:37,040 --> 00:07:39,120
Speaker 2: The thing I'll push back on is they are making

119
00:07:39,240 --> 00:07:43,000
Speaker 2: new things, but it feels like the ones I mentioned Crypto,

120
00:07:43,320 --> 00:07:48,600
Speaker 2: Metaverse and now Generative AI. They're not producing actual products

121
00:07:48,600 --> 00:07:50,280
Speaker 2: at the end. It's not so much that we couldn't

122
00:07:50,320 --> 00:07:52,560
Speaker 2: live in a virtual world or that digital money wonn't

123
00:07:52,560 --> 00:07:56,600
Speaker 2: be usedful, but it's more that the actual output from

124
00:07:56,640 --> 00:08:00,000
Speaker 2: the companies is not translating into meaningful products.

125
00:08:01,000 --> 00:08:03,640
Speaker 3: Well, again, this is a question of what we are measuring,

126
00:08:03,680 --> 00:08:05,480
Speaker 3: and whether what we're measuring is the right thing, and

127
00:08:05,520 --> 00:08:08,720
Speaker 3: whether it's really welfare relevant. But if I create a

128
00:08:08,760 --> 00:08:14,240
Speaker 3: metaverse and you're willing to pay a million dollars for it,

129
00:08:14,960 --> 00:08:17,520
Speaker 3: that will increase GDP by a million dollars. So that's

130
00:08:17,560 --> 00:08:19,920
Speaker 3: a new service, right. So a lot of things we

131
00:08:20,000 --> 00:08:24,000
Speaker 3: consume today are services. It's not something produced like a

132
00:08:24,040 --> 00:08:28,200
Speaker 3: T shirt or a car. They are based on digital services. Now,

133
00:08:28,200 --> 00:08:30,480
Speaker 3: of course, to produce those digital services, we're actually using

134
00:08:30,480 --> 00:08:33,640
Speaker 3: real resources such as energy. But what I actually buy

135
00:08:33,640 --> 00:08:35,520
Speaker 3: from you maybe, or what you buy from you maybe

136
00:08:35,559 --> 00:08:39,240
Speaker 3: just a digital service. Now, some digital services are extremely useful,

137
00:08:39,760 --> 00:08:41,440
Speaker 3: and some of them are useless. Some of them may

138
00:08:41,480 --> 00:08:42,120
Speaker 3: be bad for.

139
00:08:42,040 --> 00:08:47,000
Speaker 2: You, right. I think my point is more they're not

140
00:08:47,120 --> 00:08:50,400
Speaker 2: making particularly useful Selvis. They're doing well monetizing the things

141
00:08:50,400 --> 00:08:52,640
Speaker 2: they've had for years. But it kind of reminds me

142
00:08:52,679 --> 00:08:55,760
Speaker 2: of something he wrote in twenty nineteen where you were

143
00:08:55,760 --> 00:09:00,680
Speaker 2: talking about automation and effect on growth, but also we

144
00:09:00,760 --> 00:09:03,520
Speaker 2: may have run out of ideas for generating new high productivity,

145
00:09:03,600 --> 00:09:06,600
Speaker 2: labor intensive tasks. Do you think we're approaching that point?

146
00:09:07,240 --> 00:09:09,160
Speaker 3: Thank you for raising that, ed So. I think that

147
00:09:09,240 --> 00:09:12,240
Speaker 3: is a very very important part of my thinking. I

148
00:09:12,240 --> 00:09:16,599
Speaker 3: would say that the tech sector is not producing sufficient

149
00:09:16,960 --> 00:09:21,360
Speaker 3: new tasks for workers to use their skills and to

150
00:09:21,400 --> 00:09:25,000
Speaker 3: expand their capabilities, and firms perhaps are not demanding and

151
00:09:25,040 --> 00:09:28,960
Speaker 3: implementing enough of them. But it is not, according to me,

152
00:09:30,600 --> 00:09:33,480
Speaker 3: because we're running out of possibilities. It's just that we

153
00:09:33,520 --> 00:09:36,720
Speaker 3: haven't focused on those. And that's where the ideology reference

154
00:09:36,800 --> 00:09:39,840
Speaker 3: earlier on that I made comes in. You know, I

155
00:09:39,880 --> 00:09:46,559
Speaker 3: think the software industry would have done somewhat more productive

156
00:09:46,640 --> 00:09:52,679
Speaker 3: things if it did not become too focused on replacing

157
00:09:52,760 --> 00:09:57,920
Speaker 3: humans having machines as humans overlords, which today has of

158
00:09:57,960 --> 00:10:00,680
Speaker 3: course reached its apex with the craze on AGI.

159
00:10:05,880 --> 00:10:07,720
Speaker 2: But here's the thing. I get that that might be

160
00:10:07,760 --> 00:10:11,640
Speaker 2: what they're pushing toward, But generative AI isn't even automation

161
00:10:11,960 --> 00:10:13,520
Speaker 2: at this point. A lot of what you've written about

162
00:10:13,520 --> 00:10:16,520
Speaker 2: AI is correctly like the effects of if we all

163
00:10:16,600 --> 00:10:19,640
Speaker 2: make these tasks, but it feels that they aren't even

164
00:10:19,800 --> 00:10:21,800
Speaker 2: successful automating anything.

165
00:10:21,520 --> 00:10:25,640
Speaker 3: Absolutely one percent. Thank you for saying that. My take

166
00:10:25,800 --> 00:10:30,800
Speaker 3: is that generative AI is actually an informational tool, right,

167
00:10:31,160 --> 00:10:38,800
Speaker 3: so you should use generative AI as a way of generating, filtering, summarizing, finding,

168
00:10:39,400 --> 00:10:44,640
Speaker 3: checking information. That's actually what it's good at. If you

169
00:10:44,800 --> 00:10:47,400
Speaker 3: try to use it for other purposes, sometimes you can

170
00:10:47,440 --> 00:10:49,199
Speaker 3: get away with it, but it won't be very good

171
00:10:49,280 --> 00:10:52,000
Speaker 3: at that. So you can try to automate a lot

172
00:10:52,040 --> 00:10:56,760
Speaker 3: of warehouse tasks today by using the current crop of robots.

173
00:10:57,440 --> 00:11:00,000
Speaker 3: People don't do that because they are not good at it.

