1
00:00:02,400 --> 00:00:03,560
Speaker 1: Au Zone Media.

2
00:00:06,400 --> 00:00:09,280
Speaker 2: Hello and welcome to Better Offline. I'm your host ed zitron.

3
00:00:21,160 --> 00:00:24,200
Speaker 2: Last part I really laid down how bad things are

4
00:00:24,239 --> 00:00:27,200
Speaker 2: for open Ai, how bad a company they are, how

5
00:00:27,200 --> 00:00:30,560
Speaker 2: they spent nine billion dollars to lose five billion dollars,

6
00:00:30,920 --> 00:00:34,040
Speaker 2: how bad things look. But open Ai they have users.

7
00:00:34,200 --> 00:00:36,640
Speaker 2: I mean the users lose the money. Every single user

8
00:00:36,680 --> 00:00:39,040
Speaker 2: paying or not loses the money. They can only survive

9
00:00:39,120 --> 00:00:41,760
Speaker 2: if they're given more money. They will literally die if

10
00:00:41,760 --> 00:00:44,640
Speaker 2: not given more venture capital money. And that really bothers me,

11
00:00:45,200 --> 00:00:47,000
Speaker 2: And that bothers me a great deal. Listen to the

12
00:00:47,040 --> 00:00:49,879
Speaker 2: episode again if you want to really understand why. But

13
00:00:49,960 --> 00:00:52,560
Speaker 2: let's start this one with a very very simple question.

14
00:00:53,440 --> 00:00:58,520
Speaker 2: Is generative AI a real industry? Look, the large language

15
00:00:58,520 --> 00:01:02,040
Speaker 2: model paradigm is yet to produce a successful mass market product,

16
00:01:02,080 --> 00:01:05,039
Speaker 2: and no, large language models are not a success, nor

17
00:01:05,080 --> 00:01:07,120
Speaker 2: are they mass market. I know you're gonna say chat

18
00:01:07,200 --> 00:01:09,760
Speaker 2: GPT is huge. We've already been through that. I just

19
00:01:09,840 --> 00:01:14,240
Speaker 2: talked about that. But surely, surely, right like, it wouldn't

20
00:01:14,280 --> 00:01:17,959
Speaker 2: just be open Ai, right like, we've been through this.

21
00:01:18,000 --> 00:01:20,640
Speaker 2: But if generative AI was a real industry, there'd be

22
00:01:20,760 --> 00:01:24,600
Speaker 2: multiple other players with massive customer bases as a result

23
00:01:24,600 --> 00:01:28,039
Speaker 2: of how revolutionary it was, right, like at least close, right,

24
00:01:28,640 --> 00:01:31,039
Speaker 2: maybe like half the size, a quarter of the size,

25
00:01:31,280 --> 00:01:36,399
Speaker 2: right right, wrong, wrong, so wrong, so very wrong, so

26
00:01:36,560 --> 00:01:40,240
Speaker 2: fucking wrong. It boils my blood. It makes me scared

27
00:01:40,319 --> 00:01:43,520
Speaker 2: for the future, for the market for Silicon Valley. The

28
00:01:43,600 --> 00:01:47,400
Speaker 2: venture capitalists are so fucking wrong, and so are the hyperscalers. Wrong, wrong, wrong.

29
00:01:48,080 --> 00:01:50,640
Speaker 2: Let's look at some estimated numbers that I got from

30
00:01:50,720 --> 00:01:53,600
Speaker 2: data intelligence firm Center Tower, and that's referring to these

31
00:01:53,600 --> 00:01:56,400
Speaker 2: monthly active users on apps and similar web which is

32
00:01:56,440 --> 00:01:59,160
Speaker 2: just unique monthly active visitors to websites. And this is

33
00:01:59,160 --> 00:02:02,000
Speaker 2: for the biggest players in AI in January twenty twenty five.

34
00:02:02,200 --> 00:02:05,080
Speaker 2: And I must be clear, the following really fucked me

35
00:02:05,160 --> 00:02:08,519
Speaker 2: up a little. So open AI's chatgpt had three hundred

36
00:02:08,520 --> 00:02:11,400
Speaker 2: and thirty nine million active users on chat GPT's app

37
00:02:11,520 --> 00:02:14,000
Speaker 2: and two hundred and forty six million unique monthly visitors

38
00:02:14,000 --> 00:02:17,959
Speaker 2: to chatgpt dot com in January twenty twenty five. Pretty good, right,

39
00:02:18,040 --> 00:02:21,160
Speaker 2: One would assume that everybody else, especially the hyperscalers, they'd

40
00:02:21,200 --> 00:02:24,000
Speaker 2: be pretty close behind. Right, They've got way more money,

41
00:02:24,120 --> 00:02:27,720
Speaker 2: They've got a ton of advertising, and one would be

42
00:02:27,760 --> 00:02:32,040
Speaker 2: wrong Google Gemini. Google's Gemini had a pathetic eighteen million

43
00:02:32,080 --> 00:02:34,679
Speaker 2: monthly active users on the Gemini app and a mere

44
00:02:34,760 --> 00:02:38,919
Speaker 2: forty seven point three million unique monthly users on their website.

45
00:02:39,040 --> 00:02:41,800
Speaker 2: For COMPARISONCNN dot com has over one hundred and fifty

46
00:02:41,840 --> 00:02:45,400
Speaker 2: million unique monthly visitors and did not require purchasing billions

47
00:02:45,440 --> 00:02:51,079
Speaker 2: of dollars of GPUs. Nevertheless, though, Microsoft, who funds open Ai,

48
00:02:51,600 --> 00:02:54,639
Speaker 2: they wouldn't have blown this right. They wouldn't have got

49
00:02:54,639 --> 00:02:58,800
Speaker 2: a tiny amount of US right. Wrong, so very wrong.

50
00:02:58,919 --> 00:03:02,560
Speaker 2: Microsoft Copilot had an embarrassing eleven million monthly active users

51
00:03:02,560 --> 00:03:05,400
Speaker 2: on the Copilot app and fifteen point six million unique

52
00:03:05,400 --> 00:03:09,480
Speaker 2: monthly users visitors. I mean to copilot dot Microsoft dot com.

53
00:03:09,919 --> 00:03:12,160
Speaker 2: They had a fucking Super Bowl commercial a year ago. Man,

54
00:03:12,200 --> 00:03:14,800
Speaker 2: what the hell? These are terrible numbers for a company

55
00:03:14,800 --> 00:03:17,320
Speaker 2: with a market capitalization of three trillion dollars that spent

56
00:03:17,360 --> 00:03:20,360
Speaker 2: over seventy five billion dollars on capital expenditures in twenty

57
00:03:20,400 --> 00:03:23,160
Speaker 2: twenty four. But you know what, maybe it's just that

58
00:03:23,160 --> 00:03:25,600
Speaker 2: big tech hasn't worked it out. The plucky start of

59
00:03:25,600 --> 00:03:28,080
Speaker 2: the Silicon Valley must have, right. I mean, we've all

60
00:03:28,080 --> 00:03:30,399
Speaker 2: heard about perplexity. They seem to be giving away pro

61
00:03:30,440 --> 00:03:33,200
Speaker 2: accounts all the time. They're pretty big, right, they must

62
00:03:33,240 --> 00:03:36,560
Speaker 2: be wrong. They had an abominable eight million monthly active

63
00:03:36,640 --> 00:03:39,240
Speaker 2: users on their app in January and a poultry ten

64
00:03:39,280 --> 00:03:42,360
Speaker 2: point six million unique monthly visitors to perplexity dot ai.

65
00:03:42,960 --> 00:03:45,880
Speaker 2: This company's raised six hundred and sixty five million dollars.

66
00:03:45,880 --> 00:03:48,360
Speaker 2: They have a multi billion dollar valuation. This is absolutely

67
00:03:48,400 --> 00:03:51,640
Speaker 2: fucking pathetic. What are we doing here? But you know,

68
00:03:51,720 --> 00:03:53,440
Speaker 2: maybe I'm just being a sour pus. Maybe I'm just

69
00:03:53,440 --> 00:03:55,120
Speaker 2: being just being a hater. You know, I sit in

70
00:03:55,120 --> 00:03:56,880
Speaker 2: my hates thrown and I hate it. I'm like, ooh,

71
00:03:56,880 --> 00:04:01,080
Speaker 2: I hate the company so much. Right, that's it. It's bias. Obviously,

72
00:04:01,480 --> 00:04:04,480
Speaker 2: there's one hulking Juggernau I've been leaving out and I

73
00:04:04,520 --> 00:04:07,920
Speaker 2: would never ever ever leave them out. Because Anthropic has

74
00:04:08,000 --> 00:04:11,600
Speaker 2: raised fourteen point seven billion dollars, one would assume with

75
00:04:11,720 --> 00:04:14,440
Speaker 2: all this money, with all of the press attention, all

76
00:04:14,480 --> 00:04:17,360
Speaker 2: of the people saying that Warrio Ami Day, sorry, Dario

77
00:04:17,480 --> 00:04:21,720
Speaker 2: Ami Day, will have Agi in twenty twenty seven around there, right,

78
00:04:21,800 --> 00:04:24,279
Speaker 2: they'd be huge. They must be huge. They must be huge,

79
00:04:24,360 --> 00:04:31,240
Speaker 2: right right? Right then, So flipping wrong. Anthropics Claude had

80
00:04:31,240 --> 00:04:34,719
Speaker 2: I a shit you not two million monthly active users

81
00:04:34,720 --> 00:04:37,480
Speaker 2: on the claud app and eight point two million unique

82
00:04:37,520 --> 00:04:39,960
Speaker 2: monthly visitors the Claude the AI and that's the web

83
00:04:40,080 --> 00:04:43,800
Speaker 2: based version of their app. These numbers are absolutely abominable.

84
00:04:43,839 --> 00:04:46,560
Speaker 2: They're trash, they're garbage. Anyone's saying that Anthropic as a

85
00:04:46,600 --> 00:04:49,240
Speaker 2: real product is talking out of their asshole. I'll get

86
00:04:49,279 --> 00:04:52,080
Speaker 2: to API calls later, don't you worry. But it's time

87
00:04:52,120 --> 00:04:54,040
Speaker 2: to wake up and stop yammering about these companies like

88
00:04:54,040 --> 00:04:56,920
Speaker 2: they're building the future or have any real product market fit.

89
00:04:57,040 --> 00:05:00,560
Speaker 2: What a goddamn joke. I am furious. Getting these numbers

90
00:05:00,600 --> 00:05:02,960
Speaker 2: pissed me off so much. I have had these chonder

91
00:05:03,000 --> 00:05:05,640
Speaker 2: fucks telling me, oh Ed, you have no idea how

92
00:05:05,640 --> 00:05:08,640
Speaker 2: big this is? Oh Ed, They're tiny. This is so small.

93
00:05:08,760 --> 00:05:11,599
Speaker 2: I know. I said it'd be calm. I said it'd

94
00:05:11,600 --> 00:05:14,919
Speaker 2: be calm. But when you spend two years having people

95
00:05:14,920 --> 00:05:16,799
Speaker 2: calling you a hater and a cynic at, a pessimist

96
00:05:16,800 --> 00:05:19,359
Speaker 2: and a pig and a dog and they spray you

97
00:05:19,360 --> 00:05:21,760
Speaker 2: at their hoses, you get a little angry about this stuff.

