1
00:00:02,720 --> 00:00:13,960
Speaker 1: Bloomberg Audio Studios, Podcasts, radio News.

2
00:00:18,480 --> 00:00:21,280
Speaker 2: Hello and welcome to another episode of The Odd Lots Podcast.

3
00:00:21,400 --> 00:00:22,760
Speaker 3: I'm Tracy Alloway.

4
00:00:22,480 --> 00:00:23,599
Speaker 4: And I'm Joe Wisenthal.

5
00:00:23,920 --> 00:00:25,320
Speaker 3: Joe, we're in the media business.

6
00:00:25,320 --> 00:00:26,360
Speaker 4: That's right, that's right.

7
00:00:26,520 --> 00:00:31,040
Speaker 2: Have you ever had an article go viral unexpectedly viral?

8
00:00:31,720 --> 00:00:31,960
Speaker 5: Yeah?

9
00:00:31,960 --> 00:00:35,680
Speaker 4: I can't like trying to like remember like specifics, but yes,

10
00:00:35,800 --> 00:00:39,599
Speaker 4: And it's one of those things typically where you're like

11
00:00:39,680 --> 00:00:41,760
Speaker 4: really excited, like a lot of people are, you know,

12
00:00:41,800 --> 00:00:43,680
Speaker 4: there's getting a lot of traction. Cool people are talking

13
00:00:43,680 --> 00:00:46,040
Speaker 4: about this, and then it goes like multiple orders of

14
00:00:46,080 --> 00:00:48,400
Speaker 4: magnitude bigger, and you're like, oh, this is like super

15
00:00:48,440 --> 00:00:51,680
Speaker 4: weird and no context for what this is, and you're

16
00:00:51,760 --> 00:00:53,360
Speaker 4: like sort of want to hide in your home and

17
00:00:53,400 --> 00:00:56,080
Speaker 4: like close the laptop because then you sort of like

18
00:00:56,080 --> 00:00:57,640
Speaker 4: make it all go away and stuff like that.

19
00:00:57,920 --> 00:01:00,480
Speaker 2: Yeah, it's kind of like once you release it into

20
00:01:00,520 --> 00:01:02,639
Speaker 2: the world, you don't actually have a lot of control

21
00:01:02,800 --> 00:01:04,760
Speaker 2: over how people use it. And I think back to

22
00:01:05,080 --> 00:01:08,280
Speaker 2: I wrote a piece about some investors trying to revive

23
00:01:08,400 --> 00:01:14,120
Speaker 2: claims on Chinese imperial bonds, like antique Chinese imperial debt

24
00:01:14,200 --> 00:01:18,920
Speaker 2: from the early nineteen hundreds and somehow this went absolutely

25
00:01:19,000 --> 00:01:22,080
Speaker 2: viral in Hong Kong at the time of the pro

26
00:01:22,120 --> 00:01:25,440
Speaker 2: democracy protest. So I would walk down the street and

27
00:01:25,480 --> 00:01:28,759
Speaker 2: I would see these homemade banners that people had created

28
00:01:28,880 --> 00:01:33,480
Speaker 2: saying that China owes the US like twenty billion in

29
00:01:33,560 --> 00:01:36,080
Speaker 2: payments on old debts, And it was just so real,

30
00:01:37,120 --> 00:01:40,920
Speaker 2: absolutely surreal, and like completely unexpected, because you wouldn't think

31
00:01:41,240 --> 00:01:44,039
Speaker 2: that some like intricate debt story was suddenly going to

32
00:01:44,040 --> 00:01:46,480
Speaker 2: become a pro democracy protest slogan.

33
00:01:46,959 --> 00:01:49,160
Speaker 3: But the world works in mysterious ways.

34
00:01:49,480 --> 00:01:51,840
Speaker 2: And speaking of the world working in mysterious ways, there

35
00:01:51,880 --> 00:01:55,240
Speaker 2: is something that went viral this week. We are recording

36
00:01:55,360 --> 00:01:58,880
Speaker 2: on February twenty seventh, and if you haven't heard of

37
00:01:58,880 --> 00:02:02,720
Speaker 2: this particular thing, you have probably been living under the

38
00:02:02,800 --> 00:02:03,680
Speaker 2: proverbial rock.