174
00:11:01,200 --> 00:11:03,960
Speaker 3: If you did it, costs would go up, people would

175
00:11:04,000 --> 00:11:06,920
Speaker 3: lose their jobs, Delays would pile up. But you can

176
00:11:07,080 --> 00:11:09,520
Speaker 3: do it. It's the same thing with generative AI. Even

177
00:11:09,520 --> 00:11:12,600
Speaker 3: though it's not an automation tool. Automation wouldn't be its

178
00:11:12,640 --> 00:11:17,080
Speaker 3: best use, especially given the current unreliability, even though there

179
00:11:17,160 --> 00:11:19,480
Speaker 3: is something else that we could do better with it.

180
00:11:20,200 --> 00:11:21,960
Speaker 3: I think many people are going to use it for

181
00:11:22,000 --> 00:11:25,319
Speaker 3: automation because that's the vibe. That's what companies are being told.

182
00:11:25,720 --> 00:11:28,760
Speaker 3: You know, if you talk to business leaders today, everybody's

183
00:11:28,800 --> 00:11:32,800
Speaker 3: asking them, financial journalists, their shareholders, and their friends, where

184
00:11:32,840 --> 00:11:36,120
Speaker 3: are you with the AI investment? So that's the hype,

185
00:11:37,160 --> 00:11:39,480
Speaker 3: and then people are going to rush into use AI

186
00:11:39,600 --> 00:11:41,600
Speaker 3: to implement AI even when they don't know what to

187
00:11:41,640 --> 00:11:44,720
Speaker 3: do with it, and automation will often appeal to them.

188
00:11:44,760 --> 00:11:46,800
Speaker 3: Because it's like the easiest thing to do. It's the

189
00:11:46,840 --> 00:11:50,680
Speaker 3: thing that they may have experienced from other technologies, and

190
00:11:50,720 --> 00:11:52,440
Speaker 3: it's the thing that some people are telling them that

191
00:11:52,440 --> 00:11:55,840
Speaker 3: that's what they should do. There are companies, integrators' websites

192
00:11:55,880 --> 00:11:58,920
Speaker 3: devoted to automation AI, even though it wouldn't be very

193
00:11:58,920 --> 00:12:02,040
Speaker 3: good at it. I mean again, it could automate some tasks.

194
00:12:02,360 --> 00:12:06,760
Speaker 3: You could have more of your customer service done without people.

195
00:12:07,600 --> 00:12:10,320
Speaker 2: Right, even then, that feels like a stretch of what

196
00:12:10,400 --> 00:12:14,080
Speaker 2: automation means because sure, the customer service example, and I

197
00:12:14,120 --> 00:12:16,760
Speaker 2: think you may have raised this point as well, how

198
00:12:16,760 --> 00:12:19,320
Speaker 2: does it get better? How do you measure better in

199
00:12:19,320 --> 00:12:22,240
Speaker 2: the case of customer service? But even then, it's automation

200
00:12:22,440 --> 00:12:24,680
Speaker 2: only in so much as you can trust it, And

201
00:12:24,720 --> 00:12:28,079
Speaker 2: it feels like these core issues of hallucination almost kill

202
00:12:28,920 --> 00:12:31,000
Speaker 2: the concept of automation with generative AI.

203
00:12:32,320 --> 00:12:35,600
Speaker 3: So here's a good use case for automation of customer service,

204
00:12:35,760 --> 00:12:38,720
Speaker 3: which is you call your bank and you enter some

205
00:12:38,760 --> 00:12:42,400
Speaker 3: password and they tell you your balance. That's perfect. You

206
00:12:42,440 --> 00:12:45,120
Speaker 3: don't need a person there to tell you the balance, right,

207
00:12:45,440 --> 00:12:50,680
Speaker 3: because the current technologies can faithfully take those numbers and

208
00:12:50,720 --> 00:12:53,280
Speaker 3: communicate them to you. After the right security steps.

209
00:12:53,880 --> 00:12:56,320
Speaker 2: Yeah, and it's not a generative answer because it's a

210
00:12:56,520 --> 00:12:57,400
Speaker 2: number in a day abase.

211
00:12:57,480 --> 00:13:00,080
Speaker 3: It's not a generative answer. So now put generitive A

212
00:13:00,280 --> 00:13:02,480
Speaker 3: and there you're probably gonna get lots of incorrect answers. Yes,

213
00:13:02,960 --> 00:13:04,480
Speaker 3: but some companies might still do it.

214
00:13:04,520 --> 00:13:07,400
Speaker 2: Which is it just feels like a crazy time that

215
00:13:07,440 --> 00:13:10,320
Speaker 2: you have companies shoving this through almost like the very

216
00:13:10,400 --> 00:13:14,200
Speaker 2: much like a post Jack Welch situation where.

217
00:13:14,160 --> 00:13:16,600
Speaker 3: Yes, exactly the Jack Welsh mindset.

218
00:13:20,480 --> 00:13:22,080
Speaker 2: Do you think about the problem is that the people

219
00:13:22,160 --> 00:13:24,560
Speaker 2: running these companies aren't really technologists.

220
00:13:25,520 --> 00:13:29,000
Speaker 3: I don't know. Look, I think this is another branch

221
00:13:29,080 --> 00:13:35,839
Speaker 3: of my work. But US businesses are often led by

222
00:13:35,960 --> 00:13:41,320
Speaker 3: people who have been trained into thinking that their only

223
00:13:41,640 --> 00:13:45,800
Speaker 3: priority should be increasing short term shareholder value, right, and

224
00:13:46,520 --> 00:13:49,440
Speaker 3: a very effective way of doing that is cut labor costs.

225
00:13:50,000 --> 00:13:53,920
Speaker 3: But a that's not the right social objective. Even maximizing

226
00:13:53,960 --> 00:13:57,960
Speaker 3: long term shareholder value is not the right objective. But

227
00:14:00,760 --> 00:14:06,640
Speaker 3: even more fundamentally, cutting short run labor costs maybe an

228
00:14:06,640 --> 00:14:11,439
Speaker 3: illusion and be associated with longer term problems. So if

229
00:14:11,480 --> 00:14:15,360
Speaker 3: you have a company where your workers are skilled, talented,

230
00:14:15,400 --> 00:14:20,480
Speaker 3: they are very useful for liaising with customers. Creating new services, products, innovation.