98
00:05:21,839 --> 00:05:24,160
Speaker 2: And also, why is nobody I'm a pr guy who

99
00:05:24,160 --> 00:05:26,359
Speaker 2: does a podcast and a newslre why am I the

100
00:05:26,360 --> 00:05:31,160
Speaker 2: person to say it? But what's really funny is that

101
00:05:31,200 --> 00:05:34,440
Speaker 2: the recently emerged deep Seak had twenty seven million monthly

102
00:05:34,480 --> 00:05:37,599
Speaker 2: active users on the deep seak Act app and seventy

103
00:05:37,720 --> 00:05:41,240
Speaker 2: nine point nine million unique monthly visitors to deepseak dot

104
00:05:41,240 --> 00:05:43,839
Speaker 2: com in January. This figure, by the way, doesn't capture

105
00:05:43,839 --> 00:05:47,279
Speaker 2: deep Seak's China based users, who at least on mobile

106
00:05:47,360 --> 00:05:50,479
Speaker 2: access the app through a variety of different marketplaces. From

107
00:05:50,480 --> 00:05:53,000
Speaker 2: what I can tell, the deep Seak Act app I

108
00:05:53,080 --> 00:05:55,360
Speaker 2: keep speaking like Kathy, I guess, but I'm gonna keep

109
00:05:55,360 --> 00:05:57,200
Speaker 2: this going. We don't need to edit that the app

110
00:05:57,240 --> 00:05:59,880
Speaker 2: has nearly ten million downloads on the Vivo Store, which

111
00:05:59,920 --> 00:06:02,359
Speaker 2: is just one of the different Android app marketplaces serving

112
00:06:02,400 --> 00:06:05,120
Speaker 2: mainline China. It's not even one of the biggest. But

113
00:06:05,240 --> 00:06:07,160
Speaker 2: for the sake of simplicity, assume that all of these

114
00:06:07,240 --> 00:06:09,760
Speaker 2: numbers refer to those outside of China, where most, if

115
00:06:09,800 --> 00:06:11,880
Speaker 2: not all, of the Western made chatbots are blocked by

116
00:06:11,880 --> 00:06:14,960
Speaker 2: the Great Firewall. But let me put this all into perspective.

117
00:06:15,400 --> 00:06:19,679
Speaker 2: The entire combined monthly active users of Microsoft Copilot, Claude Gemini,

118
00:06:19,720 --> 00:06:23,039
Speaker 2: Deep Seek, and it perplexitis apps amount to sixty six

119
00:06:23,160 --> 00:06:26,320
Speaker 2: million monthly active users, or nineteen point four to seven

120
00:06:26,360 --> 00:06:29,120
Speaker 2: percent of the entire monthly active users of chat GPT's

121
00:06:29,160 --> 00:06:34,120
Speaker 2: mobile app. Web traffic slightly improves things, I say sarcastically,

122
00:06:34,279 --> 00:06:36,559
Speaker 2: with one hundred and sixty one million unique monthly visitors

123
00:06:36,600 --> 00:06:39,520
Speaker 2: that visited the websites for Copilot, Claude, Gemini, Deep Seek,

124
00:06:39,520 --> 00:06:42,440
Speaker 2: and Perplexity making up sixty five point sixty nine percent

125
00:06:42,480 --> 00:06:46,279
Speaker 2: of all the traffic that went to chatgpt dot com. However,

126
00:06:46,520 --> 00:06:50,000
Speaker 2: I'd argue that including deep Seek vastly overinflates these numbers.

127
00:06:50,040 --> 00:06:52,240
Speaker 2: They're an outlier, and they're also relatively new, and they've

128
00:06:52,279 --> 00:06:55,000
Speaker 2: enjoyed a big moment in the media, so we can

129
00:06:55,080 --> 00:06:58,520
Speaker 2: leave them out for a second. Without deep Sea, Copilot, Claude, Gemini,

130
00:06:58,560 --> 00:07:00,920
Speaker 2: and Perplexity made up a total of thirty nine million

131
00:07:00,920 --> 00:07:03,479
Speaker 2: monthly active users across their apps, and a grand total

132
00:07:03,480 --> 00:07:07,559
Speaker 2: of eighty one point seven million unique monthly visitors Without

133
00:07:07,640 --> 00:07:10,440
Speaker 2: chat gpt, it appears that the entire general IVAI app

134
00:07:10,440 --> 00:07:12,160
Speaker 2: market is a little more than half the size of

135
00:07:12,160 --> 00:07:14,720
Speaker 2: Pokemon Goo its peak, which had around one hundred and

136
00:07:14,760 --> 00:07:17,480
Speaker 2: forty seven million monthly active users. Though this number is

137
00:07:17,560 --> 00:07:21,040
Speaker 2: kind of hard to chase down. I've heard two hundred million. Nevertheless,

138
00:07:21,480 --> 00:07:23,440
Speaker 2: even if it was one hundred million, it would still

139
00:07:23,480 --> 00:07:26,280
Speaker 2: be more. And while one can say I missed a

140
00:07:26,280 --> 00:07:30,880
Speaker 2: few apps xais, Grock and Mason's rufirst character Ai, there

141
00:07:30,920 --> 00:07:32,840
Speaker 2: isn't really a chance in hell they cover the shortfall.

142
00:07:33,560 --> 00:07:36,240
Speaker 2: These numbers aren't simply pissed poor. They're a sign that

143
00:07:36,240 --> 00:07:39,080
Speaker 2: the market for general ivai is incredibly small, and based

144
00:07:39,120 --> 00:07:40,840
Speaker 2: on the fact that every single one of these apps

145
00:07:40,880 --> 00:07:44,600
Speaker 2: only loses money, they're actively harmful to their respective investors

146
00:07:44,720 --> 00:07:47,600
Speaker 2: or owners. I do not think this is a real industry,

147
00:07:47,800 --> 00:07:49,200
Speaker 2: and I believe that if we pulled the plug on

148
00:07:49,240 --> 00:07:52,400
Speaker 2: the venture capital aspect tomorrow, it would die. It would

149
00:07:52,480 --> 00:07:55,720
Speaker 2: die within a month, maybe two. But let's talk about

150
00:07:55,760 --> 00:07:58,000
Speaker 2: API course, and this is when companies plug their apps

151
00:07:58,000 --> 00:08:01,320
Speaker 2: into open AIS models, anthropics models, Google's models, any company's

152
00:08:01,320 --> 00:08:05,080
Speaker 2: models to power some sort of supposedly amazing Generalivai feature.

153
00:08:05,600 --> 00:08:07,160
Speaker 2: And the counter I hear a lot is that these

154
00:08:07,200 --> 00:08:09,720
Speaker 2: API calls are a kind of hidden adoption, that there's

155
00:08:09,720 --> 00:08:13,040
Speaker 2: this massive swell of engaged, happy customers using Genera ivai.

156
00:08:13,280 --> 00:08:15,520
Speaker 2: They're just not using it on any of the major apps,

157
00:08:15,960 --> 00:08:18,800
Speaker 2: and the connection to these models that that's the real

158
00:08:18,920 --> 00:08:22,000
Speaker 2: success story here, because people are adopting general ivai, they're

159
00:08:22,040 --> 00:08:24,880
Speaker 2: just doing it through other apps. This isn't the case.

160
00:08:25,880 --> 00:08:28,640
Speaker 2: Open Ai, as I've established, is the largest player in

161
00:08:28,720 --> 00:08:31,800
Speaker 2: general ivai, making more revenue roughly four billion dollars in

162
00:08:31,840 --> 00:08:34,000
Speaker 2: twenty twenty four, though they lost five billion dollars after

163
00:08:34,040 --> 00:08:37,679
Speaker 2: revenue again open ai lost. They spent nine billion dollars

164
00:08:37,720 --> 00:08:40,480
Speaker 2: in twenty twenty four to lose five billion dollars. Anyway,

165
00:08:40,520 --> 00:08:42,439
Speaker 2: they still made more than everyone else in every other

166
00:08:42,440 --> 00:08:45,200
Speaker 2: private AI company. The closest I can get to an

167
00:08:45,280 --> 00:08:48,400
Speaker 2: estimate on how many actual developers integrate their applications through

168
00:08:48,440 --> 00:08:51,240
Speaker 2: open Ai is a statement from open ai from October

169
00:08:51,240 --> 00:08:53,520
Speaker 2: twenty twenty four's dev Day, where they said they had

170
00:08:53,559 --> 00:08:56,840
Speaker 2: over three million developers building apps using open AI's models.

171
00:08:57,600 --> 00:08:59,880
Speaker 2: And as I've discussed in the past, open AI's revenue

172
00:08:59,920 --> 00:09:03,520
Speaker 2: is heavily weighted towards its subscription business, with licensing access

173
00:09:03,520 --> 00:09:05,800
Speaker 2: to its models like GPT four to O making up

174
00:09:05,880 --> 00:09:08,520
Speaker 2: less than thirty percent around a billion dollars of their revenue,

175
00:09:09,000 --> 00:09:12,199
Speaker 2: and subscriptions to their premium products like chech, GPD plus teams, business,

176
00:09:12,240 --> 00:09:14,760
Speaker 2: pro government and so on make up the majority, around

177
00:09:14,800 --> 00:09:18,840
Speaker 2: three billion dollars in twenty twenty four. My argument's fairly simple.

178
00:09:19,120 --> 00:09:21,760
Speaker 2: Open ai is the most well known player in GENERITIAI,

179
00:09:21,960 --> 00:09:25,040
Speaker 2: and thus we can extrapolate from it to draw conclusions

180
00:09:25,080 --> 00:09:27,480
Speaker 2: about the wider industry. In the event that there was

181
00:09:27,520 --> 00:09:30,840
Speaker 2: a huge, meaningful industry integrating generative AI into distinct products

182
00:09:30,880 --> 00:09:34,640
Speaker 2: with mass market consumer adoption, open AI's API business would

183
00:09:34,640 --> 00:09:37,520
Speaker 2: be doing far, far more revenue. But let's get a

184
00:09:37,520 --> 00:09:40,960
Speaker 2: little more specific about what an API call is. When

185
00:09:41,000 --> 00:09:44,319
Speaker 2: a business plugs open AI's models or any other GENERATIVEAI

186
00:09:44,480 --> 00:09:46,880
Speaker 2: companies models into their apps and the customer triggers one,

187
00:09:46,960 --> 00:09:49,120
Speaker 2: such as just asking the app to summarize an email,

188
00:09:49,360 --> 00:09:51,880
Speaker 2: open ai charges the business both for the prompt, which

189
00:09:51,920 --> 00:09:53,960
Speaker 2: is the input, and the result, which is the output.

190
00:09:54,600 --> 00:09:57,040
Speaker 2: As a result, where weekly active users might be indicative

191
00:09:57,040 --> 00:09:59,839
Speaker 2: of attention to open AI's products, API calls are farm

192
00:09:59,840 --> 00:10:03,600
Speaker 2: more more indicative consumer and enterprise adoption and usage. In fact,

193
00:10:05,080 --> 00:10:07,120
Speaker 2: to be clear, I acknowledge that there are a lot,

194
00:10:07,280 --> 00:10:09,560
Speaker 2: a non specific amount, but a fair amount of app

195
00:10:09,600 --> 00:10:14,040
Speaker 2: developers and companies adopting GENERATIVEAI. However, judging on the revenue

196
00:10:14,040 --> 00:10:16,679
Speaker 2: both from open AI's developer focused business and the lack

197
00:10:16,760 --> 00:10:20,600
Speaker 2: of any real revenue for any business integrating GENERATIVAI, I

198
00:10:20,720 --> 00:10:24,240
Speaker 2: hypothesize that customers, which include developers integrating open AI's models

199
00:10:24,240 --> 00:10:27,160
Speaker 2: into both consumer facing apps and enterprise focused apps, are

200
00:10:27,200 --> 00:10:30,120
Speaker 2: not actually using these features that much. I should add

201
00:10:30,160 --> 00:10:32,079
Speaker 2: the open ai makes about two hundred million dollars a

202
00:10:32,160 --> 00:10:35,120
Speaker 2: year selling their models through Microsoft, meaning that their API

203
00:10:35,400 --> 00:10:38,240
Speaker 2: business may be as small as eight hundred million dollars. Again,

204
00:10:38,480 --> 00:10:44,040
Speaker 2: this is not profit, its revenue. Before we go forward,

205
00:10:44,040 --> 00:10:46,880
Speaker 2: there is also an alternative open ai is charging way, way,

206
00:10:46,880 --> 00:10:48,599
Speaker 2: way less for their models than they should, which is

207
00:10:48,640 --> 00:10:51,440
Speaker 2: an argument I made in the subprime AI crisis last year.

