1
00:00:01,840 --> 00:00:06,720
Speaker 1: All Zone Media. Hail traveler, Welcome to this week's Better

2
00:00:06,760 --> 00:00:17,040
Speaker 1: offline monologue. I'm your host ed ZITRN, so I'm breeding

3
00:00:17,079 --> 00:00:19,120
Speaker 1: this the day after In Video put out there August

4
00:00:19,160 --> 00:00:21,919
Speaker 1: twenty seventh turnings, and I'm not sure anybody really knows

5
00:00:21,920 --> 00:00:25,120
Speaker 1: how to react. While in Video be estimates in general,

6
00:00:25,200 --> 00:00:27,920
Speaker 1: they for the second quarter running mist analyst estimates on

7
00:00:28,000 --> 00:00:30,840
Speaker 1: data center revenue, which is eighty eight percent of their revenue,

8
00:00:30,880 --> 00:00:33,279
Speaker 1: where all the GPUs are coming in at forty one

9
00:00:33,280 --> 00:00:35,840
Speaker 1: point one billion dollars on an estimated forty one point

10
00:00:35,840 --> 00:00:38,599
Speaker 1: three billion dollars. The market also has a very unhealthy

11
00:00:38,640 --> 00:00:41,800
Speaker 1: relationship with this company. In Video grew like fifty percent

12
00:00:41,880 --> 00:00:43,840
Speaker 1: or something year over year, but because they didn't grow

13
00:00:43,960 --> 00:00:46,480
Speaker 1: one hundred and twelve percent or whatever year over year,

14
00:00:46,680 --> 00:00:49,720
Speaker 1: people are the babies and the markets are soiling their diapers,

15
00:00:49,960 --> 00:00:54,320
Speaker 1: And honestly, it's because really it's unhealthy. The unrealistic valuations

16
00:00:54,320 --> 00:00:57,480
Speaker 1: that in Video has been given are finally causing problems.

17
00:00:57,880 --> 00:00:59,800
Speaker 1: And as a reminder, in Video is the company that

18
00:00:59,840 --> 00:01:03,120
Speaker 1: may the specialized GPUs used in effectively all generative AI

19
00:01:03,360 --> 00:01:05,880
Speaker 1: and has seen unprecedented stock growth as a result of

20
00:01:05,920 --> 00:01:10,959
Speaker 1: this hype. Cycle. The magnificent seven stocks Microsoft, Google, Meta, Amazon, Apple, Tesla,

21
00:01:11,000 --> 00:01:13,640
Speaker 1: and of course Nvidia make up thirty five percent of

22
00:01:13,680 --> 00:01:15,800
Speaker 1: the value of the US stock market, and as I

23
00:01:15,880 --> 00:01:18,120
Speaker 1: mentioned on the AI money Trap, they are sent a

24
00:01:18,160 --> 00:01:21,000
Speaker 1: capital expenditures, of which Nvidia makes up a large chunk.

25
00:01:21,280 --> 00:01:24,280
Speaker 1: By well, their GPUs do at least accounted for more

26
00:01:24,360 --> 00:01:29,200
Speaker 1: growth than all consumer spending combined. So on so many levels,

27
00:01:29,200 --> 00:01:32,120
Speaker 1: and Video really is the AI bubble. They're the only

28
00:01:32,160 --> 00:01:35,000
Speaker 1: company making a profit in AI, and forty two percent

29
00:01:35,000 --> 00:01:37,759
Speaker 1: of their revenue is from the magnificent seven companies buying

30
00:01:37,800 --> 00:01:41,120
Speaker 1: their GPUs, and their financial health and continued growth is

31
00:01:41,120 --> 00:01:43,880
Speaker 1: a barometer for the future growth of this nonsense, bullshit

32
00:01:43,959 --> 00:01:46,959
Speaker 1: bubble that I'm tired of talking about kind of, but

33
00:01:47,080 --> 00:01:49,840
Speaker 1: you know, there's always something to chat about yet. Luke

34
00:01:49,920 --> 00:01:53,120
Speaker 1: Carwo of Sherwood News discovered one particularly weird part of

35
00:01:53,240 --> 00:01:56,600
Speaker 1: Nvidia's earnings this week two point two billion dollars of

36
00:01:56,640 --> 00:02:00,280
Speaker 1: their revenue under net other income, which turned out to

37
00:02:00,360 --> 00:02:03,680
Speaker 1: be from the mark to market value of their investment

38
00:02:03,720 --> 00:02:09,399
Speaker 1: in unprofitable AI compute company cor Weve and by the way,

39
00:02:09,480 --> 00:02:12,200
Speaker 1: in Video invested in core Weave has a one point

40
00:02:12,280 --> 00:02:15,920
Speaker 1: three billion dollar GPU compute deal with them called Project Osprey.

41
00:02:16,600 --> 00:02:19,720
Speaker 1: They give them preferential rights their GPUs and core Weave.