231
00:14:21,320 --> 00:14:23,240
Speaker 3: You can in the short run cut your labor costs,

232
00:14:23,280 --> 00:14:25,360
Speaker 3: but it would destroy you in the long run. I

233
00:14:25,360 --> 00:14:27,880
Speaker 3: think many more companies are in this bucket than American

234
00:14:27,920 --> 00:14:29,080
Speaker 3: business leaders realize.

235
00:14:29,200 --> 00:14:30,760
Speaker 2: And it's funny you mentioned that. I've heard a lot

236
00:14:30,800 --> 00:14:33,400
Speaker 2: from Google people. In particular. I get emails from Google

237
00:14:33,400 --> 00:14:35,240
Speaker 2: people all the time because it did an episode about

238
00:14:35,240 --> 00:14:39,360
Speaker 2: a particular guy, and it's funny. They all talk about

239
00:14:39,400 --> 00:14:42,200
Speaker 2: the kind of brain drain of layoffs that you don't realize.

240
00:14:42,240 --> 00:14:45,160
Speaker 2: It's not just the output you're losing, it's the person

241
00:14:45,160 --> 00:14:47,880
Speaker 2: who knew how the staff worked and where the stuff was,

242
00:14:47,920 --> 00:14:49,920
Speaker 2: and who built the staff and why the stuff is

243
00:14:49,960 --> 00:14:54,040
Speaker 2: good or back. And it almost feels like American capitalism

244
00:14:54,120 --> 00:14:56,960
Speaker 2: is dramatically disconnected from labor in general.

245
00:14:57,680 --> 00:15:00,400
Speaker 3: Yeah. Absolutely so. Look, I mean, I think there is

246
00:15:00,520 --> 00:15:05,680
Speaker 3: a tremendous amount of tacit knowledge that workers have which

247
00:15:05,720 --> 00:15:11,400
Speaker 3: often gets unrecognized, and even bosses sometimes don't recognize that.

248
00:15:13,240 --> 00:15:16,960
Speaker 3: So both French and British trade unions in the history

249
00:15:17,280 --> 00:15:21,560
Speaker 3: experimented with these types of strikes where workers just follow

250
00:15:21,600 --> 00:15:25,080
Speaker 3: the rules. They do exactly what the rule book says

251
00:15:25,080 --> 00:15:28,760
Speaker 3: their responsibilities are, and it turns out to be quite

252
00:15:28,760 --> 00:15:32,320
Speaker 3: disastrous for the company because most of what actually workers

253
00:15:32,360 --> 00:15:35,520
Speaker 3: do is much more adaptive than just following the rule.

254
00:15:35,880 --> 00:15:37,680
Speaker 2: Like kind of outsourcing risk. Almost.

255
00:15:38,200 --> 00:15:40,440
Speaker 3: Yeah, it's just like, you know, the rule book says,

256
00:15:41,000 --> 00:15:43,200
Speaker 3: you know, operate the machinery, but you know when to

257
00:15:43,280 --> 00:15:46,800
Speaker 3: actually operate the machinery, not just operate the machinery. The

258
00:15:47,240 --> 00:15:52,160
Speaker 3: sound yeah exactly. Yeah. So that's the kind of tacit

259
00:15:52,240 --> 00:15:55,920
Speaker 3: knowledge that people acquire via training, why experienced by their

260
00:15:55,960 --> 00:16:00,280
Speaker 3: social network, talking to friends. And if we don't value that,

261
00:16:00,320 --> 00:16:02,200
Speaker 3: will lose that and it's going to be very difficult

262
00:16:02,240 --> 00:16:06,200
Speaker 3: to replace it with what machines or information technologies do.

263
00:16:06,200 --> 00:16:07,720
Speaker 2: Do you think we're in a bubble right now?

264
00:16:08,360 --> 00:16:09,160
Speaker 3: Define a bubble?

265
00:16:09,280 --> 00:16:12,760
Speaker 2: Actually, let me reframe the question. Do you think generative

266
00:16:12,760 --> 00:16:15,520
Speaker 2: AI is a trillion dollar industry? Do you actually think

267
00:16:15,520 --> 00:16:17,360
Speaker 2: it is the next hype of growth market?

268
00:16:17,640 --> 00:16:21,360
Speaker 3: Let me answer that question slightly indirectly.

269
00:16:21,560 --> 00:16:22,080
Speaker 2: Sounds good.

270
00:16:22,800 --> 00:16:26,800
Speaker 3: I believe that generative AI has the capacity to add

271
00:16:27,280 --> 00:16:32,120
Speaker 3: a trillion dollar or more over time if we use

272
00:16:32,160 --> 00:16:38,880
Speaker 3: it correctly, because as an information technology it has great capabilities.

273
00:16:38,920 --> 00:16:41,880
Speaker 3: We live in an age in which useful information is

274
00:16:41,920 --> 00:16:44,960
Speaker 3: scarce all sorts of chunk you don't want is on

275
00:16:45,000 --> 00:16:47,920
Speaker 3: the internet. But when you actually need to solve a problem,

276
00:16:48,240 --> 00:16:51,000
Speaker 3: get better at what you're doing, get more background information,

277
00:16:51,120 --> 00:16:55,000
Speaker 3: those things are very difficult to find, and generitive AI

278
00:16:55,080 --> 00:16:59,600
Speaker 3: could be a tool for providing that sort of information

279
00:16:59,680 --> 00:17:02,360
Speaker 3: to all all sorts of decision makers and workers blue

280
00:17:02,400 --> 00:17:06,159
Speaker 3: collar workers, office workers and so on. But that's not

281
00:17:06,240 --> 00:17:09,120
Speaker 3: the direction we're going in which case, I don't think

282
00:17:09,119 --> 00:17:13,280
Speaker 3: it's going to add trillions of dollars of true value.