39
00:02:04,160 --> 00:02:08,520
Speaker 4: Right So past Oddlot's guest James Van Galen, a co

40
00:02:08,600 --> 00:02:12,960
Speaker 4: author to Peace on his sub stack Treny Research, talk

41
00:02:13,000 --> 00:02:15,720
Speaker 4: about a potential AI doom scenario, which a lot of

42
00:02:15,800 --> 00:02:18,160
Speaker 4: people talk about, and there's been a lot of talk

43
00:02:18,160 --> 00:02:22,320
Speaker 4: about mass white color displacement as a possible thing that

44
00:02:22,360 --> 00:02:25,120
Speaker 4: could happen as AI gets adopted, et cetera. But you know,

45
00:02:25,520 --> 00:02:28,399
Speaker 4: we know that the market's been very skittish about this specifically,

46
00:02:28,440 --> 00:02:30,960
Speaker 4: we've been seeing the software stock sell off all year,

47
00:02:30,960 --> 00:02:33,560
Speaker 4: which we've talked about plenty on the podcast, and some

48
00:02:33,600 --> 00:02:36,640
Speaker 4: of the private insurers and all this, and something about

49
00:02:36,639 --> 00:02:40,440
Speaker 4: this moment and this particular piece. I think it came

50
00:02:40,480 --> 00:02:43,640
Speaker 4: out on Sunday. Last Sunday landed with a sort of

51
00:02:43,680 --> 00:02:46,800
Speaker 4: like unbelievable thud, and so it evidently started moving markets

52
00:02:46,840 --> 00:02:49,560
Speaker 4: on Monday and then throughout the week. And this is

53
00:02:49,560 --> 00:02:52,120
Speaker 4: the part that really flabbergassed me, was you see like

54
00:02:52,240 --> 00:02:57,519
Speaker 4: all these banks and every economists, etcetera, like weighing in

55
00:02:57,639 --> 00:03:01,119
Speaker 4: and many of the very critical and like securities which

56
00:03:01,120 --> 00:03:03,560
Speaker 4: I didn't even know they like publish stuff because that's

57
00:03:03,560 --> 00:03:06,160
Speaker 4: just a market maker, like they put out all this stuff.

58
00:03:06,200 --> 00:03:08,760
Speaker 4: So I responding to it or trying to take it out.

59
00:03:08,800 --> 00:03:11,520
Speaker 4: It was a as a market story and a media story,

60
00:03:11,840 --> 00:03:12,519
Speaker 4: a wild week.

61
00:03:12,919 --> 00:03:17,200
Speaker 2: It has become the discourse dujore. There's actually a prediction.

62
00:03:16,880 --> 00:03:19,120
Speaker 3: Market on it, which you were telling me about a

63
00:03:19,120 --> 00:03:19,880
Speaker 3: few minutes ago.

64
00:03:19,960 --> 00:03:24,080
Speaker 2: Like this thing has just become much bigger than the

65
00:03:24,120 --> 00:03:27,440
Speaker 2: initial substack, which to me again says much more about

66
00:03:27,440 --> 00:03:30,360
Speaker 2: the nervousness of the market and how little anyone actually

67
00:03:30,400 --> 00:03:33,080
Speaker 2: knows about how AI is going to unfold at the

68
00:03:33,080 --> 00:03:35,440
Speaker 2: moment that people are so keen to just like latch

69
00:03:35,560 --> 00:03:37,480
Speaker 2: onto any scenario that comes out.

70
00:03:37,600 --> 00:03:39,960
Speaker 4: I get these notes from like cell side or research

71
00:03:39,960 --> 00:03:41,960
Speaker 4: shops and they're like, client has been asking us about

72
00:03:41,960 --> 00:03:45,120
Speaker 4: the Satrini scenario, and it's just like, wow, this is wild,

73
00:03:45,280 --> 00:03:46,320
Speaker 4: Like it's really like.

74
00:03:46,320 --> 00:03:46,640
Speaker 3: A s right.

75
00:03:46,680 --> 00:03:49,160
Speaker 2: People calling up Capital Economics being like I manage a

76
00:03:49,200 --> 00:03:51,520
Speaker 2: portfolio of one hundred billion and I am concerned about

77
00:03:51,520 --> 00:03:54,640
Speaker 2: a substack. Okay, well we should talk to the author

78
00:03:54,760 --> 00:03:56,640
Speaker 2: of the substack, And as you said, we've had them

79
00:03:56,640 --> 00:03:59,760
Speaker 2: on a number of times before, often talking about AI.

80
00:04:00,040 --> 00:04:02,520
Speaker 2: It is, of course James fan Kielan, the founder of

81
00:04:02,560 --> 00:04:05,680
Speaker 2: Satrini Research. So James, thanks so much for coming back

82
00:04:05,720 --> 00:04:06,280
Speaker 2: on the podcast.

83
00:04:06,360 --> 00:04:07,080
Speaker 5: Thanks for having me.