42
00:02:19,880 --> 00:02:23,000
Speaker 1: When they raised debt, they raise by using the GPUs

43
00:02:23,040 --> 00:02:26,400
Speaker 1: they bought from a Videos as collateral in the loan,

44
00:02:26,480 --> 00:02:29,240
Speaker 1: and then they use that money from the from the

45
00:02:29,280 --> 00:02:33,720
Speaker 1: loan to buy more GPUs. Every time I think about

46
00:02:33,760 --> 00:02:38,520
Speaker 1: core Weave, I feel a little crazy. And in simple terms,

47
00:02:38,639 --> 00:02:41,240
Speaker 1: in Video book the increased value of their investment in

48
00:02:41,280 --> 00:02:44,440
Speaker 1: core Weave as revenue in a quarter where they barely

49
00:02:44,520 --> 00:02:49,160
Speaker 1: hit estimates. Hmmm. The reason I'm telling you this is

50
00:02:49,200 --> 00:02:50,959
Speaker 1: that I believe, as I've always said, that all of

51
00:02:51,000 --> 00:02:54,480
Speaker 1: this ends poorly in Video isn't doing anything dodgy per

52
00:02:54,520 --> 00:02:57,400
Speaker 1: se by including core Weave's stock. It's just that the

53
00:02:57,480 --> 00:02:59,840
Speaker 1: value of that stock and core Weave is covering up

54
00:02:59,880 --> 00:03:02,840
Speaker 1: in video is slow in growth. If Core weaves value

55
00:03:02,880 --> 00:03:04,920
Speaker 1: craters in the next few months, and Video will have

56
00:03:04,919 --> 00:03:07,880
Speaker 1: to book a massive loss in the same category. And

57
00:03:08,639 --> 00:03:10,080
Speaker 1: I know some of you are going to email me

58
00:03:10,120 --> 00:03:12,880
Speaker 1: and say, hey, where to hear mark to market before

59
00:03:13,320 --> 00:03:16,320
Speaker 1: And the answer is Enron. And I don't think this

60
00:03:16,480 --> 00:03:19,720
Speaker 1: is like Enron, because Enron was doing something somewhat different

61
00:03:19,760 --> 00:03:23,000
Speaker 1: they were valuing contracts in bizarre ways. This is not

62
00:03:23,440 --> 00:03:29,040
Speaker 1: exactly the same thing. Nevertheless, it's extremely weird to say, yeah,

63
00:03:29,080 --> 00:03:31,880
Speaker 1: the value of this stocker got over here went up.

64
00:03:31,960 --> 00:03:35,480
Speaker 1: Isn't that good? Yeah, that's revenue now. It's dangerous. It's

65
00:03:35,600 --> 00:03:39,040
Speaker 1: dangerous stuff. And it means that on some level, In

66
00:03:39,120 --> 00:03:42,560
Speaker 1: Video's financial health and future earnings are somewhat dependent on

67
00:03:42,680 --> 00:03:46,960
Speaker 1: core weave. It's all so good. I like it. It's very good.

68
00:03:47,040 --> 00:03:51,240
Speaker 1: It's normal. Everything's so good. It's great, everybody's doing well.

69
00:03:51,520 --> 00:03:53,760
Speaker 1: And what's even more better is that in Video is

70
00:03:53,800 --> 00:03:57,240
Speaker 1: effectively the only success story in AI, opening eye and

71
00:03:57,280 --> 00:03:59,600
Speaker 1: anthropic but billions of dollars each year and have no

72
00:03:59,680 --> 00:04:02,720
Speaker 1: path any kind of sustainability. And as I've argued again

73
00:04:02,760 --> 00:04:05,600
Speaker 1: and again, these products lack any kind of large scale

74
00:04:05,720 --> 00:04:07,880
Speaker 1: use cases other than the fact that every single media

75
00:04:07,880 --> 00:04:10,880
Speaker 1: outlet has been screaming chat GPT for the last three years,

76
00:04:11,080 --> 00:04:14,960
Speaker 1: and seemingly everyone just sees code generation goes, oh my god,

77
00:04:15,040 --> 00:04:18,920
Speaker 1: this is literally god. But I think after three years

78
00:04:18,920 --> 00:04:21,320
Speaker 1: of breathless hype, the worm is kind of turning on AI.

79
00:04:21,600 --> 00:04:23,800
Speaker 1: In the last week, I've been on national TV twice,

80
00:04:23,880 --> 00:04:26,240
Speaker 1: once on CNN and once on the BBC. And while

81
00:04:26,279 --> 00:04:29,000
Speaker 1: I'm still introduced as somebody that does not believed the hype,

82
00:04:29,120 --> 00:04:31,280
Speaker 1: the fact is that more people are beginning to wonder

83
00:04:31,600 --> 00:04:35,920
Speaker 1: what if Ed's right? So from here on now, I'm

84
00:04:35,920 --> 00:04:38,320
Speaker 1: going to extrapolate a little bit. I don't know the future.

85
00:04:38,360 --> 00:04:40,320
Speaker 1: I'm not saying, oh, this will definitely happen, but here

86
00:04:40,320 --> 00:04:43,840
Speaker 1: are some things that on my mind if I'm right. Well,

87
00:04:44,600 --> 00:04:47,359
Speaker 1: let's start with something simple. Open ai is toast. The

88
00:04:47,440 --> 00:04:50,760
Speaker 1: Financial Times reported that open AI's corporate restructuring is and

89
00:04:50,800 --> 00:04:53,680
Speaker 1: I quote likely to slip into next year as Microsoft

90
00:04:53,760 --> 00:04:56,080
Speaker 1: refuses to budge on the terms that of open ai

91
00:04:56,200 --> 00:04:59,480
Speaker 1: converting from a nonprofit to a for profit. Now. This

92
00:04:59,520 --> 00:05:03,200
Speaker 1: is because open ai went to Microsoft and said, okay, guys,

93
00:05:03,480 --> 00:05:05,839
Speaker 1: we need to convert to a for profit or we

94
00:05:05,880 --> 00:05:09,320
Speaker 1: will die. Microsoft said, okay, great, let's go ahead and