208
00:10:51,800 --> 00:10:54,040
Speaker 2: But accepting this argument means that at some point open

209
00:10:54,080 --> 00:10:56,640
Speaker 2: ai will have to become profitable, which they've shown no

210
00:10:56,720 --> 00:10:59,360
Speaker 2: signs of doing so, or they're going to have to

211
00:10:59,440 --> 00:11:02,160
Speaker 2: charge the action costs of running their unprofitable models. Do

212
00:11:02,200 --> 00:11:05,719
Speaker 2: you not see the problem there? If they have to

213
00:11:05,800 --> 00:11:08,120
Speaker 2: raise all the prices for this thing that people aren't

214
00:11:08,160 --> 00:11:11,880
Speaker 2: really using, why would they keep it? But you're wondering,

215
00:11:11,960 --> 00:11:16,120
Speaker 2: probably how bad is this for anthropic It's pretty disastrous.

216
00:11:16,720 --> 00:11:20,160
Speaker 2: The Information reported recently that Anthropic was projected, and I

217
00:11:20,280 --> 00:11:22,560
Speaker 2: should be clear this means made up, that they will

218
00:11:22,559 --> 00:11:24,880
Speaker 2: make at least twelve billion dollars in revenue in twenty

219
00:11:24,880 --> 00:11:26,880
Speaker 2: twenty seven, despite the fact they only made around nine

220
00:11:26,960 --> 00:11:29,600
Speaker 2: hundred million dollars in twenty twenty four and lost five

221
00:11:29,640 --> 00:11:34,000
Speaker 2: point six billion dollars. Somehow, Anthropic is currently raising two

222
00:11:34,040 --> 00:11:37,040
Speaker 2: billion dollars at a sixty billion dollar valuation for a

223
00:11:37,040 --> 00:11:39,280
Speaker 2: business that loses billions of dollars a year, with an

224
00:11:39,280 --> 00:11:41,800
Speaker 2: app within in store base of two million people and

225
00:11:41,840 --> 00:11:44,120
Speaker 2: a web present smaller than that of a niche hobbyist

226
00:11:44,160 --> 00:11:48,360
Speaker 2: news outlet. The Information also adds and CNBC reports as

227
00:11:48,400 --> 00:11:50,760
Speaker 2: well that sixty to seventy five percent of that revenue

228
00:11:50,800 --> 00:11:53,360
Speaker 2: came from API course, though this number is from September

229
00:11:53,400 --> 00:11:57,000
Speaker 2: twenty twenty four. So what they're making most of their

230
00:11:57,040 --> 00:11:59,120
Speaker 2: money from people integrating it, and they're making what a

231
00:11:59,120 --> 00:12:02,079
Speaker 2: few hundred million dollar dollars and it costs them billions

232
00:12:02,080 --> 00:12:04,840
Speaker 2: of dollars to serve this small customer base. This company

233
00:12:04,880 --> 00:12:08,120
Speaker 2: is not worth sixty billion dollars. Anthropic has raised fourteen

234
00:12:08,160 --> 00:12:11,040
Speaker 2: point seven billion dollars to create an also ran large

235
00:12:11,080 --> 00:12:14,680
Speaker 2: language model company that some people like more than open Ai,

236
00:12:14,760 --> 00:12:17,880
Speaker 2: with a competing consumer facing large language model called Claude

237
00:12:17,920 --> 00:12:19,800
Speaker 2: that has an install base of maybe two percent of

238
00:12:19,800 --> 00:12:22,240
Speaker 2: the five free to play games made by Clash of

239
00:12:22,280 --> 00:12:25,560
Speaker 2: Clans develop a super cell. Anthropic, much like open Ai,

240
00:12:25,640 --> 00:12:28,760
Speaker 2: has categorically failed to productize its large language model, but

241
00:12:28,840 --> 00:12:33,080
Speaker 2: the only product it appears to have pushed being computer use,

242
00:12:33,240 --> 00:12:36,920
Speaker 2: which is like operator by open Ai, and it's similarly useless,

243
00:12:36,960 --> 00:12:39,520
Speaker 2: and it can sometimes successfully do in minutes what would

244
00:12:39,600 --> 00:12:42,599
Speaker 2: only take you a few seconds with your hands. Anthropic,

245
00:12:42,760 --> 00:12:46,280
Speaker 2: like OpenAI, also has no mote. While they have, they've

246
00:12:46,320 --> 00:12:48,720
Speaker 2: got kind of chain of thought reasoning in their models

247
00:12:49,160 --> 00:12:51,840
Speaker 2: that has been, as I mentioned, commoditized by deep Seek.

248
00:12:52,160 --> 00:12:56,160
Speaker 2: Its models, again like OpenAI, are totally unprofitable. They're unsustainable

249
00:12:56,160 --> 00:12:59,080
Speaker 2: and heavily dependent on training data that has either run

250
00:12:59,120 --> 00:13:02,920
Speaker 2: out or is running out. Anthropic CEO Dario ami Day

251
00:13:02,960 --> 00:13:05,360
Speaker 2: is also a sleazy con man who, like Sam Mortmon,

252
00:13:05,400 --> 00:13:08,880
Speaker 2: continually promises that his company's AI systems will become powerful

253
00:13:08,880 --> 00:13:11,320
Speaker 2: and autonomous in a way that they've never shown they

254
00:13:11,360 --> 00:13:14,520
Speaker 2: have any possibility of becoming. He loves talking about AGI.

255
00:13:14,600 --> 00:13:16,960
Speaker 2: He's just like Sam Mortmon, He's just as big a conman.

256
00:13:17,040 --> 00:13:20,280
Speaker 2: Fuck Dario ama Day. Any investor in Anthropic needs to

257
00:13:20,320 --> 00:13:23,520
Speaker 2: seriously consider what it is they're investing in. Anthropic has,

258
00:13:23,600 --> 00:13:26,240
Speaker 2: other than iterating on its large language model, claud shown

259
00:13:26,280 --> 00:13:30,640
Speaker 2: little fundamental differentiation from the rest of the industry. Anthropics business, again,

260
00:13:30,720 --> 00:13:33,600
Speaker 2: like OpenAI, is entirely propped up by venture capital and

261
00:13:33,679 --> 00:13:36,679
Speaker 2: hyperscalar dollars Google and Amazon in this case, and without

262
00:13:36,720 --> 00:13:40,400
Speaker 2: them it would absolutely die almost immediately because they have

263
00:13:40,559 --> 00:13:46,360
Speaker 2: only ever lost money. Anthropics products are both unpopular and commoditized,

264
00:13:46,520 --> 00:13:49,320
Speaker 2: and they lost five point six billion dollars last year.

265
00:13:49,600 --> 00:13:52,880
Speaker 2: Stop dancing around this fact, stop it. Stop doing this.

266
00:13:52,920 --> 00:13:54,440
Speaker 2: We need to stop. If you remember of the media

267
00:13:54,480 --> 00:13:56,199
Speaker 2: writing about this, listening to this, I need you to

268
00:13:56,280 --> 00:14:01,439
Speaker 2: fucking stop. We by not reporting this every article, it's

269
00:14:01,520 --> 00:14:05,000
Speaker 2: journalistic malpractice. These companies will die. How can you not

270
00:14:05,080 --> 00:14:19,120
Speaker 2: see this? Let's talk about perplexity, and my general view

271
00:14:19,120 --> 00:14:22,120
Speaker 2: on perplexity is who gives a shit? Who cares perplexity?

272
00:14:22,200 --> 00:14:24,400
Speaker 2: A company valued at nine billion dollars towards the end

273
00:14:24,400 --> 00:14:26,840
Speaker 2: of twenty twenty four has eight million people a month

274
00:14:26,920 --> 00:14:29,320
Speaker 2: using its app, but the Financial Times reporting that they

275
00:14:29,320 --> 00:14:32,360
Speaker 2: have a grand total of fifteen million monthly active users

276
00:14:32,440 --> 00:14:36,960
Speaker 2: for an unprofitable search engine. Perplexity, like every Generativai company,

277
00:14:37,040 --> 00:14:40,520
Speaker 2: only ever loses money, and its product, Generaivai powered search

278
00:14:40,880 --> 00:14:43,720
Speaker 2: is so commoditized that it's actually remarkable that they still exist.

279
00:14:44,240 --> 00:14:47,320
Speaker 2: I mean, they're bigger than Anthropic. That's crazy. Other than

280
00:14:47,360 --> 00:14:49,560
Speaker 2: the slick design, there's little to be excited about here,

281
00:14:49,680 --> 00:14:52,880
Speaker 2: and eight million monthly active users is pathetic. It's embarrassing,

282
00:14:53,200 --> 00:14:55,360
Speaker 2: deeply embarrassing for a company with the majority of its

283
00:14:55,440 --> 00:14:59,120
Speaker 2: users on mobile aravins. Rivinus is a desperate man with

284
00:14:59,200 --> 00:15:02,000
Speaker 2: questionable intent that made a half hearted attempt to merge

285
00:15:02,040 --> 00:15:04,600
Speaker 2: with TikTok in January. Really funny bother, It's like, hey,

286
00:15:04,840 --> 00:15:07,000
Speaker 2: I have a really shitty company that loses a bunch

287
00:15:07,040 --> 00:15:09,200
Speaker 2: of money. Can I merge with your beloved app for

288
00:15:09,240 --> 00:15:12,040
Speaker 2: some reason? Like you need to do this. Also, their

289
00:15:12,040 --> 00:15:14,520
Speaker 2: product rips off journalists. By the way, they had a

290
00:15:14,560 --> 00:15:16,880
Speaker 2: whole thing for so they were just ripping fucking content.

291
00:15:16,920 --> 00:15:19,960
Speaker 2: Did it with Business Insider two. It's disgusting, But any

292
00:15:19,960 --> 00:15:22,720
Speaker 2: investor in Perplexity needs to ask themselves, what is it

293
00:15:22,760 --> 00:15:26,240
Speaker 2: I'm investing in an unprofitable search engine, an unprofitable large

294
00:15:26,320 --> 00:15:28,880
Speaker 2: language model company, A company that has such poor adoption

295
00:15:28,920 --> 00:15:30,640
Speaker 2: of its product that was prepared to become the shell

296
00:15:30,680 --> 00:15:34,280
Speaker 2: corporation for TikTok. Hmmm, Personally, I'd be concerned about the

297
00:15:34,320 --> 00:15:37,000
Speaker 2: bullshit numbers they keep making up. The information reported to

298
00:15:37,000 --> 00:15:39,000
Speaker 2: Perplexity said they'd make one hundred and twenty seven million

299
00:15:39,040 --> 00:15:41,320
Speaker 2: dollars in twenty twenty five and six hundred and fifty

300
00:15:41,360 --> 00:15:44,120
Speaker 2: six million dollars in twenty twenty six. How much money

301
00:15:44,120 --> 00:15:46,480
Speaker 2: did it make in twenty twenty four just over fifty

302
00:15:46,520 --> 00:15:48,440
Speaker 2: six million dollars. Is it profitable?

303
00:15:48,560 --> 00:15:48,800
Speaker 1: Fuck?

304
00:15:48,880 --> 00:15:52,680
Speaker 2: No, Perplexiti's product is commoditized, and they make less than

305
00:15:52,680 --> 00:15:54,400
Speaker 2: a quarter of the revenue of the baseball team in

306
00:15:54,400 --> 00:15:57,080
Speaker 2: the Oakland Athletics in twenty twenty four at least, though

307
00:15:57,080 --> 00:15:59,880
Speaker 2: I should add the Perplexity's app is marginally more popular.

308
00:16:01,000 --> 00:16:03,640
Speaker 2: It really is time to stop humoring these companies, though.

309
00:16:03,680 --> 00:16:06,040
Speaker 2: It's time to stop writing about them like their gifted children.

310
00:16:06,680 --> 00:16:10,560
Speaker 2: They are horrible. They are abominations of startups. They are

311
00:16:10,600 --> 00:16:15,320
Speaker 2: abominations of capitalism, which is already fairly abominable. I'm really

312
00:16:15,360 --> 00:16:18,600
Speaker 2: just disgusted reading these numbers. Jokeified me one hundred times.