84
00:04:07,520 --> 00:04:11,760
Speaker 2: Why don't we start with what Satrini Research actually is

85
00:04:12,640 --> 00:04:14,600
Speaker 2: and what it is that you actually do in some

86
00:04:14,640 --> 00:04:18,040
Speaker 2: of your other enterprises, because I think this has become

87
00:04:18,120 --> 00:04:21,479
Speaker 2: also a source of confusion or at least interest for

88
00:04:21,880 --> 00:04:22,840
Speaker 2: people who are reading this.

89
00:04:23,360 --> 00:04:27,599
Speaker 5: Satrini Research is a pure investment research firm. We focus

90
00:04:27,680 --> 00:04:32,560
Speaker 5: primarily on thematic equity and macro research. The progression of

91
00:04:32,600 --> 00:04:35,760
Speaker 5: it was. I started it as a newsletter just speaking

92
00:04:35,839 --> 00:04:40,719
Speaker 5: about stocks and bonds and whatever else. And as we

93
00:04:40,839 --> 00:04:43,040
Speaker 5: had a kind of string of good calls, which you

94
00:04:43,120 --> 00:04:45,560
Speaker 5: were kind enough to have us on with the GLP

95
00:04:45,680 --> 00:04:47,760
Speaker 5: one early July twenty twenty three, I think.

96
00:04:47,720 --> 00:04:48,600
Speaker 3: Yeah, that was a great call.

97
00:04:48,720 --> 00:04:51,400
Speaker 5: Yeah, And the first piece we ever published was a

98
00:04:51,400 --> 00:04:54,200
Speaker 5: piece that was very bullish on the AI infrastructure complex.

99
00:04:54,600 --> 00:04:57,680
Speaker 5: So that's been an area that AI robotics has been

100
00:04:57,680 --> 00:04:59,880
Speaker 5: a big area for us. In terms of thematic equity.

101
00:05:00,440 --> 00:05:03,560
Speaker 5: We've kind of covered this winding road of bottlenecks in

102
00:05:03,680 --> 00:05:07,800
Speaker 5: terms of optics, memory power, whatever else you can possibly

103
00:05:08,160 --> 00:05:11,520
Speaker 5: allude to. We've probably covered from a what stories are

104
00:05:11,520 --> 00:05:14,400
Speaker 5: people telling about the movements that are going on in stocks.

105
00:05:14,440 --> 00:05:15,920
Speaker 5: I remember the last time that I was on odd

106
00:05:15,920 --> 00:05:19,640
Speaker 5: lots it was about this massive Stargate data center build out. Yeah,

107
00:05:19,680 --> 00:05:22,359
Speaker 5: and Joe was very surprised to see that Caterpillar was

108
00:05:22,720 --> 00:05:24,880
Speaker 5: and I think very happy that the old economy was

109
00:05:24,880 --> 00:05:25,240
Speaker 5: getting a.

110
00:05:26,880 --> 00:05:28,920
Speaker 3: He's an old economy standards.

111
00:05:28,920 --> 00:05:31,440
Speaker 5: And really that's what we've been doing for the past

112
00:05:31,440 --> 00:05:34,560
Speaker 5: three years. I've built out the team, and this piece

113
00:05:34,680 --> 00:05:38,599
Speaker 5: very much was just a response to what the market

114
00:05:38,600 --> 00:05:41,599
Speaker 5: has done here today, which is bonds of ra allied

115
00:05:42,080 --> 00:05:44,839
Speaker 5: software companies have gotten sold off, a lot of fintech

116
00:05:44,839 --> 00:05:48,039
Speaker 5: companies have gotten sold off, Private equity has sold off,

117
00:05:48,360 --> 00:05:51,600
Speaker 5: and we're always kind of looking for the cohesive narrative

118
00:05:51,720 --> 00:05:55,600
Speaker 5: that can connect disparate market moves, and the pieces co

119
00:05:55,680 --> 00:05:58,880
Speaker 5: author all up posed to me a question which was

120
00:05:59,360 --> 00:06:02,599
Speaker 5: we've been folkocused on the bullishness surrounding AI infrastructure for

121
00:06:02,640 --> 00:06:06,360
Speaker 5: a while and it's translated into this capability curve that

122
00:06:06,680 --> 00:06:09,680
Speaker 5: is moving a lot faster than anyone could expect. If

123
00:06:09,720 --> 00:06:14,800
Speaker 5: you imagine this exponential anologulorhythmic chart, it's just a diagonal line.