95
00:05:09,360 --> 00:05:11,839
Speaker 1: do that. We don't mind. Open Ai said, well, what

96
00:05:11,880 --> 00:05:13,160
Speaker 1: we're going to need is, we're going to need you

97
00:05:13,200 --> 00:05:15,599
Speaker 1: to end your exclusivity deal over our models. We're going

98
00:05:15,640 --> 00:05:18,400
Speaker 1: to need to cut your revenue share, We're going to

99
00:05:18,480 --> 00:05:21,560
Speaker 1: need to not share our IP and our research from

100
00:05:21,800 --> 00:05:24,359
Speaker 1: with you. What do you think? And Microsoft said, go

101
00:05:24,440 --> 00:05:27,960
Speaker 1: fuck yourself. Microsoft back in I think May said that

102
00:05:27,960 --> 00:05:29,599
Speaker 1: they were willing to walk away from the table. I

103
00:05:29,640 --> 00:05:32,480
Speaker 1: don't fucking blame them. Now people will say, well, Microsoft

104
00:05:32,560 --> 00:05:35,080
Speaker 1: wants open ai to go public, they'll get a bunch

105
00:05:35,160 --> 00:05:37,440
Speaker 1: of money. Kind of luck. Cool, we've different video, right,

106
00:05:37,960 --> 00:05:40,920
Speaker 1: Except Microsoft already owns all their IP and already runs

107
00:05:40,920 --> 00:05:43,760
Speaker 1: all their infrastructure. Remember open ai doesn't own any of

108
00:05:43,800 --> 00:05:46,440
Speaker 1: their info at all. They're trying to build that bullshit

109
00:05:46,480 --> 00:05:50,040
Speaker 1: out in Abilene, Texas, of course, but I mean that's

110
00:05:50,080 --> 00:05:53,039
Speaker 1: not going to be built this year or even next

111
00:05:53,040 --> 00:05:56,760
Speaker 1: at this point. So you know, Microsoft is not understandably

112
00:05:56,839 --> 00:05:59,320
Speaker 1: jazz to take this deal and could let open ai

113
00:05:59,480 --> 00:06:01,839
Speaker 1: die on the vie. And if open ai fails to

114
00:06:01,839 --> 00:06:03,880
Speaker 1: convert by the end of the year, soft Bank cups

115
00:06:03,880 --> 00:06:06,880
Speaker 1: they're funding round in half, meaning that by my calculations,

116
00:06:07,120 --> 00:06:09,560
Speaker 1: soft Bank's part of the remaining thirty billion dollars, which

117
00:06:09,560 --> 00:06:12,120
Speaker 1: would at this point become ten billion dollars, would drop

118
00:06:12,160 --> 00:06:15,080
Speaker 1: two seven hundred billion dollars thanks to open ai already

119
00:06:15,160 --> 00:06:19,400
Speaker 1: raising eight point three billion dollars from other investors. Some

120
00:06:19,440 --> 00:06:22,600
Speaker 1: people are saying it's ten billion dollars just from soft Bank.

121
00:06:22,760 --> 00:06:24,919
Speaker 1: I don't think people are good at math. I swear

122
00:06:24,920 --> 00:06:26,920
Speaker 1: to God, I've been looking at this ship for years,

123
00:06:26,960 --> 00:06:31,320
Speaker 1: and nevertheless things don't look good there. And again, soft

124
00:06:31,320 --> 00:06:33,599
Speaker 1: Bank has been very clear that if by the end

125
00:06:33,640 --> 00:06:35,480
Speaker 1: of the year open ai is not converted, they're cutting

126
00:06:35,520 --> 00:06:38,120
Speaker 1: that round in half. Even if open Ai and Microsoft

127
00:06:38,160 --> 00:06:40,400
Speaker 1: work this out today, I don't think it's possible that

128
00:06:40,440 --> 00:06:42,000
Speaker 1: they convert by the end of the year, if it's

129
00:06:42,040 --> 00:06:46,240
Speaker 1: even possible at all. Nonprofits are very specific things, and

130
00:06:46,240 --> 00:06:48,200
Speaker 1: no one has ever done what open ai is trying

131
00:06:48,240 --> 00:06:51,159
Speaker 1: to do. And if open ai cannot convert, they aren't

132
00:06:51,200 --> 00:06:54,080
Speaker 1: dead because they can never IPO and they're too big

133
00:06:54,120 --> 00:06:56,200
Speaker 1: to sell to anyone else. And thus investors would have

134
00:06:56,200 --> 00:06:59,000
Speaker 1: little reason to put further capital in other than believing

135
00:06:59,040 --> 00:07:01,920
Speaker 1: that open ai will so somehow stop burning billions of dollars,

136
00:07:01,960 --> 00:07:04,400
Speaker 1: something they have never proven is possible and nobody has

137
00:07:04,440 --> 00:07:07,599
Speaker 1: in fact, and fucking I hear the cost of inference

138
00:07:07,680 --> 00:07:10,000
Speaker 1: is going down again, and going to screen, it's going up.

139
00:07:10,040 --> 00:07:12,560
Speaker 1: I'm getting to it later and in a future episode.