283
00:17:14,040 --> 00:17:17,359
Speaker 3: But that also doesn't mean that generative AAI companies are

284
00:17:17,400 --> 00:17:19,560
Speaker 3: going to go bust, right, because they're going to be

285
00:17:19,560 --> 00:17:22,840
Speaker 3: able to monetize this in other ways. So if generative

286
00:17:22,840 --> 00:17:26,399
Speaker 3: AI enables you to take over the search market from Google,

287
00:17:27,480 --> 00:17:30,760
Speaker 3: that's a huge amount of money. It may take over

288
00:17:30,800 --> 00:17:34,439
Speaker 3: the search market from Google without providing much better service

289
00:17:34,480 --> 00:17:37,879
Speaker 3: to consumers, but it might still be hugely profitable. If

290
00:17:37,960 --> 00:17:42,480
Speaker 3: GENITTIVAI companies convince businesses to invest in generative AI, that's

291
00:17:42,480 --> 00:17:44,080
Speaker 3: going to be very profitable for them, but not so

292
00:17:44,160 --> 00:17:46,199
Speaker 3: good for the businesses that misimplement it. Right.

293
00:17:46,800 --> 00:17:49,840
Speaker 2: So the thing is, and I understand why you're making

294
00:17:49,880 --> 00:17:54,600
Speaker 2: these assumptions, but what if it doesn't get cheaper because

295
00:17:54,680 --> 00:17:56,639
Speaker 2: right now, the thing I've been on about with generative

296
00:17:56,640 --> 00:18:00,159
Speaker 2: AI is, on top of not being super useful, it's

297
00:18:00,240 --> 00:18:04,280
Speaker 2: so unprofitable and every report seems to be suggesting it

298
00:18:04,320 --> 00:18:07,840
Speaker 2: isn't making people money. What if it stays where it

299
00:18:07,880 --> 00:18:10,119
Speaker 2: is because in the last eighteen months, GPT four row

300
00:18:10,240 --> 00:18:13,960
Speaker 2: is not significantly different. What if they've stalled. What if

301
00:18:13,960 --> 00:18:14,800
Speaker 2: this is all we've got.

302
00:18:15,760 --> 00:18:19,400
Speaker 3: Yeah, my guess is that it will get somewhat cheaper

303
00:18:19,720 --> 00:18:24,840
Speaker 3: because right now it's very costly to even answer queries,

304
00:18:26,400 --> 00:18:29,439
Speaker 3: and with more GPU capacity it will get somewhat cheaper.

305
00:18:29,800 --> 00:18:33,000
Speaker 3: With better designs, it will get somewhat cheaper. But I

306
00:18:33,040 --> 00:18:36,000
Speaker 3: do not believe that there is a scaling law in there.

307
00:18:36,640 --> 00:18:40,240
Speaker 3: So many people in the industry believe in this mysterious

308
00:18:40,240 --> 00:18:43,159
Speaker 3: scaling law, which is that you double the GPU capacity

309
00:18:43,240 --> 00:18:47,320
Speaker 3: or comput capacity, you double the data, and you get

310
00:18:47,600 --> 00:18:48,560
Speaker 3: twice the performance.

311
00:18:48,720 --> 00:18:51,240
Speaker 2: Just an aberration of Moore's law by people who don't

312
00:18:51,280 --> 00:18:52,280
Speaker 2: necessarily understand it.

313
00:18:53,160 --> 00:18:54,920
Speaker 3: But first of all, what does it mean to say

314
00:18:55,000 --> 00:18:58,040
Speaker 3: double the data? We're going to throw more Reddit at it.

315
00:18:59,480 --> 00:19:03,119
Speaker 3: So even if there were such a scaling law, you

316
00:19:03,119 --> 00:19:05,960
Speaker 3: would require high quality data, which we're not producing. Is

317
00:19:06,000 --> 00:19:07,600
Speaker 3: run out. We're not paying for it.

318
00:19:07,720 --> 00:19:18,120
Speaker 2: Yeah, there's something just very nihilistic about the whole thing

319
00:19:18,560 --> 00:19:21,520
Speaker 2: as it stands. There's not really much of a social output.

320
00:19:21,520 --> 00:19:25,960
Speaker 2: It's helping automate away jobs prolominantly held by contractors, which

321
00:19:26,000 --> 00:19:29,040
Speaker 2: is already a problem onto itself. But also it doesn't

322
00:19:29,080 --> 00:19:31,560
Speaker 2: seem to be making the money. I don't think I've

323
00:19:31,680 --> 00:19:35,760
Speaker 2: ever seen anything like this in tech, and I'm just wondering. Indeed,

324
00:19:35,760 --> 00:19:38,480
Speaker 2: maybe this is the question what happens if this is

325
00:19:38,520 --> 00:19:42,480
Speaker 2: not it? But also TICK doesn't have a next step.

326
00:19:42,720 --> 00:19:44,600
Speaker 2: Do you think one of these big companies could die?

327
00:19:44,720 --> 00:19:46,960
Speaker 2: Do you think that there is actually an existential risk

328
00:19:47,119 --> 00:19:49,560
Speaker 2: if generative AI and all this force apart.

329
00:19:50,480 --> 00:19:53,639
Speaker 3: No, I don't think so. I think none of these companies,

330
00:19:54,320 --> 00:19:58,119
Speaker 3: you know, are just committed to generative AI. They have

331
00:19:58,240 --> 00:20:01,920
Speaker 3: other businesses that are making money, and even Nvidia can

332
00:20:01,960 --> 00:20:04,119
Speaker 3: make still a lot of money with the GPUs.

333
00:20:04,280 --> 00:20:06,840
Speaker 2: Let me rephrase it then, So right now, all of

334
00:20:06,880 --> 00:20:09,280
Speaker 2: these tech companies, they do very well in their multiples

335
00:20:09,320 --> 00:20:12,160
Speaker 2: in the markets because they have a relatively low cost

336
00:20:12,200 --> 00:20:15,040
Speaker 2: of goods, like their actual costs are pretty great, but

337
00:20:15,080 --> 00:20:18,840
Speaker 2: they're predicated on this ongoing growth. They must always grow.

338
00:20:18,880 --> 00:20:22,080
Speaker 2: But what happens if they don't have a new growth

339
00:20:22,080 --> 00:20:26,679
Speaker 2: thing because they haven't for a while, and what if

340
00:20:27,040 --> 00:20:29,320
Speaker 2: they turn on generative AI? Like this feels like this

341
00:20:29,359 --> 00:20:31,400
Speaker 2: could be an economic panic onto itself.

342
00:20:32,320 --> 00:20:35,200
Speaker 3: Yeah, it could be. It could be some drops in valuation.