313
00:16:18,720 --> 00:16:20,600
Speaker 2: I didn't even need to put on the joker makeup.

314
00:16:20,600 --> 00:16:23,400
Speaker 2: It just appeared on my skin naturally. I'm currently high

315
00:16:23,440 --> 00:16:27,280
Speaker 2: kicking around this sound cube I record everything in. But

316
00:16:27,440 --> 00:16:30,560
Speaker 2: really all of this is far more apocalyptic for the hyperscalers.

317
00:16:31,240 --> 00:16:33,760
Speaker 2: The Wall Street Journal reports that Microsoft intends to spend

318
00:16:33,840 --> 00:16:36,840
Speaker 2: ninety three point seven billion dollars in capital expenditures in

319
00:16:36,880 --> 00:16:40,040
Speaker 2: twenty twenty five, or roughly eighty five hundred and eighteen

320
00:16:40,160 --> 00:16:42,600
Speaker 2: dollars per monthly active user of the co Pilot app

321
00:16:42,600 --> 00:16:45,400
Speaker 2: in January twenty twenty five. Google is planning to spend

322
00:16:45,440 --> 00:16:48,360
Speaker 2: seventy five billion dollars on capital expenditures in twenty twenty five,

323
00:16:48,560 --> 00:16:51,040
Speaker 2: or roughly foury one hundred and sixty seven per monthly

324
00:16:51,240 --> 00:16:54,600
Speaker 2: active user of the Gemini app in January twenty twenty five.

325
00:16:54,960 --> 00:16:57,640
Speaker 2: Sundhapashai wants Gemini to be used by five hundred million

326
00:16:57,640 --> 00:16:59,840
Speaker 2: people before the end of twenty twenty five, and number

327
00:16:59,880 --> 00:17:02,320
Speaker 2: five so unrealistic that someone at Google should be fired,

328
00:17:02,840 --> 00:17:05,320
Speaker 2: and that someone is sunned up as shy. The fact

329
00:17:05,359 --> 00:17:07,399
Speaker 2: of the matter is that if Google and Microsoft can't

330
00:17:07,400 --> 00:17:10,720
Speaker 2: make generative AI apps work, if they can't get meaningful

331
00:17:10,720 --> 00:17:14,800
Speaker 2: consumer penetration. This entire industry is screwed. There really are

332
00:17:14,840 --> 00:17:17,240
Speaker 2: no optimistic ways to look at these numbers, and yes

333
00:17:17,280 --> 00:17:21,119
Speaker 2: I'm repeating myself. Microsoft Copilot had eleven million monthly active

334
00:17:21,160 --> 00:17:23,920
Speaker 2: users on the Copilot app and fifteen point six million

335
00:17:23,960 --> 00:17:27,320
Speaker 2: unique monthly visitors the coopilot dot Microsoft dot com. Google

336
00:17:27,359 --> 00:17:30,080
Speaker 2: Gemini had eighty million monthly active users on the Gemini

337
00:17:30,119 --> 00:17:33,040
Speaker 2: app and forty seven point three million unique monthly users

338
00:17:33,640 --> 00:17:38,119
Speaker 2: visitors even to their website. These are utterly pathetic considering

339
00:17:38,160 --> 00:17:42,080
Speaker 2: Microsoft and Google scale, especially given the latters complete dominance

340
00:17:42,119 --> 00:17:44,919
Speaker 2: over Google Search and web search in general, and the

341
00:17:44,920 --> 00:17:48,800
Speaker 2: ability to funnel customers to Gemini for millions, perhaps billions,

342
00:17:49,119 --> 00:17:51,280
Speaker 2: Google is the first page that they see when they

343
00:17:51,359 --> 00:17:55,440
Speaker 2: open a web browser. Google should be owning this by now. Look,

344
00:17:55,560 --> 00:17:58,040
Speaker 2: forty seven point three million unique monthly visitors is a

345
00:17:58,040 --> 00:18:00,320
Speaker 2: lot of people, But considering that Google spent five fifty

346
00:18:00,359 --> 00:18:04,520
Speaker 2: two zero point five four billion dollars in capital expenditures

347
00:18:04,520 --> 00:18:06,720
Speaker 2: in twenty twenty four, it's hard to see whether return

348
00:18:06,800 --> 00:18:09,720
Speaker 2: is or even whether a tone could be. Google, like

349
00:18:09,760 --> 00:18:12,760
Speaker 2: most companies, does not break out revenue from AI. Just

350
00:18:12,800 --> 00:18:14,840
Speaker 2: to be clear, if they were doing well, they would,

351
00:18:15,119 --> 00:18:17,040
Speaker 2: though they do love to say stuff like a strong

352
00:18:17,119 --> 00:18:19,480
Speaker 2: quarter was driven by our leadership in AI and momentum

353
00:18:19,480 --> 00:18:22,800
Speaker 2: across the business, which means nothing. By the way that

354
00:18:23,320 --> 00:18:25,840
Speaker 2: shit is made for journalisticy and go oh, that means

355
00:18:25,840 --> 00:18:28,200
Speaker 2: they're making money in AI. When a company's making money

356
00:18:28,240 --> 00:18:31,720
Speaker 2: in something, they'll tell you directly. And as a result

357
00:18:31,800 --> 00:18:34,439
Speaker 2: of its unwillingness to share hard numbers, all we have

358
00:18:34,560 --> 00:18:36,679
Speaker 2: to look at are numbers like those that received from

359
00:18:36,720 --> 00:18:39,280
Speaker 2: similar webons Centre Tower, and it's fair to suggest that

360
00:18:39,359 --> 00:18:42,800
Speaker 2: Gemini and its associated products have been a complete flop.

361
00:18:43,359 --> 00:18:46,520
Speaker 2: We're still Google spent one hundred and twenty seven point

362
00:18:46,600 --> 00:18:49,760
Speaker 2: five four billion dollars in capital expenditures in twenty twenty

363
00:18:49,760 --> 00:18:52,520
Speaker 2: three and twenty twenty four, combined with an estimated seventy

364
00:18:52,560 --> 00:18:54,960
Speaker 2: five billion like I said for twenty twenty five, what

365
00:18:55,000 --> 00:18:58,160
Speaker 2: the fuck is going on? Yes, Google is likely making

366
00:18:58,160 --> 00:19:01,239
Speaker 2: revenue from people running generally AIM and Google Cloud, and

367
00:19:01,359 --> 00:19:04,480
Speaker 2: yes they're likely making money from forcing AI onto Google

368
00:19:04,520 --> 00:19:07,200
Speaker 2: Workspace customers by raising the prices and saying you get

369
00:19:07,200 --> 00:19:11,240
Speaker 2: this for quote free. But Google, like every single other

370
00:19:11,320 --> 00:19:14,399
Speaker 2: general IVAI company, is losing money on every single general

371
00:19:14,440 --> 00:19:17,520
Speaker 2: IVAI prompt, and based on these monthly active user numbers,

372
00:19:17,560 --> 00:19:21,320
Speaker 2: nobody really cares about Gemini at all. Actually, I take

373
00:19:21,359 --> 00:19:24,520
Speaker 2: that back. Some people care about Gemini. Not that many,

374
00:19:24,560 --> 00:19:26,959
Speaker 2: but some. And it's far more fair to say that

375
00:19:27,000 --> 00:19:30,080
Speaker 2: nobody cares about Microsoft Copilot, despite Microsoft shoving it in

376
00:19:30,119 --> 00:19:33,359
Speaker 2: every corner of our lives. Eleven million monthly active users

377
00:19:33,400 --> 00:19:36,159
Speaker 2: for its unprofitable, heavily commoditized large language model app is

378
00:19:36,200 --> 00:19:38,600
Speaker 2: a joke, as are the fifteen point six million monthly

379
00:19:38,640 --> 00:19:41,320
Speaker 2: active users for its web presents, probably because it does

380
00:19:41,359 --> 00:19:43,800
Speaker 2: exactly the same shit that every other LMM does, and

381
00:19:43,800 --> 00:19:46,080
Speaker 2: if one knows it's powered by chet GPT, it's just

382
00:19:46,920 --> 00:19:52,160
Speaker 2: it's remarkable. Microsoft's copil app isn't just unpopular, it's irrelevant.

383
00:19:52,680 --> 00:19:55,960
Speaker 2: For comparison, Microsoft Teams has, according to a post for

384
00:19:56,119 --> 00:19:58,440
Speaker 2: Microsoft from the end of twenty twenty three, over three

385
00:19:58,520 --> 00:20:02,160
Speaker 2: hundred and twenty million months active users. That's more than

386
00:20:02,200 --> 00:20:04,399
Speaker 2: ten times the amount of monthly active users of the

387
00:20:04,480 --> 00:20:07,680
Speaker 2: Copilot app in January twenty twenty five, and the Copilot

388
00:20:07,680 --> 00:20:13,120
Speaker 2: website combined, and unlike Copilot teams, makes Microsoft money now.

389
00:20:13,119 --> 00:20:15,360
Speaker 2: I obviously don't have the numbers on people that accidentally

390
00:20:15,359 --> 00:20:18,200
Speaker 2: click the Copilot button in Microsoft Office or bing dot com.

391
00:20:18,240 --> 00:20:20,160
Speaker 2: But I do know that Microsoft isn't making much money

392
00:20:20,160 --> 00:20:23,159
Speaker 2: on AI at all. Microsoft reported in its last earnings

393
00:20:23,160 --> 00:20:25,960
Speaker 2: that it was making thirteen billion dollars of annual revenue,

394
00:20:26,040 --> 00:20:29,119
Speaker 2: a projected number based on current contracts versus booked money.

395
00:20:29,480 --> 00:20:33,119
Speaker 2: And this was on their artificial intelligence products. Now I've

396
00:20:33,160 --> 00:20:35,320
Speaker 2: made this point again and again and again, and I'm

397
00:20:35,320 --> 00:20:37,040
Speaker 2: going to keep making it. But revenue is not the

398
00:20:37,080 --> 00:20:39,520
Speaker 2: same thing as profit, and Microsoft does not have an

399
00:20:39,560 --> 00:20:42,879
Speaker 2: artificial intelligence part of its earnings breakdowns. These numbers are

400
00:20:42,960 --> 00:20:45,560
Speaker 2: cherry picked from across the entire suite of Microsoft products,

401
00:20:45,560 --> 00:20:47,719
Speaker 2: such as selling Copilot add ons to their Microsoft three

402
00:20:47,760 --> 00:20:49,720
Speaker 2: sixty five enter price suite. And by the way, The

403
00:20:49,760 --> 00:20:52,399
Speaker 2: Information reported in September twenty twenty four that Microsoft had

404
00:20:52,400 --> 00:20:55,720
Speaker 2: only sold Copilot to around one percent of their customers

405
00:20:55,760 --> 00:20:58,760
Speaker 2: buying three sixty five. They also make it selling access

406
00:20:58,760 --> 00:21:01,200
Speaker 2: to open AIS models on this roughly a billion dollars

407
00:21:01,240 --> 00:21:03,359
Speaker 2: in revenue, and people running their own models, and the

408
00:21:03,440 --> 00:21:07,320
Speaker 2: zore cloud, Microsoft's cloud compute platform for context. By the way,

409
00:21:07,320 --> 00:21:09,800
Speaker 2: Microsoft made sixty nine point sixty three billion dollars in

410
00:21:09,840 --> 00:21:13,800
Speaker 2: revenue in its last quarter, thirteen billion dollars of annual

411
00:21:14,119 --> 00:21:17,280
Speaker 2: revenue not profit, is about three point twenty five billion

412
00:21:17,320 --> 00:21:20,240
Speaker 2: dollars in quarterly revenue off of upwards of two hundred

413
00:21:20,240 --> 00:21:24,439
Speaker 2: billion dollars of capital expenditure since twenty twenty three. The

414
00:21:24,520 --> 00:21:28,320
Speaker 2: fact that neither Gemini nor Copilot has any meaningful consumer

415
00:21:28,359 --> 00:21:31,280
Speaker 2: penetration isn't just a joke. It should be sending alarm

416
00:21:31,320 --> 00:21:34,400
Speaker 2: bells through Wall Street. While Microsoft and Google may make

417
00:21:34,440 --> 00:21:37,600
Speaker 2: money outside of consumer software, both companies have desperately tried

418
00:21:37,640 --> 00:21:40,480
Speaker 2: to cram Copilot and Gemini down consumer's throats, and they

419
00:21:40,480 --> 00:21:44,800
Speaker 2: have categorically unquestionably failed or while burning billions of dollars

420
00:21:44,880 --> 00:21:48,520
Speaker 2: to do so. But ed ed, what about get hubcopilot?