140
00:07:12,600 --> 00:07:15,160
Speaker 1: And if the conversion fails, open Ai can probably cobble

141
00:07:15,200 --> 00:07:17,880
Speaker 1: together another fifteen billion dollars of funding before they shit

142
00:07:17,920 --> 00:07:20,760
Speaker 1: their pants and die. And as an aside, Open Ai

143
00:07:20,880 --> 00:07:23,080
Speaker 1: is currently doing an insider share sale at a five

144
00:07:23,160 --> 00:07:26,840
Speaker 1: hundred billion dollar valuation, with current and former employees selling

145
00:07:27,040 --> 00:07:30,400
Speaker 1: eight billion dollars worth of stock. This is not something

146
00:07:30,480 --> 00:07:33,040
Speaker 1: you do if you think you're going to go public

147
00:07:33,120 --> 00:07:37,200
Speaker 1: anytime soon. If ever, I need to be clear about this,

148
00:07:37,760 --> 00:07:41,680
Speaker 1: it looks bad and outside of Nvidia, AI is failing

149
00:07:41,720 --> 00:07:44,280
Speaker 1: to produce any kind of profit. And once venture capital

150
00:07:44,360 --> 00:07:48,720
Speaker 1: realizes this, they'll stop investing in these giant, loss ridden monstrosities.

151
00:07:48,960 --> 00:07:51,760
Speaker 1: This will naturally bring things to a headed Anthropic, another

152
00:07:51,800 --> 00:07:54,560
Speaker 1: company that burns billions of dollars, who will inevitably raise

153
00:07:54,600 --> 00:07:58,640
Speaker 1: maybe another ten to fifteen themselves, and perhaps another roundup after

154
00:07:58,720 --> 00:08:03,040
Speaker 1: that with nouffittability, though there is no reason, really no

155
00:08:03,240 --> 00:08:05,320
Speaker 1: reason to keep throwing cash in the furnace. And I

156
00:08:05,360 --> 00:08:08,160
Speaker 1: believe that Anthropic eventually gets absorbed into Amazon, like at

157
00:08:08,160 --> 00:08:11,520
Speaker 1: the end of that horrible game inside, and maybe it

158
00:08:11,600 --> 00:08:14,520
Speaker 1: will also flop around like a big blob in a field.

159
00:08:14,520 --> 00:08:17,120
Speaker 1: And if I spoiled inside of you, it's an old game,

160
00:08:17,160 --> 00:08:19,120
Speaker 1: you should have beaten it by now. Then when the

161
00:08:19,120 --> 00:08:22,200
Speaker 1: public markets, eventually the AI trade will fail. Despite what

162
00:08:22,240 --> 00:08:25,040
Speaker 1: the headlines may say, outside of Nvidia, the Magnificent seven

163
00:08:25,080 --> 00:08:27,520
Speaker 1: has not been making money on AI and has instead

164
00:08:27,600 --> 00:08:30,400
Speaker 1: been growing there already existing businesses, which is something that

165
00:08:30,440 --> 00:08:34,400
Speaker 1: reporters should know better. Yet they're claiming that any kind

166
00:08:34,440 --> 00:08:37,040
Speaker 1: of growth proves that AI bets are paying off. I'm

167
00:08:37,040 --> 00:08:39,640
Speaker 1: not singling you out yet, but you've gotta fucking stop.

168
00:08:40,200 --> 00:08:42,880
Speaker 1: Once other revenue sources stop papering over the flimsy or

169
00:08:42,960 --> 00:08:46,280
Speaker 1: unexistent revenue and I mean revenue, not profit from AI,

170
00:08:46,400 --> 00:08:49,000
Speaker 1: the street will realize and they will panic. But let's

171
00:08:49,000 --> 00:08:52,200
Speaker 1: get serious. If I'm right, we're all bearing witness to

172
00:08:52,240 --> 00:08:54,240
Speaker 1: one of the most catastrophic failures in the history of

173
00:08:54,280 --> 00:08:58,959
Speaker 1: the markets, journalism, and society itself. Generative AI was sold

174
00:08:59,000 --> 00:09:01,840
Speaker 1: as a myth. It's been blatantly obvious from the very

175
00:09:01,840 --> 00:09:04,400
Speaker 1: beginning that large language models do not have mass market

176
00:09:04,520 --> 00:09:08,720
Speaker 1: use cases and no coding llms do not goddamn count

177
00:09:09,160 --> 00:09:12,000
Speaker 1: they don't. They don't produce the kind of revenues. And

178
00:09:12,080 --> 00:09:14,960
Speaker 1: guess what. The coding lms are so expensive to run

179
00:09:15,040 --> 00:09:17,160
Speaker 1: that if you took away the subsidies, they would be

180
00:09:17,240 --> 00:09:20,720
Speaker 1: economically impossible. Stop arguing with me on this point. I

181
00:09:20,760 --> 00:09:24,280
Speaker 1: will crush you. Yet many many, many, many, many journalists

182
00:09:24,320 --> 00:09:26,720
Speaker 1: choose to blindly repeat the open AI and other companies

183
00:09:26,720 --> 00:09:29,960
Speaker 1: are building powerful AI that will replace human workers, choosing

184
00:09:30,000 --> 00:09:32,960
Speaker 1: to deliberately ignore the pisspoor outcomes from using these products

185
00:09:33,080 --> 00:09:36,559
Speaker 1: and I don't know, failing to define what powerful means

186
00:09:36,559 --> 00:09:52,920
Speaker 1: because powerful does not seem to mean goddamn anything. Journalists

187
00:09:52,960 --> 00:09:55,320
Speaker 1: had an opportunity to tell the truth, to look at

188
00:09:55,360 --> 00:09:57,320
Speaker 1: the products and say, hey, I don't really get why