343
00:20:35,960 --> 00:20:39,720
Speaker 3: The general pattern we have seen with many other products

344
00:20:39,760 --> 00:20:42,240
Speaker 3: and technologies is that it looks a little bit like

345
00:20:42,280 --> 00:20:46,600
Speaker 3: an S curve. Right, you're an acceleration and then you plateau,

346
00:20:47,040 --> 00:20:50,960
Speaker 3: and that's when new products are invented, new investors move

347
00:20:51,000 --> 00:20:54,400
Speaker 3: on to other things. And that hasn't happened with tech.

348
00:20:54,840 --> 00:20:59,879
Speaker 3: You know, Microsoft is living its fourth life or whatever

349
00:21:00,119 --> 00:21:04,359
Speaker 3: since a mestos, partly because they have acquired new businesses,

350
00:21:05,400 --> 00:21:10,320
Speaker 3: some competitors, some competing technologies, and sometimes some tech companies

351
00:21:10,320 --> 00:21:13,399
Speaker 3: have invested in the wrong things. I mean, cryptocurrency was

352
00:21:13,520 --> 00:21:16,920
Speaker 3: more crazy than AI. There. I really didn't see the

353
00:21:17,040 --> 00:21:17,639
Speaker 3: use case.

354
00:21:17,840 --> 00:21:19,800
Speaker 2: The question I keep and have asked a lot of people,

355
00:21:19,840 --> 00:21:23,159
Speaker 2: this is just what happens if there's nothing though, because

356
00:21:23,160 --> 00:21:27,480
Speaker 2: growth is slowing. There is a pattern of slowing growth

357
00:21:27,480 --> 00:21:31,520
Speaker 2: within these companies and there isn't a new thing that

358
00:21:31,520 --> 00:21:34,400
Speaker 2: they can pick up and acquire. I don't know whether

359
00:21:34,480 --> 00:21:37,440
Speaker 2: tech has ever had this happen is the problem.

360
00:21:37,680 --> 00:21:41,120
Speaker 3: Yeah, it's a good point, but it's even deeper than that.

361
00:21:41,280 --> 00:21:45,480
Speaker 3: Growth has slowed in the industrialized world, and it's not

362
00:21:45,520 --> 00:21:49,040
Speaker 3: a new phenomenon. Since one of these paradoxes, which needs

363
00:21:49,080 --> 00:21:51,960
Speaker 3: to be repeated more and more, the tech age has

364
00:21:52,000 --> 00:21:55,080
Speaker 3: also coincided with a slow down of aggregate growth and

365
00:21:55,160 --> 00:21:59,600
Speaker 3: every indicator of aggregate growth. So we are growing much

366
00:21:59,680 --> 00:22:03,200
Speaker 3: less today than we did in the seventies or sixties.

367
00:22:04,280 --> 00:22:09,359
Speaker 3: Productivity is growing less, and I think this is also

368
00:22:09,440 --> 00:22:12,320
Speaker 3: related to the fact that we're not getting enough out

369
00:22:12,320 --> 00:22:14,440
Speaker 3: of the new technologies and the new ideas and the

370
00:22:14,480 --> 00:22:18,399
Speaker 3: new scientific discoveries that we are making. And part of

371
00:22:18,440 --> 00:22:21,960
Speaker 3: the reason why there is so much hunger for AI

372
00:22:22,200 --> 00:22:25,240
Speaker 3: hype is that many people, including policymakers by the way,

373
00:22:25,400 --> 00:22:29,520
Speaker 3: are wishfully thinking, oh, well, this could be a solution

374
00:22:29,640 --> 00:22:32,720
Speaker 3: to our productivity slow down. So perhaps in the next

375
00:22:32,760 --> 00:22:36,240
Speaker 3: decade we can have a much faster productivity growth thanks

376
00:22:36,240 --> 00:22:37,480
Speaker 3: to generative AI or thanks to.

377
00:22:37,480 --> 00:22:39,720
Speaker 2: AI right, new jobs and such.

378
00:22:40,359 --> 00:22:42,680
Speaker 3: Yeah, new jobs, sudden discoveries.

379
00:22:43,200 --> 00:22:46,240
Speaker 2: So almost history is kind of slowing down. I've not

380
00:22:46,320 --> 00:22:48,560
Speaker 2: really heard anyone really discuss it in these terms, but

381
00:22:48,600 --> 00:22:51,160
Speaker 2: it's interesting. So you've seen that this is the growth

382
00:22:51,200 --> 00:22:54,320
Speaker 2: of all costs is everywhere, and growth is slowing. But

383
00:22:54,440 --> 00:22:56,760
Speaker 2: it sounds like growth isn't just about a money thing though.

384
00:22:56,960 --> 00:22:58,840
Speaker 3: No, no, growth is not just about money thing, and

385
00:22:58,880 --> 00:23:01,640
Speaker 3: I think if you do look at other indicators, we're

386
00:23:01,680 --> 00:23:05,680
Speaker 3: doing worse. One of the regularities of the twentieth century

387
00:23:05,840 --> 00:23:10,280
Speaker 3: across the world is that health and life expectancy have

388
00:23:10,359 --> 00:23:16,680
Speaker 3: improved every today. People in Sub Saharan Africa have twice

389
00:23:16,720 --> 00:23:19,720
Speaker 3: the life expectancy at birth as people who lived in

390
00:23:19,760 --> 00:23:23,800
Speaker 3: London or Manchester in the eighteen hundreds, and Americans have

391
00:23:23,920 --> 00:23:28,280
Speaker 3: had tremendous improvements in life expectancy and health until the

392
00:23:28,359 --> 00:23:31,480
Speaker 3: last decade when it's slow. Then it started getting reversed.

393
00:23:31,800 --> 00:23:34,960
Speaker 3: So on many indicators were actually doing even worse than

394
00:23:35,160 --> 00:23:37,240
Speaker 3: GDP suggests.

395
00:23:37,680 --> 00:23:40,920
Speaker 2: So what's contributing to it? Is it a welfare issue,

396
00:23:41,000 --> 00:23:42,480
Speaker 2: is a societal one?

397
00:23:42,560 --> 00:23:46,040
Speaker 3: Is it? Well? I don't think there is a clear answer.