421
00:21:48,560 --> 00:21:50,800
Speaker 2: All right, let's talk about git hub coopilot, shall we.

422
00:21:51,200 --> 00:21:53,080
Speaker 2: According to a report from The Wall Street Journal from

423
00:21:53,080 --> 00:21:56,000
Speaker 2: October twenty twenty three, Microsoft was losing an average of

424
00:21:56,440 --> 00:21:59,439
Speaker 2: more than twenty dollars a month per user on the

425
00:21:59,480 --> 00:22:03,119
Speaker 2: paid of gitthub copilot, what with some users costing them

426
00:22:03,160 --> 00:22:07,640
Speaker 2: more than eighty dollars a month. Jesus christ. Microsoft said

427
00:22:07,640 --> 00:22:09,840
Speaker 2: a year later that gethub copilot had one point eight

428
00:22:09,880 --> 00:22:13,040
Speaker 2: million paid subscribers, which is pretty good, except like all

429
00:22:13,040 --> 00:22:15,720
Speaker 2: generativio products, it loses money. Like I just bloody said,

430
00:22:16,320 --> 00:22:18,719
Speaker 2: I must repeat that Microsoft will have spent over two

431
00:22:18,840 --> 00:22:21,280
Speaker 2: hundred billion dollars in capital expenditures by the end of

432
00:22:21,280 --> 00:22:24,360
Speaker 2: twenty twenty five. In return, Microsoft got one point eight

433
00:22:24,400 --> 00:22:27,240
Speaker 2: million paying customers for a product that, like everything else

434
00:22:27,240 --> 00:22:30,000
Speaker 2: I'm talking about, is heavily commoditized. Basically, every LM can

435
00:22:30,040 --> 00:22:32,920
Speaker 2: generate code that some are better than others, by which

436
00:22:32,960 --> 00:22:35,320
Speaker 2: I mean they all introduced security issues into your code.

437
00:22:35,359 --> 00:22:39,320
Speaker 2: But nevertheless, and somehow, Microsoft loses money even when the

438
00:22:39,400 --> 00:22:42,760
Speaker 2: users use it paid. Am I getting through to you yet?

439
00:22:42,880 --> 00:22:45,320
Speaker 2: Is this working? If you are working for a hedge

440
00:22:45,359 --> 00:22:47,520
Speaker 2: fund and investment bank or anyone like that, please get

441
00:22:47,560 --> 00:22:50,280
Speaker 2: in touch. I will protect your identity. Is anyone around

442
00:22:50,320 --> 00:22:53,920
Speaker 2: you freaking out because they should be? They should be. Man,

443
00:22:54,200 --> 00:22:56,919
Speaker 2: I'm freaking out a little, and I just keep all

444
00:22:57,000 --> 00:22:59,680
Speaker 2: my money in a big box under my bed. I don't

445
00:22:59,760 --> 00:23:03,600
Speaker 2: have that I do anyway, not going to do that. Jug. So,

446
00:23:03,680 --> 00:23:06,480
Speaker 2: one of the arguments people make is that AI is everywhere,

447
00:23:06,480 --> 00:23:08,480
Speaker 2: But it's important to remember that the prevalence of AI

448
00:23:08,560 --> 00:23:11,040
Speaker 2: you seeing it in different apps, is not proof of

449
00:23:11,080 --> 00:23:13,560
Speaker 2: its adoption, but the intent of companies to shove it

450
00:23:13,560 --> 00:23:16,760
Speaker 2: into everything, And the same goes for businesses integrating AI

451
00:23:16,800 --> 00:23:19,360
Speaker 2: that are really just mandating people dick around with Copilot

452
00:23:19,400 --> 00:23:22,520
Speaker 2: or chat GPT and I'm really not kidding, no really.

453
00:23:22,720 --> 00:23:26,680
Speaker 2: KPMG bought forty seven thousand Microsoft Copilot subscriptions last year,

454
00:23:26,720 --> 00:23:30,240
Speaker 2: a significant discount to be familiar with any AI questions

455
00:23:30,480 --> 00:23:34,359
Speaker 2: their customers may have. Management consultancy PwC bought one hundred

456
00:23:34,440 --> 00:23:38,159
Speaker 2: thousand enterprise subscriptions, becoming open AI's largest customer in the process,

457
00:23:38,160 --> 00:23:40,320
Speaker 2: as well as their first reseller, and have created their

458
00:23:40,359 --> 00:23:45,960
Speaker 2: own internal generative AI called Chach PwC. The PWSC stuff

459
00:23:46,000 --> 00:23:49,000
Speaker 2: as absolutely hate. It's really cool that when you actually

460
00:23:49,040 --> 00:23:51,600
Speaker 2: talk to the users, they just fucking hate it. And

461
00:23:51,640 --> 00:23:54,680
Speaker 2: while you may see AI everywhere, integrations of generative AI

462
00:23:54,920 --> 00:23:57,920
Speaker 2: are indicative more of the decision making of the management

463
00:23:57,920 --> 00:24:00,280
Speaker 2: behind the platforms and the demands of the market more

464
00:24:00,320 --> 00:24:03,719
Speaker 2: than any consumer demand. Enterprise software is more often than

465
00:24:03,760 --> 00:24:06,639
Speaker 2: not sold in bolt to managers or c suite executive tasks,

466
00:24:06,720 --> 00:24:10,159
Speaker 2: less with company operations and messy things like doing staff

467
00:24:10,200 --> 00:24:12,760
Speaker 2: for making sure the company runs, more with seeming on

468
00:24:12,800 --> 00:24:16,119
Speaker 2: the forefront of technology. In practical terms, this means that

469
00:24:16,160 --> 00:24:18,400
Speaker 2: there is a lot of demand to put AI in stuff,

470
00:24:18,760 --> 00:24:21,480
Speaker 2: and some demand to buy stuff with AI on it

471
00:24:21,960 --> 00:24:25,760
Speaker 2: by enterprise buying software, but little evidence that this actually

472
00:24:25,840 --> 00:24:30,400
Speaker 2: leads the significant user adoption or usage. I'd argue this

473
00:24:30,440 --> 00:24:32,840
Speaker 2: is because large language models do not really lend themselves

474
00:24:32,880 --> 00:24:35,800
Speaker 2: to features that would provide meaningful business returns. And I

475
00:24:35,800 --> 00:24:37,840
Speaker 2: think everyone can agree on that, Like there are things

476
00:24:38,040 --> 00:24:40,480
Speaker 2: like summarizing emails, which I'll get to get to that

477
00:24:40,560 --> 00:24:43,600
Speaker 2: in a second. Look. In fact, let's do it now.

478
00:24:44,119 --> 00:24:47,320
Speaker 2: Let's briefly talk about where large language models work, where

479
00:24:47,760 --> 00:24:49,720
Speaker 2: they are actually good. And some of you are not

480
00:24:49,720 --> 00:24:51,000
Speaker 2: going to love this, but I know there's one of

481
00:24:51,040 --> 00:24:53,960
Speaker 2: you is that yes, yes, now I will get ahead.

482
00:24:54,040 --> 00:24:57,600
Speaker 2: I've got him now, I have him in my sights,

483
00:24:57,640 --> 00:24:59,560
Speaker 2: to be clear, And this is really dealing with the

484
00:24:59,600 --> 00:25:02,680
Speaker 2: am actually responses. I'm not saying and really have never

485
00:25:02,760 --> 00:25:05,560
Speaker 2: meant to say that large language models have no use

486
00:25:05,640 --> 00:25:09,560
Speaker 2: cases or no customers. People really do use them. They

487
00:25:09,640 --> 00:25:12,800
Speaker 2: use them for coding, for searching defined libraries of documents,

488
00:25:12,800 --> 00:25:16,200
Speaker 2: for generating draft materials, for brainstorming, for summarizing and searching documents,

489
00:25:16,240 --> 00:25:20,240
Speaker 2: These are useful, but they're not magical. They're cool, but

490
00:25:20,320 --> 00:25:24,760
Speaker 2: that's about it, and their coolness or usefulness is a

491
00:25:24,800 --> 00:25:28,120
Speaker 2: tiny little ant compared to the costs and stealing from

492
00:25:28,119 --> 00:25:32,919
Speaker 2: millions of people and damaging our power grid in our planet. Okay, okay,

493
00:25:32,960 --> 00:25:36,200
Speaker 2: so you're probably wondering. I brought it up earlier. Agents.

494
00:25:36,240 --> 00:25:39,000
Speaker 2: You've heard about agents. Mark Benioff wankin Off about agents.

495
00:25:38,960 --> 00:25:42,000
Speaker 2: Sam Allman talking about agents. They loved, They loved talking

496
00:25:42,040 --> 00:25:45,600
Speaker 2: about agents, right, they love saying agents of the future.

497
00:25:46,000 --> 00:25:48,760
Speaker 2: When a company uses the term agent, they're intentionally trying

498
00:25:48,800 --> 00:25:51,520
Speaker 2: to be deceitful, because the term agent means autonomous AI

499
00:25:51,560 --> 00:25:53,520
Speaker 2: that does stuff without you touching. It goes off and

500
00:25:53,560 --> 00:25:56,359
Speaker 2: does things for you with one command, and it knows

501
00:25:56,359 --> 00:25:59,560
Speaker 2: what to do. Remember, these models don't know anything. The

502
00:25:59,600 --> 00:26:02,000
Speaker 2: problem this definition is that everybody has used it to

503
00:26:02,080 --> 00:26:04,399
Speaker 2: refer to what is actually a chatbot that can do

504
00:26:04,480 --> 00:26:07,119
Speaker 2: some things while connected to a database, which I would

505
00:26:07,160 --> 00:26:11,040
Speaker 2: regularly called a chatbot personally. In open AI and anthropics case,

506
00:26:11,080 --> 00:26:13,439
Speaker 2: agents refer to a model that controls the computer. This

507
00:26:13,480 --> 00:26:15,320
Speaker 2: is closer to the truth other than the fact that

508
00:26:15,359 --> 00:26:18,040
Speaker 2: their agents are so unreliable as to be disqualifying, and

509
00:26:18,080 --> 00:26:20,840
Speaker 2: the tasks they succeed at, like searching trip advisor, are

510
00:26:20,960 --> 00:26:23,960
Speaker 2: very simple and did not need automating. Next time you

511
00:26:24,000 --> 00:26:25,960
Speaker 2: hear the agent actually look at what the product does,

512
00:26:26,000 --> 00:26:29,200
Speaker 2: and maybe you flick a booger at the person. But Ed, Ed,

513
00:26:29,440 --> 00:26:31,720
Speaker 2: you just burst into my door and having a nice

514
00:26:31,720 --> 00:26:33,359
Speaker 2: diet coke and you're in my house. What are you

515
00:26:33,440 --> 00:26:33,920
Speaker 2: doing here?

516
00:26:34,240 --> 00:26:34,320
Speaker 1: Ed?