189
00:09:57,320 --> 00:10:01,160
Speaker 1: this is so amazing or powerful, exactly, question the valuations

190
00:10:01,200 --> 00:10:03,560
Speaker 1: of companies that make little or no revenue, while burning

191
00:10:03,679 --> 00:10:06,760
Speaker 1: hundreds of millions or billions of dollars to interrogate the

192
00:10:06,760 --> 00:10:11,120
Speaker 1: outright lies of Sam Moultman and WARRYO Darryo ama Day. Instead,

193
00:10:11,160 --> 00:10:14,360
Speaker 1: they choose to fuel the fantastical fire, repeating the promises

194
00:10:14,360 --> 00:10:16,320
Speaker 1: of people like Sam Moultman in the hopes they'll be

195
00:10:16,320 --> 00:10:18,920
Speaker 1: the first to be able to parrot his next lie

196
00:10:19,200 --> 00:10:22,360
Speaker 1: or talk to his clammiass about nothing for three hours.

197
00:10:23,280 --> 00:10:25,880
Speaker 1: I worry that things are so much worse than we know.

198
00:10:26,480 --> 00:10:29,160
Speaker 1: I have an inkling that the price of serving GPU compute,

199
00:10:29,160 --> 00:10:32,959
Speaker 1: the underlying infrastructure of all generative AI, is horribly unprofitable.

200
00:10:33,040 --> 00:10:34,560
Speaker 1: And if you're a person who's going to say, oh,

201
00:10:34,600 --> 00:10:37,080
Speaker 1: it's only a few dollars an hour. Hey, what if

202
00:10:37,080 --> 00:10:39,720
Speaker 1: they're selling that few dollars an hour at a loss.

203
00:10:40,720 --> 00:10:43,720
Speaker 1: We know now that the cost of inference is increasing,

204
00:10:43,960 --> 00:10:45,920
Speaker 1: and that's coming up in future episodes, but it was

205
00:10:45,960 --> 00:10:48,320
Speaker 1: in the GPT five one, and thus the costs of

206
00:10:48,400 --> 00:10:53,079
Speaker 1: providing AI services, which were already unprofitable, is increasing too. Yet,

207
00:10:53,120 --> 00:10:56,440
Speaker 1: by constantly accepting a myth that AI would get more powerful,

208
00:10:56,480 --> 00:10:58,680
Speaker 1: that these companies would somehow work out how to stop

209
00:10:58,720 --> 00:11:00,839
Speaker 1: burning way more money than they may, the big tech

210
00:11:00,880 --> 00:11:04,360
Speaker 1: would not simply lie about what was possible. Journalism pedaled

211
00:11:04,360 --> 00:11:06,839
Speaker 1: outright lies and help sustain a hype cycle based on

212
00:11:06,920 --> 00:11:09,559
Speaker 1: little more than believing the last smart sounding guy you

213
00:11:09,640 --> 00:11:12,480
Speaker 1: ran into, or what that guy told someone you kind

214
00:11:12,480 --> 00:11:16,479
Speaker 1: of respect. You can dress this up by being an optimist.

215
00:11:16,720 --> 00:11:19,760
Speaker 1: You can dress this up by pretending that you thought

216
00:11:19,800 --> 00:11:23,280
Speaker 1: they would extrapolate from here, that these companies would just

217
00:11:23,440 --> 00:11:25,960
Speaker 1: work it out, that they were just like Uber, that

218
00:11:26,000 --> 00:11:29,000
Speaker 1: they were just like Amazon Web services, that they would

219
00:11:29,040 --> 00:11:32,720
Speaker 1: innovate and contribute you so much the economy based on

220
00:11:32,760 --> 00:11:34,760
Speaker 1: the things that they told you that they wanted you

221
00:11:34,800 --> 00:11:37,440
Speaker 1: to repeat, so their investors would give them more money.

222
00:11:38,080 --> 00:11:40,559
Speaker 1: You can pretend that all of this made sense at

223
00:11:40,600 --> 00:11:44,280
Speaker 1: any goddamn time, but you know you either half asked

224
00:11:44,280 --> 00:11:47,960
Speaker 1: it or you're lying to yourself at this point. And

225
00:11:48,040 --> 00:11:51,160
Speaker 1: I understand also that there are some people writing about

226
00:11:51,160 --> 00:11:54,280
Speaker 1: this who maybe don't want to write such fantastical things.

227
00:11:54,520 --> 00:11:57,559
Speaker 1: They feel pressured or they're scared, And I actually understand

228
00:11:57,600 --> 00:12:00,000
Speaker 1: the fear. Then, when everyone around you is pissed down

229
00:12:00,200 --> 00:12:02,960
Speaker 1: pants and they're saying, why haven't you pissed your pants?

230
00:12:03,040 --> 00:12:06,080
Speaker 1: Perhaps you wouldn't want to do it. But if everyone's

231
00:12:06,080 --> 00:12:08,319
Speaker 1: screaming at you, what are you meant to do? If

232
00:12:08,320 --> 00:12:11,120
Speaker 1: your editors are saying write a positive AI story, if

233
00:12:11,160 --> 00:12:14,120
Speaker 1: your editors are saying that thirteen percent of people in

234
00:12:14,160 --> 00:12:16,920
Speaker 1: the Stanford there was this. Sorry, I'm actually going to

235
00:12:17,000 --> 00:12:19,520
Speaker 1: reframe this point and not retake it, because this is important.