398
00:23:46,119 --> 00:23:48,400
Speaker 3: Some people think it's because of you know, the life

399
00:23:48,400 --> 00:23:51,919
Speaker 3: expectancy part is because of early deaths due to alcoholism,

400
00:23:52,200 --> 00:23:55,720
Speaker 3: opioids and drugs, But there is a more general deterioration

401
00:23:55,880 --> 00:23:59,000
Speaker 3: mental health. There's a mental health crisis, So if you

402
00:23:59,119 --> 00:24:02,680
Speaker 3: look at health of surviving people, it's much worse if

403
00:24:02,720 --> 00:24:04,520
Speaker 3: you're factoring that mental health issue.

404
00:24:04,600 --> 00:24:07,040
Speaker 2: I wonder if it's also where tech falls into this,

405
00:24:07,080 --> 00:24:09,880
Speaker 2: as well as the exposure to social media. I've had

406
00:24:09,880 --> 00:24:12,919
Speaker 2: this overall, which is one of my Flimsiert theories. I

407
00:24:12,920 --> 00:24:15,160
Speaker 2: don't think people should be thinking about politics as much

408
00:24:15,160 --> 00:24:17,119
Speaker 2: as they do. Not saying people shouldn't be political, but

409
00:24:17,240 --> 00:24:21,879
Speaker 2: just the immediacy of political discussion has been irrosive to

410
00:24:21,920 --> 00:24:23,040
Speaker 2: people's mental health.

411
00:24:23,680 --> 00:24:26,880
Speaker 3: Well, I'll give you two factois that might perhaps support

412
00:24:26,920 --> 00:24:30,160
Speaker 3: your ideal, although I'm not sure whether I completely agree

413
00:24:30,200 --> 00:24:32,879
Speaker 3: with it. But one is that if you look at

414
00:24:32,920 --> 00:24:37,040
Speaker 3: when the mental health crisis seems to start, it coincides

415
00:24:37,040 --> 00:24:41,840
Speaker 3: with smartphones. Ah, so people accessing social media and other

416
00:24:41,880 --> 00:24:44,200
Speaker 3: things on their smart forms twenty four hours a day

417
00:24:44,880 --> 00:24:47,439
Speaker 3: might have something to do with it. Another one is

418
00:24:47,560 --> 00:24:52,359
Speaker 3: due to economists Alcott and Matthew Genskau. They did this

419
00:24:52,440 --> 00:24:58,240
Speaker 3: experiment where they incentivized Facebook users to stop using the platform.

420
00:24:58,359 --> 00:25:02,240
Speaker 3: So when people stop using the platform, their mental health improvement,

421
00:25:02,920 --> 00:25:05,280
Speaker 3: but they can answer questions about what's going on in

422
00:25:05,400 --> 00:25:09,560
Speaker 3: current politics much less well. So there at least immediate

423
00:25:09,640 --> 00:25:13,879
Speaker 3: superficial knowledge of what's going on in politics also declines.

424
00:25:14,560 --> 00:25:18,439
Speaker 2: Interesting, Yeah, it does feel like there is a wider discussion.

425
00:25:19,480 --> 00:25:22,160
Speaker 2: Discussion perhaps is the wrong word. Within the tech industry,

426
00:25:22,240 --> 00:25:25,760
Speaker 2: there is almost no consideration of the social aspects, of

427
00:25:25,800 --> 00:25:29,640
Speaker 2: the welfare aspects of any technology being buil the metaphus,

428
00:25:29,680 --> 00:25:33,359
Speaker 2: for example. As ridiculous as that was, I can understand

429
00:25:33,400 --> 00:25:35,600
Speaker 2: an executive being like, yeah, we use the internet, now

430
00:25:35,600 --> 00:25:39,199
Speaker 2: what if we use more Internet? But just no consideration

431
00:25:39,440 --> 00:25:42,560
Speaker 2: to whether people wanted to. It feels like there's just

432
00:25:42,600 --> 00:25:45,040
Speaker 2: a disconnection between capitalism and people.

433
00:25:45,400 --> 00:25:48,520
Speaker 3: I mean, I think tech is much more complicated. Oftentimes

434
00:25:48,600 --> 00:25:51,480
Speaker 3: it's multi use, so something that may appear to have

435
00:25:51,560 --> 00:25:55,640
Speaker 3: good users also has bad users. But I do think

436
00:25:55,680 --> 00:25:59,480
Speaker 3: that tech workers also need to own up to greater

437
00:25:59,520 --> 00:26:03,960
Speaker 3: social reds responsibility. Right, So, if you are a physicist

438
00:26:04,080 --> 00:26:08,040
Speaker 3: nuclear physicist today, it's unthinkable that you do not have

439
00:26:08,160 --> 00:26:12,800
Speaker 3: some social responsibility related knowledge as well as training about,

440
00:26:12,880 --> 00:26:14,160
Speaker 3: you know, nuclear weapons.

441
00:26:14,160 --> 00:26:16,960
Speaker 2: My best mate is a nuclear health and safety person,

442
00:26:17,000 --> 00:26:18,560
Speaker 2: so I feel will appreciate that.

443
00:26:19,480 --> 00:26:24,560
Speaker 3: But the same degree of thinking about ethical implications, social implications.

444
00:26:24,560 --> 00:26:28,280
Speaker 3: What happens if I unleash this on humanity doesn't quite

445
00:26:28,359 --> 00:26:30,840
Speaker 3: exist to the same extent in the tech industry, and

446
00:26:30,840 --> 00:26:33,000
Speaker 3: I think it's going to develop. There are many people

447
00:26:33,000 --> 00:26:36,600
Speaker 3: who are very socially minded in the tech sector, but

448
00:26:36,680 --> 00:26:39,159
Speaker 3: I think we may need something more systemic.

449
00:26:39,400 --> 00:26:41,960
Speaker 2: And what would regulation look like? On a better note,

450
00:26:42,000 --> 00:26:44,720
Speaker 2: I suppose what can we do to kind of reverse

451
00:26:44,760 --> 00:26:49,400
Speaker 2: this disconnected trend? Is it regulation? Is it better safety culture?