517
00:26:34,359 --> 00:26:37,200
Speaker 2: What about artificial general intelligence? Aren't they going to turn

518
00:26:37,240 --> 00:26:39,560
Speaker 2: this into artificial general intelligence? No, they're not. Get out

519
00:26:39,600 --> 00:26:43,199
Speaker 2: of my house. Generative AI is probabilistic and large language

520
00:26:43,200 --> 00:26:45,359
Speaker 2: models do not know anything because they are guessing what

521
00:26:45,400 --> 00:26:47,679
Speaker 2: the next part of a particular output would be based

522
00:26:47,680 --> 00:26:50,359
Speaker 2: on the imput in reasoning models. They might look at

523
00:26:50,359 --> 00:26:51,919
Speaker 2: that a few times go oh, maybe it's not this,

524
00:26:52,000 --> 00:26:55,399
Speaker 2: maybe it's this. They are not making decisions. Generative AI

525
00:26:55,480 --> 00:26:59,240
Speaker 2: does not make decisions. They are probability machines, which in

526
00:26:59,280 --> 00:27:01,919
Speaker 2: turn makes them own only as reliable as probability can

527
00:27:02,000 --> 00:27:04,800
Speaker 2: be and as conscious no matter how intricate the system

528
00:27:04,840 --> 00:27:07,159
Speaker 2: may be or how much infrastructure is built. As a

529
00:27:07,200 --> 00:27:11,560
Speaker 2: pair of dice, we do not understand how human intelligence works,

530
00:27:11,640 --> 00:27:14,160
Speaker 2: and as a result, it's completely laughable to imagine we'd

531
00:27:14,160 --> 00:27:17,000
Speaker 2: be able to simulate it. Large language models do not

532
00:27:17,080 --> 00:27:21,520
Speaker 2: create or resemble, or they're not artificial intelligence. They are

533
00:27:21,920 --> 00:27:24,880
Speaker 2: at most the most powerful parrot in the world, trained

534
00:27:24,920 --> 00:27:27,119
Speaker 2: to respond to stimulus with what they guess is the

535
00:27:27,119 --> 00:27:30,760
Speaker 2: correct answer, and they're pretty good at it. They're pretty good. Right,

536
00:27:31,720 --> 00:27:35,000
Speaker 2: It's pretty cool, except we shouldn't be burning hundreds of

537
00:27:35,000 --> 00:27:37,600
Speaker 2: billions of dollars to make them slightly better at this.

538
00:27:38,000 --> 00:27:39,840
Speaker 2: Let me put it in simpler terms. Imagine if you

539
00:27:39,880 --> 00:27:41,679
Speaker 2: made a machine that threw a bouncy ball down a

540
00:27:41,680 --> 00:27:44,440
Speaker 2: hallway and it was really really you got really good

541
00:27:44,440 --> 00:27:46,000
Speaker 2: at dialing in to throw the ball so that it

542
00:27:46,000 --> 00:27:49,280
Speaker 2: followed a fairly exact trajectory. Would you think the arm

543
00:27:49,400 --> 00:27:53,600
Speaker 2: was intelligent? Would you think the ball was intelligent? Would

544
00:27:53,680 --> 00:27:56,879
Speaker 2: you think that the ability to precisely do something or

545
00:27:57,440 --> 00:28:01,440
Speaker 2: more reliably do something would make it smart. The point

546
00:28:01,480 --> 00:28:04,680
Speaker 2: I'm making about large language models is that they're a

547
00:28:04,760 --> 00:28:07,560
Speaker 2: cool concept with some interesting things they can do. But

548
00:28:07,600 --> 00:28:09,960
Speaker 2: they've been used as a cynical marketing vehicle to raise

549
00:28:09,960 --> 00:28:12,359
Speaker 2: money for open AI by lying about what they're capable

550
00:28:12,400 --> 00:28:30,080
Speaker 2: of doing, starting with calling them artificial intelligence. All right, really, though,

551
00:28:30,359 --> 00:28:32,960
Speaker 2: at this point, I need to ask a very fucking

552
00:28:33,000 --> 00:28:36,159
Speaker 2: simple question. Where is the goddamn money? Where is the

553
00:28:36,160 --> 00:28:39,760
Speaker 2: goddamn money? Where's the money? Sammy, Sammy, give me the money.

554
00:28:39,800 --> 00:28:42,560
Speaker 2: Where's the money? Money me money now? Sam Mortman, where

555
00:28:42,640 --> 00:28:45,160
Speaker 2: is the money? Dario Ama Day, where's the money? Satya Nadella,

556
00:28:45,200 --> 00:28:47,960
Speaker 2: where's the money? Summed up Ashai? Where's the money? Mark Benioff,

557
00:28:47,960 --> 00:28:51,160
Speaker 2: where's the money? Where is the goddamn money? Because revenue

558
00:28:51,200 --> 00:28:53,560
Speaker 2: is not the same as profit, I will say it again,

559
00:28:53,840 --> 00:28:58,160
Speaker 2: revenue is not the same as profit. And even then, Google, Amazon,

560
00:28:58,200 --> 00:29:00,640
Speaker 2: and to an extent, Microsoft, the company's making the most

561
00:29:00,680 --> 00:29:03,200
Speaker 2: investments in AI, do not want to state what the

562
00:29:03,240 --> 00:29:06,960
Speaker 2: revenue is on AI. I hypothesize the reason that they

563
00:29:06,960 --> 00:29:09,160
Speaker 2: do not want to disclose it is that it's pretty

564
00:29:09,200 --> 00:29:13,640
Speaker 2: goddamn small. It is extremely worrying that so few companies

565
00:29:13,680 --> 00:29:16,600
Speaker 2: are willing to directly disclose their revenue from selling services

566
00:29:16,600 --> 00:29:21,240
Speaker 2: that are allegedly revolutionary. Why Salesforce says they close two

567
00:29:21,280 --> 00:29:24,160
Speaker 2: hundred AI related deals in their last earnings. How much

568
00:29:24,200 --> 00:29:26,440
Speaker 2: money did they make? Why does Google get away with

569
00:29:26,480 --> 00:29:29,240
Speaker 2: saying that they have growing demand for AI and nothing else?

570
00:29:29,760 --> 00:29:33,880
Speaker 2: Is it? Because nobody's making that much money. As a sidebar,

571
00:29:34,080 --> 00:29:36,960
Speaker 2: I can find and I've really really looked like one

572
00:29:37,040 --> 00:29:39,600
Speaker 2: company that appears to be making profit from genera ivai

573
00:29:39,880 --> 00:29:42,800
Speaker 2: during a consultancy that helps generaivai companies find people to

574
00:29:42,800 --> 00:29:45,200
Speaker 2: train their models. That made three hundred million dollars in

575
00:29:45,240 --> 00:29:48,840
Speaker 2: revenue in twenty twenty four and reach an indeterminate amount

576
00:29:48,840 --> 00:29:50,560
Speaker 2: of profitability. We don't know if it was like a

577
00:29:50,600 --> 00:29:54,120
Speaker 2: million dollars or not, while Microsoft may disclose it made

578
00:29:54,200 --> 00:29:57,560
Speaker 2: thirteen million dollars in AI revenue that's annualized, so projected

579
00:29:57,600 --> 00:30:00,640
Speaker 2: based on current contracts rather than actual money in accounts,

580
00:30:00,840 --> 00:30:03,320
Speaker 2: and does not speak to the specific line items like

581
00:30:03,360 --> 00:30:05,800
Speaker 2: one would say. If said line items, we're not going

582
00:30:05,880 --> 00:30:08,280
Speaker 2: to make. The market say, hey, what the fuck? Put

583
00:30:08,280 --> 00:30:10,920
Speaker 2: aside whatever fantastical beliefs you may have about the future,

584
00:30:10,920 --> 00:30:13,360
Speaker 2: and tell me right now what business use case exists

585
00:30:13,520 --> 00:30:16,440
Speaker 2: that justifies burning hundreds of billions of dollars, damaging our

586
00:30:16,480 --> 00:30:18,880
Speaker 2: power grid, hurting our planet, and stealing from millions of

587
00:30:18,880 --> 00:30:21,720
Speaker 2: people to train these models. Even if you can put

588
00:30:21,720 --> 00:30:25,960
Speaker 2: troublesome things like morals or the basic principles of finance aside,

589
00:30:26,040 --> 00:30:28,960
Speaker 2: can AI evangelists not see that their dream is failing.

590
00:30:29,240 --> 00:30:31,880
Speaker 2: Can they not see that nothing is really happening, That

591
00:30:31,960 --> 00:30:34,520
Speaker 2: general ivai at best can be kind of cool, yet

592
00:30:34,520 --> 00:30:38,000
Speaker 2: mostly sucks and comes at this unbearable moral, financial, and

593
00:30:38,080 --> 00:30:41,880
Speaker 2: environmental cost. Is any of this really worth it? And

594
00:30:41,920 --> 00:30:45,640
Speaker 2: where exactly does this end? Do you AI evangelist gone

595
00:30:45,680 --> 00:30:48,080
Speaker 2: to your head, your life contingent on the truth, leaving

596
00:30:48,080 --> 00:30:50,640
Speaker 2: your lips believe that this goes much further than you

597
00:30:50,680 --> 00:30:54,440
Speaker 2: see today? Do not know if these AI people see

598
00:30:54,480 --> 00:30:57,000
Speaker 2: that this kind of sucks? Do they not see that

599
00:30:57,040 --> 00:30:59,719
Speaker 2: general ifai runs contrary to the basic tenets of what

600
00:30:59,760 --> 00:31:03,280
Speaker 2: makes science fiction call It doesn't make humans better. It

601
00:31:03,320 --> 00:31:06,080
Speaker 2: reduces their work to a stagnant, unremarkable slop in every

602
00:31:06,120 --> 00:31:08,560
Speaker 2: way it can, and reduces the cognition of those who

603
00:31:08,560 --> 00:31:10,320
Speaker 2: come to rely on it, And it costs hundreds of

604
00:31:10,320 --> 00:31:12,720
Speaker 2: billions of dollars and a return to fossil fuels. For

605
00:31:12,760 --> 00:31:16,600
Speaker 2: some fucking reason, it isn't working. The users aren't there,

606
00:31:16,720 --> 00:31:19,000
Speaker 2: the revenue isn't there. The best time to stop this

607
00:31:19,040 --> 00:31:20,680
Speaker 2: was two years ago, and the next best time to

608
00:31:20,680 --> 00:31:24,080
Speaker 2: stop is as soon as humanly possible. Generatifai is a

609
00:31:24,080 --> 00:31:27,840
Speaker 2: group delusion, its own kind of real life hallucination. What

610
00:31:27,880 --> 00:31:29,920
Speaker 2: you're seeing in the news is not the success of

611
00:31:29,960 --> 00:31:33,640
Speaker 2: the artificial intelligence industry, but a runaway narrative created by

612
00:31:33,640 --> 00:31:37,560
Speaker 2: and sustained by Sam Altman, Open Ai, Dario Amadee, and

613
00:31:37,680 --> 00:31:41,440
Speaker 2: of course satch In Nedella these fucking people. What you're

614
00:31:41,480 --> 00:31:44,640
Speaker 2: watching is not a revolution but a repetitious public relations

615
00:31:44,640 --> 00:31:47,240
Speaker 2: campaign for one company that accidentally time the launch of

616
00:31:47,280 --> 00:31:50,320
Speaker 2: chat GBT with a period of deep desperation in big tech,

617
00:31:50,560 --> 00:31:52,640
Speaker 2: one so profound that it will likely drag half a

618
00:31:52,680 --> 00:31:56,760
Speaker 2: trillion dollars worth of capital expenditures along with it. The

619
00:31:56,760 --> 00:31:58,920
Speaker 2: bubble will only burst when either the markets or the

620
00:31:59,000 --> 00:32:02,360
Speaker 2: hyperscalers accept that they've chased their own tales toward oblivion.