236
00:12:19,920 --> 00:12:21,839
Speaker 1: There was a study that went out that said that

237
00:12:22,240 --> 00:12:25,560
Speaker 1: thirteen percent I think there's a thirteen percent drop in

238
00:12:25,679 --> 00:12:28,800
Speaker 1: jobs that were affected by AI, and that this was

239
00:12:28,920 --> 00:12:31,360
Speaker 1: proof that AI was taking these jobs. When you look

240
00:12:31,360 --> 00:12:34,679
Speaker 1: at this goddamn study, you can see that it doesn't

241
00:12:34,720 --> 00:12:37,360
Speaker 1: actually prove that AI was involved and says that yeah,

242
00:12:37,440 --> 00:12:42,720
Speaker 1: thirteen percent of come of career paths and jobs that

243
00:12:42,880 --> 00:12:46,480
Speaker 1: were somehow being replaced by AI were dropping, and as

244
00:12:46,480 --> 00:12:49,440
Speaker 1: the result, AI was involved. It seems like a fumbled point,

245
00:12:49,520 --> 00:12:52,439
Speaker 1: but what it actually just said was that there was

246
00:12:52,480 --> 00:12:54,840
Speaker 1: a thirteen percent drop in certain employment and we'd like

247
00:12:54,880 --> 00:12:58,040
Speaker 1: AI to be involved. So it's on some level, I

248
00:12:58,080 --> 00:13:00,800
Speaker 1: realize i'm ranting. I realize I'm going a little bit offscript.

249
00:13:01,200 --> 00:13:06,240
Speaker 1: I'm just fucking tired of this. I'm tired because even now,

250
00:13:07,040 --> 00:13:10,319
Speaker 1: even today, even as you look at these companies burning

251
00:13:10,360 --> 00:13:12,560
Speaker 1: all this money, even as you look at the fact

252
00:13:12,559 --> 00:13:15,080
Speaker 1: that there are no real returns here, and when you

253
00:13:15,120 --> 00:13:17,280
Speaker 1: look at the products, you can see they don't really

254
00:13:17,320 --> 00:13:21,640
Speaker 1: do anything super futuristic. Even now, there are people who

255
00:13:21,720 --> 00:13:25,320
Speaker 1: still talk to me and repeatedly say shit like, well,

256
00:13:25,320 --> 00:13:28,599
Speaker 1: what if you're wrong, What if you're wrong about this? Well? Ed?

257
00:13:28,880 --> 00:13:31,280
Speaker 1: What if you turn around? And then the computer wakes

258
00:13:31,360 --> 00:13:33,440
Speaker 1: up and says, I love you, Ed, I love you.

259
00:13:34,559 --> 00:13:38,000
Speaker 1: Nothing happens because that won't happen, And we're living a lie.

260
00:13:38,320 --> 00:13:41,600
Speaker 1: And we built our economy on a lie. And it's scary.

261
00:13:41,920 --> 00:13:44,880
Speaker 1: It's scary because the economy is built on data center

262
00:13:44,920 --> 00:13:48,240
Speaker 1: growth and It's scary because so much of journalism just

263
00:13:48,520 --> 00:13:53,560
Speaker 1: missed this. They just missed it. And I myself have

264
00:13:53,920 --> 00:13:55,960
Speaker 1: done very well as a result of this, because I've

265
00:13:55,960 --> 00:13:59,520
Speaker 1: been elucidating hundreds of thousands words at a time, hours

266
00:13:59,520 --> 00:14:02,520
Speaker 1: a podcast a time. So I'm doing fine. I guess.

267
00:14:02,640 --> 00:14:05,400
Speaker 1: I don't have any stock positions in this crap. I'm

268
00:14:05,440 --> 00:14:09,199
Speaker 1: not investing. I'm not fiscally profiting off of this, but

269
00:14:09,280 --> 00:14:12,920
Speaker 1: my career is doing well, I guess. But on some level,

270
00:14:12,920 --> 00:14:15,439
Speaker 1: we've seen that at the highest levels of journalism there

271
00:14:15,480 --> 00:14:19,280
Speaker 1: are people who are willing to peddle bullshit. And as

272
00:14:19,320 --> 00:14:23,720
Speaker 1: this falls apart, as this collapses, there must be a

273
00:14:23,760 --> 00:14:27,240
Speaker 1: point at which people are held accountable for lying because

274
00:14:27,280 --> 00:14:30,120
Speaker 1: it is lying. And I believe the majority of journalists

275
00:14:30,120 --> 00:14:32,120
Speaker 1: are not trying to lie. I don't think that they

276
00:14:32,200 --> 00:14:34,760
Speaker 1: are deliberately doing this. I think they are weasel wording

277
00:14:34,800 --> 00:14:37,200
Speaker 1: around stuff so that they don't stand out from their

278
00:14:37,240 --> 00:14:40,080
Speaker 1: peers and get in trouble and face peer pressure for

279
00:14:40,200 --> 00:14:45,080
Speaker 1: not falling behind the AI bubble cycle. Except everyone's gonna

280
00:14:45,120 --> 00:14:49,120
Speaker 1: look fucking stupid if this collapses. If I'm right, everyone's

281
00:14:49,160 --> 00:14:52,200
Speaker 1: gonna look really silly, and you can feel however you

282
00:14:52,360 --> 00:14:55,680
Speaker 1: feel about me. This isn't about anyone specific other than

283
00:14:56,200 --> 00:15:00,600
Speaker 1: the very obvious hard forky folks. But if you're a

284
00:15:00,640 --> 00:15:04,640
Speaker 1: journous listening, this probably isn't abalue. What it is about, however,