124
00:06:14,800 --> 00:06:16,920
Speaker 5: It goes up into the right. People have been trying

125
00:06:16,960 --> 00:06:19,640
Speaker 5: to put stigmoids or kind of level that curve off

126
00:06:19,680 --> 00:06:22,479
Speaker 5: for a long time and it hasn't. So we basically

127
00:06:22,600 --> 00:06:25,919
Speaker 5: drew that line out and said, what could be the

128
00:06:25,920 --> 00:06:29,640
Speaker 5: implications of this happening. It's a scenario which we would

129
00:06:29,680 --> 00:06:34,800
Speaker 5: ascribe maybe ten to fifteen percent towards and it comes

130
00:06:34,800 --> 00:06:37,560
Speaker 5: from a place of that. Everybody talks about equity markets

131
00:06:37,560 --> 00:06:40,080
Speaker 5: being forward looking, but really a lot more of what

132
00:06:40,120 --> 00:06:45,000
Speaker 5: you see as people justifying historical moves with new narratives

133
00:06:45,000 --> 00:06:47,800
Speaker 5: that they come up with afterwards. Very little of it

134
00:06:47,839 --> 00:06:51,080
Speaker 5: is driven by let me think of potential future outcomes

135
00:06:51,920 --> 00:06:55,120
Speaker 5: as an investor, which was the audience that this was

136
00:06:55,160 --> 00:06:57,760
Speaker 5: meant to go out to. I feel a lot more

137
00:06:57,839 --> 00:07:00,919
Speaker 5: comfortable when I can envision the bulk, the barecase, the

138
00:07:00,960 --> 00:07:04,159
Speaker 5: base case, and the most uncomfortable that you can be

139
00:07:04,160 --> 00:07:07,040
Speaker 5: as investors when you can't see the barecase at all.

140
00:07:07,200 --> 00:07:10,240
Speaker 5: So every time that we get into a market that's

141
00:07:10,240 --> 00:07:12,840
Speaker 5: similar to this, people start asking what if this time

142
00:07:12,920 --> 00:07:16,320
Speaker 5: is different? And I guess the thing that this piece

143
00:07:16,400 --> 00:07:19,160
Speaker 5: did differently was it asked what if this time is different?

144
00:07:19,200 --> 00:07:22,560
Speaker 5: But not so much in a Skehiynex and Micron are

145
00:07:22,560 --> 00:07:25,040
Speaker 5: going from price to book to price to earnings, but

146
00:07:25,200 --> 00:07:27,560
Speaker 5: in a way where what if this time is different?

147
00:07:27,600 --> 00:07:31,120
Speaker 5: Where the period of transition has to respond to a

148
00:07:31,280 --> 00:07:37,880
Speaker 5: very very fast, accelerating capability curve, and you start from

149
00:07:37,920 --> 00:07:41,960
Speaker 5: a place where there's a strong kind of historical precedent

150
00:07:42,440 --> 00:07:44,960
Speaker 5: for the past century or or two centuries. Every time

151
00:07:44,960 --> 00:07:48,160
Speaker 5: you've had a technological revolution, it's been great, it's been awesome.

152
00:07:48,480 --> 00:07:51,320
Speaker 5: And you see that when you go from ninety five

153
00:07:51,320 --> 00:07:54,040
Speaker 5: percent of the population working in agriculture to five percent

154
00:07:54,080 --> 00:07:56,480
Speaker 5: of the population and you create all these amazing jobs,

155
00:07:56,840 --> 00:08:00,120
Speaker 5: but it happens over a period of fifty years. Now

156
00:08:00,480 --> 00:08:03,160
Speaker 5: we have this capability curve where you go from two

157
00:08:03,200 --> 00:08:05,760
Speaker 5: minutes agents are capable of two minutes of autonomy on

158
00:08:05,880 --> 00:08:08,960
Speaker 5: intellectually complex tasks, and now depending on who you ask,

159
00:08:08,960 --> 00:08:11,920
Speaker 5: it's eight to sixteen hours. And that's happened in two years.

160
00:08:12,160 --> 00:08:15,000
Speaker 5: That is an exponential curve. What happens when we get

161
00:08:15,040 --> 00:08:17,360
Speaker 5: to multi day. You know what happens when we get

162
00:08:17,360 --> 00:08:20,080
Speaker 5: to multi week. And really the core of this is

163
00:08:20,160 --> 00:08:23,640
Speaker 5: if this capability curve continues being as fast an expansion

164
00:08:23,640 --> 00:08:26,640
Speaker 5: as it is, what does the world look like. There

165
00:08:26,680 --> 00:08:28,840
Speaker 5: are a lot of very good reasons why that capability

166
00:08:28,840 --> 00:08:31,120
Speaker 5: curve could level off, but that is the core of

167
00:08:31,160 --> 00:08:31,640
Speaker 5: the argument.