452
00:26:49,880 --> 00:26:52,800
Speaker 3: Well, all of the above. But here's a problem I

453
00:26:52,880 --> 00:26:56,040
Speaker 3: have with both regulation and the discussion that we have

454
00:26:56,119 --> 00:27:00,760
Speaker 3: about regulation. It is very reactive, right. Something happens and

455
00:27:00,800 --> 00:27:04,040
Speaker 3: we react to it by thinking of how can we

456
00:27:04,080 --> 00:27:08,280
Speaker 3: regulate so that we reduce the harms. But the problem,

457
00:27:08,440 --> 00:27:12,520
Speaker 3: as I try to articulate, including in the earlier parts

458
00:27:12,520 --> 00:27:16,000
Speaker 3: of this conversation, is about what types of technologies we

459
00:27:16,040 --> 00:27:20,040
Speaker 3: are developing, where we're putting our efforts. Expost regulation that's

460
00:27:20,119 --> 00:27:23,360
Speaker 3: reactive is not going to achieve that. So I think

461
00:27:23,400 --> 00:27:27,400
Speaker 3: we need a new tech culture and as well as

462
00:27:27,400 --> 00:27:32,680
Speaker 3: societal norms and priorities that says there is an alternative

463
00:27:33,040 --> 00:27:37,440
Speaker 3: that is technically feasible and socially desirable for technology, especially

464
00:27:37,480 --> 00:27:41,800
Speaker 3: for AI. Articulate what this is. Let's have a conversation

465
00:27:41,840 --> 00:27:44,040
Speaker 3: about how we can get there. What we can do

466
00:27:44,119 --> 00:27:47,080
Speaker 3: to encourage researchers, what we can do to encourage engineers,

467
00:27:47,119 --> 00:27:49,119
Speaker 3: what we can do to encourage businesses to actually go

468
00:27:49,200 --> 00:27:51,280
Speaker 3: in that direction. What does the government need to do,

469
00:27:51,280 --> 00:27:53,320
Speaker 3: What does civil society need to do? What does the

470
00:27:53,400 --> 00:27:55,080
Speaker 3: media need? By the way, I think media is a

471
00:27:55,080 --> 00:27:59,520
Speaker 3: big part of the problem. Media often sort of increases

472
00:27:59,600 --> 00:28:04,479
Speaker 3: the appeal of the tech industry. It sort of paints

473
00:28:04,480 --> 00:28:09,600
Speaker 3: a picture of tech leaders as these geniuses who are

474
00:28:10,119 --> 00:28:14,359
Speaker 3: revolutionizing things, and it personalizes that their power, and it

475
00:28:14,440 --> 00:28:17,919
Speaker 3: makes it harder for the public right to keep the

476
00:28:17,960 --> 00:28:22,200
Speaker 3: tech sector accountable. Also, on the AI field, I think

477
00:28:22,240 --> 00:28:24,959
Speaker 3: the media is part of the reason why there is

478
00:28:24,960 --> 00:28:29,119
Speaker 3: so much hype. Many of the leading publications, such as

479
00:28:29,160 --> 00:28:32,639
Speaker 3: the Economists or the New York Times, every week prints

480
00:28:32,680 --> 00:28:35,520
Speaker 3: something about AI will solve this problem or that problem.

481
00:28:35,640 --> 00:28:36,800
Speaker 3: AI is going to revolutionize.

482
00:28:36,840 --> 00:28:38,760
Speaker 2: It's always will solve it, Yes.

483
00:28:38,760 --> 00:28:40,800
Speaker 3: Will solve it. Yet it hasn't solved anything yet yet.

484
00:28:41,200 --> 00:28:44,240
Speaker 2: Yeah, And that's I mean, part of the reason the

485
00:28:44,240 --> 00:28:47,360
Speaker 2: show exists. And I think it comes down to I

486
00:28:47,400 --> 00:28:49,040
Speaker 2: do blame a lot of this on the growth of

487
00:28:49,080 --> 00:28:51,800
Speaker 2: all coast economy, but it's also it's almost like there

488
00:28:51,920 --> 00:28:54,160
Speaker 2: is no long termism anymore in a lot of the

489
00:28:54,200 --> 00:28:57,800
Speaker 2: tech economy, it's all this will happen. Just trust us

490
00:28:58,120 --> 00:29:00,000
Speaker 2: and give us as much time and money as possible.

491
00:29:00,440 --> 00:29:02,000
Speaker 2: But we're not going to invest in R and D.

492
00:29:02,200 --> 00:29:03,680
Speaker 2: It's just a bizarre.

493
00:29:04,000 --> 00:29:06,920
Speaker 3: Well, look, let's also think about the world at large.

494
00:29:07,760 --> 00:29:11,480
Speaker 3: There are six billion people who live outside of Europe, US,

495
00:29:11,560 --> 00:29:15,719
Speaker 3: Canada and China. That includes the weakest, the poorest people

496
00:29:15,800 --> 00:29:19,640
Speaker 3: in the world. How can we improve their lives? Nothing

497
00:29:19,680 --> 00:29:21,760
Speaker 3: we're talking about AI here is going to do that.

498
00:29:21,960 --> 00:29:25,160
Speaker 2: And that's actually the thing. It's connecting back to what

499
00:29:25,200 --> 00:29:28,400
Speaker 2: you were saying earlier. The problems being solved don't feel

500
00:29:28,400 --> 00:29:30,840
Speaker 2: like they're solving for everyone. It's solving what's very much

501
00:29:30,880 --> 00:29:36,080
Speaker 2: in front of us, the latest iPhone, latest computer. What

502
00:29:36,240 --> 00:29:38,560
Speaker 2: problems can that solve? And thus generative AI kind of

503
00:29:38,560 --> 00:29:41,440
Speaker 2: makes sense because it's like, oh, more computer, but more

504
00:29:41,440 --> 00:29:42,720
Speaker 2: computer isn't fixing.

505
00:29:42,440 --> 00:29:45,239
Speaker 3: Anything metaverse as a solution to you know, people who

506
00:29:45,280 --> 00:29:45,880
Speaker 3: are starving.

507
00:29:46,440 --> 00:29:48,720
Speaker 2: Actually, this leads me to a question, what did you

508
00:29:48,760 --> 00:29:50,760
Speaker 2: think of cryptocurrency? I wish I would have had this

509
00:29:50,800 --> 00:29:52,440
Speaker 2: podcast asked you about this at the time.