285
00:15:05,200 --> 00:15:08,200
Speaker 1: is the truth, and it is about the fact that

286
00:15:08,960 --> 00:15:13,480
Speaker 1: everyone can say something that's wrong. And we as people

287
00:15:13,560 --> 00:15:17,400
Speaker 1: that report the truth, ostensibly must look at this when

288
00:15:17,440 --> 00:15:20,640
Speaker 1: it collapses, and I really do think it will, and

289
00:15:20,720 --> 00:15:23,520
Speaker 1: we must take an account of what has been said

290
00:15:23,680 --> 00:15:27,040
Speaker 1: and how it has been said, of how many affordances

291
00:15:27,080 --> 00:15:29,840
Speaker 1: were given to people like Sam Altman and Wario Ama Day,

292
00:15:30,360 --> 00:15:33,720
Speaker 1: of how many people said, wow, AI is going to

293
00:15:33,760 --> 00:15:37,360
Speaker 1: get more powerful. We must take an account of why

294
00:15:37,400 --> 00:15:39,600
Speaker 1: it was said it would get more powerful, and if

295
00:15:39,600 --> 00:15:42,200
Speaker 1: there was no reason to other than I think it

296
00:15:42,240 --> 00:15:45,440
Speaker 1: is number is going up that these people keep telling

297
00:15:45,480 --> 00:15:47,840
Speaker 1: me they have met smart people in the valley with

298
00:15:47,920 --> 00:15:50,720
Speaker 1: tears in their eyes, big strong men that told them

299
00:15:50,880 --> 00:15:55,040
Speaker 1: that AGI was coming. If that is the flimsy, fucking

300
00:15:55,160 --> 00:15:58,560
Speaker 1: reasoning we have used to send our economy into the toilet,

301
00:15:58,600 --> 00:16:01,800
Speaker 1: which is what I fear, we should all be goddamn

302
00:16:01,840 --> 00:16:05,760
Speaker 1: ashamed of ourselves. I'm not. I'm not ashamed because I've

303
00:16:05,760 --> 00:16:07,960
Speaker 1: admitted what I am from the very goddamn beginning, which

304
00:16:08,000 --> 00:16:10,320
Speaker 1: is just a guy with a microphone and a fucking blog,

305
00:16:10,560 --> 00:16:12,760
Speaker 1: and I'm doing my best out here to tell the truth.

306
00:16:12,840 --> 00:16:16,120
Speaker 1: If I am wrong, if I am truly wrong, if

307
00:16:16,120 --> 00:16:19,000
Speaker 1: I got all of this wrong, and AGI comes and

308
00:16:19,200 --> 00:16:22,360
Speaker 1: the computer eats me whole and shits me out like Kirby,

309
00:16:22,560 --> 00:16:24,280
Speaker 1: then I will eat crow and I will tell you

310
00:16:24,360 --> 00:16:28,200
Speaker 1: why I'm wrong in detail. Do you think other people

311
00:16:28,240 --> 00:16:30,880
Speaker 1: will do the same when the bubble bursts? Do you

312
00:16:31,080 --> 00:16:34,840
Speaker 1: think that these people will hold themselves accountable or do

313
00:16:34,880 --> 00:16:37,240
Speaker 1: you think they will sweep this under the fucking rug.

314
00:16:38,400 --> 00:16:42,160
Speaker 1: I think they will, and I don't believe listeners should

315
00:16:42,240 --> 00:16:44,480
Speaker 1: let them, and I will not be letting them either.

316
00:16:45,520 --> 00:16:48,840
Speaker 1: This isn't going to be an industry takedown, nor will

317
00:16:48,880 --> 00:16:52,680
Speaker 1: one be coming in the future, but I think journalism

318
00:16:52,760 --> 00:16:56,880
Speaker 1: needs to do a great deal better. And my frustration

319
00:16:57,240 --> 00:17:00,040
Speaker 1: is that in the end of all of this, the

320
00:17:00,080 --> 00:17:02,800
Speaker 1: people that will get hurt will be regular people. The

321
00:17:02,960 --> 00:17:08,600
Speaker 1: environmental destruction cannot be unwritten, the theft cannot be undone.

322
00:17:08,840 --> 00:17:11,880
Speaker 1: No class action suit can make up for the art directors,

323
00:17:11,920 --> 00:17:15,840
Speaker 1: the translators, the transcribers who are losing their jobs to

324
00:17:15,880 --> 00:17:18,800
Speaker 1: this right now, and who will even if these things

325
00:17:18,880 --> 00:17:23,200
Speaker 1: become too expensive to actually run at scale from here

326
00:17:23,240 --> 00:17:26,520
Speaker 1: on it, which is very possible. At some point, the

327
00:17:26,560 --> 00:17:29,480
Speaker 1: price pressure on these careers as they come back will

328
00:17:29,520 --> 00:17:31,760
Speaker 1: match the price of the AI, not the price of

329
00:17:31,800 --> 00:17:35,320
Speaker 1: the jobs that they had before. The labor destruction has

330
00:17:35,359 --> 00:17:38,080
Speaker 1: already happened on a much smaller level than they want

331
00:17:38,080 --> 00:17:41,800
Speaker 1: you to believe it, but it has happened. But fundamentally,

332
00:17:41,840 --> 00:17:44,719
Speaker 1: the AI bubble has damaged the concept of the truth

333
00:17:45,920 --> 00:17:48,560
Speaker 1: and has damaged the concept of what reporting is meant