621
00:32:02,680 --> 00:32:05,800
Speaker 2: There is no justification for any of the capital expenditures

622
00:32:05,840 --> 00:32:08,560
Speaker 2: related to generative AI. We are approaching the limit of

623
00:32:08,560 --> 00:32:11,720
Speaker 2: what transformer based architecture can do, if we haven't already

624
00:32:11,800 --> 00:32:15,280
Speaker 2: reached it. No amount of beating off about test time, compute,

625
00:32:15,280 --> 00:32:18,040
Speaker 2: and connecting large language models to other large language models

626
00:32:18,080 --> 00:32:20,240
Speaker 2: is going to create a new use case for this technology,

627
00:32:20,360 --> 00:32:22,280
Speaker 2: and even if it did, it's unlikely that it ever

628
00:32:22,320 --> 00:32:25,320
Speaker 2: makes enough money to make it profitable. I will keep

629
00:32:25,360 --> 00:32:28,120
Speaker 2: talking about this stuff until I'm proven wrong. I do

630
00:32:28,120 --> 00:32:30,240
Speaker 2: not know why more people aren't more worried about this.

631
00:32:30,600 --> 00:32:33,440
Speaker 2: The financials are truly damning. The user numbers are so

632
00:32:33,480 --> 00:32:36,280
Speaker 2: small as to be insignificant. The costs are so ruinous

633
00:32:36,440 --> 00:32:38,560
Speaker 2: that they will likely cost tens of thousands of people

634
00:32:38,600 --> 00:32:41,200
Speaker 2: their jobs, and one of the hyperscalar CEOs their job

635
00:32:41,240 --> 00:32:43,920
Speaker 2: along with it, although admittedly I'm a lot less upset

636
00:32:43,960 --> 00:32:46,360
Speaker 2: about that. And they're going to inflict damage on tech

637
00:32:46,440 --> 00:32:49,760
Speaker 2: valuations that may well rival the dot com boom or worse.

638
00:32:50,320 --> 00:32:53,240
Speaker 2: And if the last point feels distant to you, ask yourself,

639
00:32:53,680 --> 00:32:57,240
Speaker 2: what's in your retirement savings? That's right, Google, Microsoft, and

640
00:32:57,320 --> 00:32:59,000
Speaker 2: hundreds of other companies that will be hurt by the

641
00:32:59,000 --> 00:33:03,720
Speaker 2: contagion of an aiding. I should also not be the

642
00:33:03,840 --> 00:33:06,960
Speaker 2: person saying this, or at least I shouldn't be one

643
00:33:06,960 --> 00:33:09,720
Speaker 2: of the first. These numbers are horrifying, and I have

644
00:33:09,800 --> 00:33:13,320
Speaker 2: no idea why nobody else is worried. There's no industry here,

645
00:33:13,400 --> 00:33:15,440
Speaker 2: there is no money. There's no proof that this will

646
00:33:15,480 --> 00:33:18,120
Speaker 2: ever turn into a real industry, and far more proof

647
00:33:18,120 --> 00:33:20,040
Speaker 2: that it will cost more money than it will ever make.

648
00:33:20,120 --> 00:33:24,040
Speaker 2: Im perpetuity open AI and anthropic are not real companies.

649
00:33:24,280 --> 00:33:27,120
Speaker 2: They're freeloaders living on venture backed welfare for an indeterminate

650
00:33:27,160 --> 00:33:29,640
Speaker 2: amount of time. Because the entire tech industry has agreed

651
00:33:29,680 --> 00:33:32,680
Speaker 2: to rally around the world's most unprofitable software, and, like

652
00:33:32,800 --> 00:33:35,400
Speaker 2: any other free ride that doesn't actually produce anything, when

653
00:33:35,440 --> 00:33:39,720
Speaker 2: the money goes away, they're fucked. Seriously, why are investors

654
00:33:39,800 --> 00:33:43,120
Speaker 2: funding open Ai? Do they seriously believe it's necessary to

655
00:33:43,200 --> 00:33:45,760
Speaker 2: let Sam Altman and open Ai continue to burn five

656
00:33:45,880 --> 00:33:48,080
Speaker 2: or more billion dollars a year on the off chance

657
00:33:48,080 --> 00:33:52,600
Speaker 2: he's able to create something that's alive. This motherfucker can't

658
00:33:52,600 --> 00:33:55,760
Speaker 2: create something that's profitable. What's the end point here? How

659
00:33:55,760 --> 00:33:58,840
Speaker 2: many more billions? Where's the fucking money? Sammy? Where is it?

660
00:33:58,880 --> 00:34:03,040
Speaker 2: Sam Moorman? Where's my god damn money? Where's my money? Sam?

661
00:34:03,280 --> 00:34:06,680
Speaker 2: I say all this because generativeai is open Ai. The

662
00:34:06,680 --> 00:34:10,279
Speaker 2: consumer adoption of this software is completely failed, and it's

663
00:34:10,280 --> 00:34:14,080
Speaker 2: going nowhere fast Chat GPT is sustained entirely on deranged,

664
00:34:14,080 --> 00:34:16,799
Speaker 2: specious hype drummed up by a media industry that thinks

665
00:34:16,840 --> 00:34:18,920
Speaker 2: it's more remarkable to write down the last lie that

666
00:34:18,960 --> 00:34:21,600
Speaker 2: Sam Moltman told than say that open ai has lost

667
00:34:21,640 --> 00:34:24,480
Speaker 2: nine billion dollars in the last year to Sorry, they

668
00:34:24,560 --> 00:34:27,800
Speaker 2: spent nine billion dollars to lose five billion dollars in

669
00:34:27,920 --> 00:34:29,520
Speaker 2: last year, and they intend to more than double that

670
00:34:29,600 --> 00:34:34,799
Speaker 2: number in twenty twenty five for absolutely no reason. Look, look,

671
00:34:35,560 --> 00:34:37,800
Speaker 2: it's time to stop humoring open ai and time to

672
00:34:37,840 --> 00:34:41,080
Speaker 2: stop directly stating that it is a bad business without

673
00:34:41,120 --> 00:34:44,200
Speaker 2: a meaningful product. We also really need to be clear

674
00:34:44,200 --> 00:34:46,680
Speaker 2: that the generative AI industry does not really exist without

675
00:34:46,719 --> 00:34:50,080
Speaker 2: open Ai, and thus this company must justify its existence.

676
00:34:50,360 --> 00:34:54,560
Speaker 2: And let's be abundantly clear, open Ai cannot exist any

677
00:34:54,640 --> 00:34:59,040
Speaker 2: further without further venture capital investment. This company has absolutely

678
00:34:59,080 --> 00:35:01,680
Speaker 2: no path to sustain in itself, no mode, and loses

679
00:35:01,719 --> 00:35:03,960
Speaker 2: so much money that it will need more than fifty

680
00:35:03,960 --> 00:35:06,239
Speaker 2: billion dollars to continue in its current form in the

681
00:35:06,280 --> 00:35:10,080
Speaker 2: next year. I don't know how I'm wrong, and I've

682
00:35:10,080 --> 00:35:12,680
Speaker 2: sat and thought through and researched a great deal on

683
00:35:12,760 --> 00:35:15,680
Speaker 2: how I might be. I can't find any compelling arguments.

684
00:35:15,880 --> 00:35:17,160
Speaker 2: I don't know what to do, but tell you what

685
00:35:17,200 --> 00:35:18,640
Speaker 2: I think and why I think that way, and hope

686
00:35:18,680 --> 00:35:21,120
Speaker 2: that you the listener understand a little bit more about

687
00:35:21,120 --> 00:35:24,600
Speaker 2: what I think is going on, because this really bothers me.

688
00:35:25,200 --> 00:35:27,240
Speaker 2: As I've said before, I grew up on the compute.

689
00:35:27,239 --> 00:35:29,239
Speaker 2: The computer made me who I am. Seeing the tech

690
00:35:29,320 --> 00:35:32,000
Speaker 2: industry like this sickens me because it's getting money away

691
00:35:32,040 --> 00:35:35,160
Speaker 2: from people doing cool shit to the least cool people

692
00:35:35,160 --> 00:35:39,080
Speaker 2: doing the least coolshit possible, and it's frustrating. But I'll

693
00:35:39,120 --> 00:35:41,240
Speaker 2: leave you with one thought and one thing that particularly

694
00:35:41,280 --> 00:35:44,719
Speaker 2: bothers me about GENERATIVAI. Regular people, for the most part

695
00:35:44,719 --> 00:35:47,160
Speaker 2: of my experience, do not seem to want this. While

696
00:35:47,160 --> 00:35:49,440
Speaker 2: there are occasionally people I'll mate you use chat GPT

697
00:35:49,600 --> 00:35:52,080
Speaker 2: to rewrite part of an email, most of the people

698
00:35:52,120 --> 00:35:54,640
Speaker 2: I may feel like AI was forced upon them. With

699
00:35:54,719 --> 00:35:57,360
Speaker 2: that in mind, I believe that Apple is actually radicalizing

700
00:35:57,400 --> 00:36:00,520
Speaker 2: millions of people against GENERATIVEAI by forcing them to reckon

701
00:36:00,560 --> 00:36:04,080
Speaker 2: with the terrible summaries, awful suggested texts, and horribly designed

702
00:36:04,160 --> 00:36:07,800
Speaker 2: user interface choices from Apple Intelligence, one of the worst

703
00:36:07,800 --> 00:36:10,800
Speaker 2: product launches I think I've seen in my life. Something

704
00:36:10,800 --> 00:36:13,560
Speaker 2: about generative AI has caused the hyperscalers to truly lose

705
00:36:13,560 --> 00:36:16,680
Speaker 2: their minds, and the intrusion of GENERATIVAI into both Microsoft

706
00:36:16,680 --> 00:36:19,440
Speaker 2: Office and Google Docs has turned just about everybody I

707
00:36:19,480 --> 00:36:22,440
Speaker 2: know in the business world against them. The resentment boiling

708
00:36:22,480 --> 00:36:24,880
Speaker 2: against this software is profound because the tech industry has

709
00:36:24,920 --> 00:36:28,320
Speaker 2: become desperate and violative, showing such contempt for their customers

710
00:36:28,320 --> 00:36:31,600
Speaker 2: that even Apple will force an inferior experience upon their

711
00:36:31,640 --> 00:36:34,040
Speaker 2: customers to please the will of the rot economy and

712
00:36:34,040 --> 00:36:36,920
Speaker 2: the growth at all cost mindset of the markets. Let's

713
00:36:36,920 --> 00:36:41,560
Speaker 2: be frank, nobody really needs anything generative AI does. Large

714
00:36:41,640 --> 00:36:44,440
Speaker 2: language models hallucinate too much to be truly reliable, a

715
00:36:44,480 --> 00:36:48,120
Speaker 2: problem that will require entirely new branches of mathematics to solve,

716
00:36:48,320 --> 00:36:51,920
Speaker 2: and their most common consumer facing functions like summarizing an article,

717
00:36:52,120 --> 00:36:54,719
Speaker 2: practicing for a job interview, or writing a business plan

718
00:36:54,920 --> 00:36:57,760
Speaker 2: are not really things people need or massively benefit from.

719
00:36:57,800 --> 00:37:01,280
Speaker 2: Even if these things weren't ruinously expensive damaging to the environment.

720
00:37:02,120 --> 00:37:05,040
Speaker 2: I believe general if AI is turning regular people against

721
00:37:05,080 --> 00:37:07,239
Speaker 2: the tech industry thanks to how much they're trying to

722
00:37:07,239 --> 00:37:10,120
Speaker 2: force it upon them and make them use the bad idea,

723
00:37:10,840 --> 00:37:14,800
Speaker 2: and it isn't working. Nobody wants this ship. They're intrigued

724
00:37:14,800 --> 00:37:16,719
Speaker 2: by the idea, then immediately bounce off of it in

725
00:37:16,760 --> 00:37:19,400
Speaker 2: many cases once they see what it can or can't do.