510
00:29:52,760 --> 00:29:56,160
Speaker 3: Well, I said that I see the positive for jener

511
00:29:56,280 --> 00:29:59,240
Speaker 3: to AI. I think it's actually a promising technology. I

512
00:29:59,320 --> 00:30:02,680
Speaker 3: do not see any positive for cryptocurrency. I never did.

513
00:30:03,920 --> 00:30:08,960
Speaker 3: When I read the manifesto first about bitcoin, it was interesting,

514
00:30:09,000 --> 00:30:12,680
Speaker 3: it was thought provoking, But two days later I was

515
00:30:12,920 --> 00:30:14,040
Speaker 3: inoculated against it.

516
00:30:14,160 --> 00:30:16,920
Speaker 2: Yeah, well you kind of remembered real money exists, and

517
00:30:17,120 --> 00:30:17,480
Speaker 2: at that.

518
00:30:17,400 --> 00:30:21,280
Speaker 3: Point we cannot trust the government. Yes, we cannot trust politicians. Yes,

519
00:30:22,040 --> 00:30:25,440
Speaker 3: but as long as we keep politicians and the government

520
00:30:25,560 --> 00:30:28,640
Speaker 3: under some sort of check with true democratic means, you know,

521
00:30:29,000 --> 00:30:32,680
Speaker 3: the money is not the most important problem. So that's

522
00:30:32,720 --> 00:30:34,880
Speaker 3: not the biggest issue that we have to worry about.

523
00:30:35,000 --> 00:30:38,000
Speaker 2: So a wrap up question, I really appreciate your time,

524
00:30:38,160 --> 00:30:40,840
Speaker 2: of course, Ed. Are you optimistic about the future for

525
00:30:40,880 --> 00:30:41,680
Speaker 2: the tech industry.

526
00:30:42,480 --> 00:30:48,240
Speaker 3: No, I am not a techno optimist, and I'm not

527
00:30:48,480 --> 00:30:52,600
Speaker 3: a market optimist, meaning that if I define optimism as

528
00:30:53,200 --> 00:30:56,040
Speaker 3: things are gonna work out there is an arc of progress.

529
00:30:56,360 --> 00:30:58,960
Speaker 3: I am not an optimist. I think we have serious

530
00:30:59,000 --> 00:31:01,720
Speaker 3: problems with the tech industry. We have serious problems with

531
00:31:01,800 --> 00:31:03,880
Speaker 3: the market process in the United States right now, with

532
00:31:04,000 --> 00:31:08,320
Speaker 3: social processes. But I'm hopeful. I believe that there is

533
00:31:08,360 --> 00:31:11,800
Speaker 3: a direction in which we could use technology that would

534
00:31:11,840 --> 00:31:13,520
Speaker 3: make things better. And there is a way in which

535
00:31:13,520 --> 00:31:18,760
Speaker 3: we can introduce better regulation, better worker organizations, better training

536
00:31:19,400 --> 00:31:22,880
Speaker 3: that would make the market system work better. But that's

537
00:31:22,920 --> 00:31:26,160
Speaker 3: the hope that we could achieve that if we did

538
00:31:26,240 --> 00:31:28,720
Speaker 3: the right things. But I don't think that we are

539
00:31:28,720 --> 00:31:31,280
Speaker 3: heading there. Left to our own devices.

540
00:31:31,080 --> 00:31:33,280
Speaker 2: So where does it head if we're heading in that direction?

541
00:31:33,760 --> 00:31:35,880
Speaker 3: Oh, I prefer not to answer that question.

542
00:31:37,080 --> 00:31:38,960
Speaker 2: That's a perfectly fine way to end it.

543
00:31:39,040 --> 00:31:39,400
Speaker 3: Darn.

544
00:31:39,400 --> 00:31:41,240
Speaker 2: Thank you so much for joining me today.

545
00:31:41,720 --> 00:31:44,240
Speaker 3: Thank you, ed. This was really excellent conversation. I really

546
00:31:44,320 --> 00:31:46,760
Speaker 3: enjoyed it, and I'm sorry that my voice was a

547
00:31:46,800 --> 00:31:47,880
Speaker 3: bit of a downer.

548
00:31:47,920 --> 00:31:50,200
Speaker 2: Oh, don't worry. The listeners have just be glad to

549
00:31:50,200 --> 00:31:53,560
Speaker 2: hear someone else have than me talking. You've been listening

550
00:31:53,600 --> 00:32:04,400
Speaker 2: to Better Offlines. Thank you for listening, everyone, Thank you

551
00:32:04,440 --> 00:32:07,160
Speaker 2: for listening to Better Offline. The editor and composer of

552
00:32:07,160 --> 00:32:10,160
Speaker 2: the Better Offline theme song is Matasowski. You can check

553
00:32:10,160 --> 00:32:13,080
Speaker 2: out more of his music and audio projects at Matasowski

554
00:32:13,160 --> 00:32:16,640
Speaker 2: dot com, M A T T O S O W

555
00:32:17,000 --> 00:32:20,320
Speaker 2: s KI dot com. You can email me at easy

556
00:32:20,360 --> 00:32:23,000
Speaker 2: at better offline dot com, or visit Better Offline dot

557
00:32:23,040 --> 00:32:25,640
Speaker 2: com to find more podcast links and of course my newsletter.

558
00:32:26,040 --> 00:32:28,560
Speaker 2: I also really recommend you go to chat dot Where's

559
00:32:28,560 --> 00:32:31,040
Speaker 2: youreed dot at to visit the discord, and go to

560
00:32:31,080 --> 00:32:34,440
Speaker 2: our slash Better Offline to check out our reddit. Thank

561
00:32:34,520 --> 00:32:35,640
Speaker 2: you so much for listening.

562
00:32:36,480 --> 00:32:39,160
Speaker 1: Better Offline is a production of cool Zone Media. For

563
00:32:39,280 --> 00:32:42,480
Speaker 1: more from cool Zone Media, visit our website cool Zonemedia

564
00:32:42,520 --> 00:32:45,360
Speaker 1: dot com, or check us out on the iHeartRadio app,

565
00:32:45,400 --> 00:32:47,800
Speaker 1: Apple Podcasts, or wherever you get your podcasts.

566
00:33:00,760 --> 00:33:04,480
Speaker 2: School