334
00:17:48,680 --> 00:17:52,120
Speaker 1: to do, because an alarming amount of reporting has been

335
00:17:52,240 --> 00:17:55,399
Speaker 1: there to prove the AI bubble was real, rather than

336
00:17:55,520 --> 00:17:59,639
Speaker 1: prove what was actually real, to actually hold the accountable powerful,

337
00:17:59,800 --> 00:18:02,600
Speaker 1: to actually make sure that the truth was being said,

338
00:18:02,680 --> 00:18:05,240
Speaker 1: even if the truth was ikey, even if the truth

339
00:18:05,400 --> 00:18:08,239
Speaker 1: was wow. Tens of billions of dollars each quarter are

340
00:18:08,240 --> 00:18:11,800
Speaker 1: going into GPUs for products that only lose money, for

341
00:18:11,960 --> 00:18:15,480
Speaker 1: products that don't really resemble actual labor, and thus can

342
00:18:15,520 --> 00:18:17,720
Speaker 1: never do any of the things that people are promising.

343
00:18:19,160 --> 00:18:23,600
Speaker 1: And I like to believe that most people here mean well,

344
00:18:23,800 --> 00:18:26,000
Speaker 1: that they've just been going on with the market consensus

345
00:18:26,040 --> 00:18:29,040
Speaker 1: and I can understand that when everyone's honking in the

346
00:18:29,080 --> 00:18:32,240
Speaker 1: same way, do you want to oink? Do you want

347
00:18:32,280 --> 00:18:35,119
Speaker 1: to stand out? At a time when journalists have a

348
00:18:35,200 --> 00:18:40,040
Speaker 1: tough time keeping their jobs as it is, I blame

349
00:18:40,119 --> 00:18:43,080
Speaker 1: the people on the top. I blame the people with

350
00:18:43,160 --> 00:18:48,760
Speaker 1: the largest megaphones. I blame the people who are constantly

351
00:18:48,800 --> 00:18:51,600
Speaker 1: talking about this as the future and the people that

352
00:18:51,640 --> 00:18:56,600
Speaker 1: have derision for those who wouldare critique the powerful. And

353
00:18:56,680 --> 00:19:01,840
Speaker 1: if I'm right, really, if I'm right, there must be

354
00:19:02,480 --> 00:19:05,240
Speaker 1: some degree of accountability within journalism. There has to be

355
00:19:05,480 --> 00:19:09,280
Speaker 1: an apology to readers and listeners and viewers. There won't

356
00:19:09,320 --> 00:19:13,560
Speaker 1: be there, never will be there, never is, but something

357
00:19:13,600 --> 00:19:17,240
Speaker 1: has to change once the bubble bursts. And if i'm again,

358
00:19:17,720 --> 00:19:20,920
Speaker 1: if I'm wrong, I'll explain why I don't do this

359
00:19:21,000 --> 00:19:22,879
Speaker 1: to be right, I do it because I enjoy it,

360
00:19:23,240 --> 00:19:25,159
Speaker 1: and I'm very grateful you've give me your time. This

361
00:19:25,200 --> 00:19:26,760
Speaker 1: is meant to be a regular five to ten minute

362
00:19:26,760 --> 00:19:30,040
Speaker 1: monologue and it's going to go over twenty minutes. But

363
00:19:30,080 --> 00:19:34,360
Speaker 1: I'm frustrated, as you can tell, I'm furious. I'm furious

364
00:19:34,400 --> 00:19:37,160
Speaker 1: because the markets will take a nasty shit after this,

365
00:19:38,000 --> 00:19:41,199
Speaker 1: and once the markets realize that big tech doesn't have

366
00:19:41,280 --> 00:19:43,719
Speaker 1: any other ideas, and the one idea they had that

367
00:19:43,720 --> 00:19:48,159
Speaker 1: they've been wanking off about for years was nothing. Was

368
00:19:48,200 --> 00:19:51,320
Speaker 1: a myth that burned millions or billions of dollars. The

369
00:19:51,400 --> 00:19:54,120
Speaker 1: markets are not going to trust the tech industry anymore.

370
00:19:54,600 --> 00:19:57,480
Speaker 1: They're not going to trust them as the ultimate growth vehicles.

371
00:19:58,200 --> 00:20:01,000
Speaker 1: What do you think that does the startup funding? What

372
00:20:01,119 --> 00:20:05,400
Speaker 1: do you think that does two jobs in tech? Nothing good?

373
00:20:06,640 --> 00:20:10,639
Speaker 1: Nothing good at all. I want a better world, and

374
00:20:10,680 --> 00:20:13,480
Speaker 1: I think you do too. I'm very grateful for your time.

375
00:20:13,840 --> 00:20:17,040
Speaker 1: It's been a little different. I realized this week. Next

376
00:20:17,080 --> 00:20:19,560
Speaker 1: week we'll have Steve from Gamers Nextus, who will be

377
00:20:19,600 --> 00:20:22,960
Speaker 1: talking about the GPU black market in China and of

378
00:20:23,000 --> 00:20:25,760
Speaker 1: course the things that Bloomberg has done to pull down

379
00:20:25,800 --> 00:20:28,960
Speaker 1: that video. And then after that, I'm gonna have a

380
00:20:29,080 --> 00:20:32,680
Speaker 1: three part episode about how to argue with AI boosters

381
00:20:32,760 --> 00:20:34,920
Speaker 1: because it's about time we take the walter to these

382
00:20:34,920 --> 00:20:37,840
Speaker 1: fucking people. Thank you for listening.