726
00:37:19,920 --> 00:37:22,160
Speaker 2: This software is being forced on people at scale by

727
00:37:22,200 --> 00:37:26,000
Speaker 2: corporations desperate to seem futuristic without any real understanding as

728
00:37:26,040 --> 00:37:28,200
Speaker 2: to why they need to do so, and whatever use

729
00:37:28,280 --> 00:37:31,000
Speaker 2: cases may exist for large language models are dwarfed by

730
00:37:31,000 --> 00:37:34,600
Speaker 2: how utterly unprofitable this whole fucking fiasco is. But I

731
00:37:34,640 --> 00:37:36,960
Speaker 2: want you to do something. If this whole thing has

732
00:37:37,040 --> 00:37:40,200
Speaker 2: pissed you off, if hearing about these companies has enraged

733
00:37:40,239 --> 00:37:43,080
Speaker 2: you as it has enraged me. I want you to

734
00:37:43,239 --> 00:37:46,919
Speaker 2: tell your friends, your family that open Ai spent nine

735
00:37:46,960 --> 00:37:49,480
Speaker 2: billion dollars to lose five billion dollars. I want you

736
00:37:49,520 --> 00:37:51,680
Speaker 2: to talk about the fact that large language models don't

737
00:37:51,680 --> 00:37:54,920
Speaker 2: really have a market. That chat GPT is a marketing

738
00:37:54,960 --> 00:37:59,680
Speaker 2: con That Sachinidella of Microsoft has burned probably will burn too,

739
00:37:59,760 --> 00:38:02,560
Speaker 2: un billion dollars chasing software that does the same thing

740
00:38:02,560 --> 00:38:05,879
Speaker 2: as everyone else. That Tim Cook of Apple has forced

741
00:38:05,880 --> 00:38:09,080
Speaker 2: Apple Intelligence barely functional software on millions and millions of

742
00:38:09,080 --> 00:38:12,040
Speaker 2: people because Apple has no more ideas. That Mark Zuckerberg

743
00:38:12,120 --> 00:38:14,759
Speaker 2: of Meta has pushed AI on everything because he has

744
00:38:14,800 --> 00:38:17,200
Speaker 2: no more ideas left, and it's going to burn so

745
00:38:17,320 --> 00:38:20,360
Speaker 2: much money in Meta and Sam Altman and Daria Amaday

746
00:38:20,520 --> 00:38:23,400
Speaker 2: are two fucking liars, two liars who will tell you

747
00:38:23,440 --> 00:38:27,200
Speaker 2: that their software will turn into artificial general intelligence. They're lying.

748
00:38:27,840 --> 00:38:31,560
Speaker 2: They're all lying. And Sundharpeshai of Google is the arc

749
00:38:31,680 --> 00:38:34,440
Speaker 2: liar who went in Google io and talked about a

750
00:38:34,440 --> 00:38:39,240
Speaker 2: completely fictional AI agent returning some shoes. The only reason

751
00:38:39,440 --> 00:38:41,520
Speaker 2: that companies like this lie like this is because the

752
00:38:41,520 --> 00:38:44,920
Speaker 2: truth is boring. The truth is mediocre. Believe your goddamn

753
00:38:44,960 --> 00:38:47,080
Speaker 2: eyes when you use chat GPT and you say, what

754
00:38:47,120 --> 00:38:50,160
Speaker 2: the fuck is this? Why does this matter? You're not crazy,

755
00:38:51,160 --> 00:38:53,560
Speaker 2: You're not an idiot because you can't see the magic

756
00:38:53,560 --> 00:38:57,520
Speaker 2: of generative AI. The people pushing this stuff are rather credulous.

757
00:38:58,600 --> 00:39:01,799
Speaker 2: They're either bought or they're making a shit ton of

758
00:39:01,800 --> 00:39:04,520
Speaker 2: money and lying to you. I refuse to let them

759
00:39:04,560 --> 00:39:07,239
Speaker 2: keep doing so. Better off Lene exists to kind of

760
00:39:07,520 --> 00:39:09,759
Speaker 2: explain this stuff as plainly as possible. You can reach

761
00:39:09,800 --> 00:39:11,600
Speaker 2: out to me if I wasn't clear about something. I

762
00:39:11,680 --> 00:39:14,200
Speaker 2: love feedback. Please be nice to me. I guess is

763
00:39:14,239 --> 00:39:16,799
Speaker 2: what I ask. But I love doing this. I am

764
00:39:16,840 --> 00:39:18,560
Speaker 2: scared of how this ends. I think it's going to

765
00:39:18,640 --> 00:39:21,400
Speaker 2: really hurt the markets. Could be three months, could be

766
00:39:21,440 --> 00:39:23,759
Speaker 2: six months, could be eighteen months. I don't know, and

767
00:39:23,880 --> 00:39:27,400
Speaker 2: I don't really think it's the right thing to do

768
00:39:27,480 --> 00:39:30,759
Speaker 2: to predict, but I'll tell you this, a better tech

769
00:39:30,800 --> 00:39:32,759
Speaker 2: industry can come out of this collapse. I don't know

770
00:39:32,800 --> 00:39:34,720
Speaker 2: how severe the collapse is. I don't think like Google's

771
00:39:34,719 --> 00:39:36,719
Speaker 2: going to die ornything. It's going to hit tens of

772
00:39:36,719 --> 00:39:39,240
Speaker 2: thousands of people's jobs, It's going to hit the markets hard.

773
00:39:39,880 --> 00:39:41,719
Speaker 2: I don't know if a recession is possible. I'm not

774
00:39:41,760 --> 00:39:45,399
Speaker 2: a financial guy, but I'll tell you this, the tech

775
00:39:45,400 --> 00:39:49,440
Speaker 2: industry needs. They need to realize how bad this is.

776
00:39:49,480 --> 00:39:52,400
Speaker 2: They need consumers to tell them, and honestly, consumers not

777
00:39:52,480 --> 00:39:54,759
Speaker 2: using it should have told them. But they clearly haven't

778
00:39:54,760 --> 00:39:57,360
Speaker 2: worked that shit out, have they. But I'll leave you

779
00:39:57,400 --> 00:40:01,600
Speaker 2: with this. I think when this there's going to be

780
00:40:01,600 --> 00:40:05,359
Speaker 2: a lot of people that need to apologize. A lot

781
00:40:05,400 --> 00:40:07,920
Speaker 2: of people in the media, the Casey Newtons of the world,

782
00:40:08,120 --> 00:40:09,880
Speaker 2: the Caros Wishes of the world, the people claim in

783
00:40:09,920 --> 00:40:14,279
Speaker 2: this the early days of the internet again boo. I

784
00:40:14,320 --> 00:40:16,359
Speaker 2: think the thing we really need to do is make

785
00:40:16,400 --> 00:40:19,480
Speaker 2: our dispersure known to these tech companies. Find the feedback forms,

786
00:40:20,360 --> 00:40:22,839
Speaker 2: go online and talk about them. Say their names. Say

787
00:40:22,840 --> 00:40:25,399
Speaker 2: their names to the friends you have who are mad

788
00:40:25,440 --> 00:40:27,239
Speaker 2: at this too, which should be a lot of them.

789
00:40:27,680 --> 00:40:29,960
Speaker 2: Say their names again and again and again. You ask me,

790
00:40:30,280 --> 00:40:33,960
Speaker 2: how can things change? These people only have their names.

791
00:40:34,000 --> 00:40:35,840
Speaker 2: They have more money than any of us can mind,

792
00:40:36,080 --> 00:40:38,239
Speaker 2: they have all of this power, and as a result,

793
00:40:38,280 --> 00:40:41,000
Speaker 2: they actually only have their seo their names. How do

794
00:40:41,040 --> 00:40:43,920
Speaker 2: you think propagar Ragavan feels, by the way, when you

795
00:40:43,960 --> 00:40:46,640
Speaker 2: google his name and all you see is me, the

796
00:40:46,719 --> 00:40:50,000
Speaker 2: smiling man. I believe that enough of us just talking

797
00:40:50,040 --> 00:40:52,440
Speaker 2: shit on these people. And I don't mean being irate.

798
00:40:52,520 --> 00:40:55,080
Speaker 2: I don't mean making stuff up. We're not them. We

799
00:40:55,120 --> 00:40:57,840
Speaker 2: don't have to operate with lies. We can operate with truths,

800
00:40:58,239 --> 00:41:02,280
Speaker 2: such as Microsoft is propping up AI a venture welfare client.

801
00:41:02,840 --> 00:41:05,040
Speaker 2: All these people hate the poor so much, but they

802
00:41:05,080 --> 00:41:07,680
Speaker 2: love welfare when it means maybe making money or maybe

803
00:41:07,680 --> 00:41:10,640
Speaker 2: burning billions of dollars. So fuck their name's up. Talk

804
00:41:10,680 --> 00:41:12,359
Speaker 2: shit about them, Talk them to your friends. I don't

805
00:41:12,360 --> 00:41:14,600
Speaker 2: care if you mentioned better offline. Just tell people what

806
00:41:14,600 --> 00:41:17,520
Speaker 2: they've done. Tell people that Generative AI loses money on

807
00:41:17,600 --> 00:41:22,040
Speaker 2: every single fucking move. It's so sickening, it's so boring,

808
00:41:22,200 --> 00:41:25,360
Speaker 2: and it's most decidedly not the future. You are not

809
00:41:25,480 --> 00:41:28,719
Speaker 2: crazy if you don't think this is amazing. You're being

810
00:41:28,800 --> 00:41:32,520
Speaker 2: lied to. What you're seeing is a marketing campaign. And

811
00:41:32,600 --> 00:41:35,080
Speaker 2: I will continue to walk you through this stuff, and

812
00:41:35,160 --> 00:41:38,239
Speaker 2: if I'm wrong somehow, I will correct myself every goddamn time,

813
00:41:38,320 --> 00:41:41,239
Speaker 2: because I deeply care about giving you what I know

814
00:41:41,560 --> 00:41:44,080
Speaker 2: and what I've seen, and the things I've reported in

815
00:41:44,120 --> 00:41:46,960
Speaker 2: these two episodes scared the shit out of me, both

816
00:41:47,000 --> 00:41:49,279
Speaker 2: for the markets and the fact that so many rich

817
00:41:49,320 --> 00:41:52,840
Speaker 2: people are willing to be so goddamn stupid, and it

818
00:41:52,920 --> 00:41:55,760
Speaker 2: made me really angry because it shows such contempt, contempt

819
00:41:55,800 --> 00:41:58,239
Speaker 2: for the media and contempt for the user, and I

820
00:41:58,320 --> 00:42:01,840
Speaker 2: will not bloody accept it. It's such a pleasure to

821
00:42:01,880 --> 00:42:12,560
Speaker 2: do this show. I'm very lucky to do so. Thank

822
00:42:12,560 --> 00:42:15,280
Speaker 2: you for listening to Better Offline. The editor and composer

823
00:42:15,320 --> 00:42:18,160
Speaker 2: of the Better Offline theme song is Mattersowski. You can

824
00:42:18,200 --> 00:42:20,520
Speaker 2: check out more of his music and audio projects at

825
00:42:20,600 --> 00:42:24,080
Speaker 2: Mattasowski dot com M A T T O S O

826
00:42:24,760 --> 00:42:28,239
Speaker 2: W s KI dot com. You can email me at

827
00:42:28,239 --> 00:42:31,080
Speaker 2: easy at Better offline dot com, or visit Better Offline

828
00:42:31,080 --> 00:42:33,160
Speaker 2: dot com to find more podcast links and of course,

829
00:42:33,200 --> 00:42:36,319
Speaker 2: my newsletter. I also really recommend you go to chat

830
00:42:36,360 --> 00:42:39,000
Speaker 2: dot Where's youreed dot at to visit the discord, and

831
00:42:39,040 --> 00:42:41,720
Speaker 2: go to our slash Better Offline to check out our reddit.

832
00:42:42,520 --> 00:42:43,880
Speaker 2: Thank you so much for listening.

833
00:42:44,719 --> 00:42:47,400
Speaker 1: Better Offline is a production of cool Zone Media. For

834
00:42:47,520 --> 00:42:50,719
Speaker 1: more from cool Zone Media, visit our website cool Zonemedia

835
00:42:50,760 --> 00:42:53,600
Speaker 1: dot com, or check us out on the iHeartRadio app,

836
00:42:53,640 --> 00:43:17,440
Speaker 1: Apple Podcasts, or wherever you get your podcasts.