168
00:08:31,800 --> 00:08:34,040
Speaker 4: I do think that's just like an important sort of

169
00:08:34,400 --> 00:08:38,400
Speaker 4: level set for people here, which is that the progress

170
00:08:38,480 --> 00:08:42,320
Speaker 4: that we've seen since chet GPT came out whenever that

171
00:08:42,559 --> 00:08:45,960
Speaker 4: was late twenty twenty two, has exceeded all of the

172
00:08:46,000 --> 00:08:49,560
Speaker 4: expectations of everyone who's working on it at the time,

173
00:08:49,640 --> 00:08:52,320
Speaker 4: including the people who are in the space and the

174
00:08:52,360 --> 00:08:54,840
Speaker 4: most bullish and like the true believers. And there are

175
00:08:54,920 --> 00:08:57,560
Speaker 4: various like measures and stuff, but you know, you mentioned

176
00:08:57,600 --> 00:09:00,000
Speaker 4: the length of time you know that it could replic

177
00:09:00,320 --> 00:09:03,560
Speaker 4: a human focused on stuff, like all the people like

178
00:09:03,600 --> 00:09:06,040
Speaker 4: they made like these bets, right, and there were even

179
00:09:06,080 --> 00:09:09,719
Speaker 4: prediction markets on their capabilities, and so like, as you say, like,

180
00:09:09,880 --> 00:09:13,199
Speaker 4: it seems very plausible that the gains will level out

181
00:09:13,280 --> 00:09:18,200
Speaker 4: in some way, or that perhaps simple computer tasks don't

182
00:09:18,240 --> 00:09:21,160
Speaker 4: actually replace a lot of white color work because there's

183
00:09:21,200 --> 00:09:22,920
Speaker 4: more to white color work than what could be done

184
00:09:22,960 --> 00:09:25,920
Speaker 4: on a computer, including personality and all kinds of stuff.

185
00:09:26,440 --> 00:09:28,600
Speaker 4: All of that seems very plausible and I probably even

186
00:09:28,720 --> 00:09:30,920
Speaker 4: buy some of that. But this point that you make,

187
00:09:31,080 --> 00:09:34,920
Speaker 4: it's like, yeah, sure, but it is still proving very fast.

188
00:09:34,600 --> 00:09:39,560
Speaker 5: And it's something where the overall trend of the cost

189
00:09:39,679 --> 00:09:44,640
Speaker 5: of inference per cognitive task has gone down so significantly,

190
00:09:44,800 --> 00:09:48,360
Speaker 5: maybe depending on the forecast, ten to thirty times over

191
00:09:48,400 --> 00:09:52,120
Speaker 5: the past year, and a task that was uneconomical in

192
00:09:52,320 --> 00:09:55,280
Speaker 5: the first quarter twenty six might cross that threshold in

193
00:09:55,320 --> 00:09:58,880
Speaker 5: the third quarter. And the other interesting thing is this

194
00:09:59,000 --> 00:10:03,000
Speaker 5: capability gap where AI is capable of a lot of

195
00:10:03,040 --> 00:10:05,719
Speaker 5: things and a lot of people don't know that it's

196
00:10:05,720 --> 00:10:08,160
Speaker 5: capable of that, right, So is it about the capability

197
00:10:08,160 --> 00:10:10,360
Speaker 5: improving or is it about people becoming more familiar with

198
00:10:10,400 --> 00:10:14,960
Speaker 5: that and as AI infrastructure. It's been a great trade

199
00:10:15,040 --> 00:10:16,920
Speaker 5: and it continues to stay tight, and I think the

200
00:10:16,960 --> 00:10:20,600
Speaker 5: best rebuttal to this piece has been well, I think

201
00:10:21,280 --> 00:10:23,800
Speaker 5: Gavin Baker made this point, which is the world is

202
00:10:23,840 --> 00:10:27,600
Speaker 5: short on Watson waivers, and that's true, absolutely true. But

203
00:10:28,080 --> 00:10:32,079
Speaker 5: technological revolutions are volatile, right, Improvements come from places that

204
00:10:32,120 --> 00:10:35,320
Speaker 5: you don't really expect them to and I think you

205
00:10:35,360 --> 00:10:38,480
Speaker 5: can't fully underwrite the idea that there aren't algorithmic improvements

206
00:10:38,559 --> 00:10:42,520
Speaker 5: or there aren't improvements to the compute infrastructure. So we

207
00:10:42,559 --> 00:10:45,840
Speaker 5: should look at Okay, if this capability curve continues improving,

208
00:10:46,480 --> 00:10:50,400
Speaker 5: what are the downstream impacts there? And has the financial

209
00:10:50,440 --> 00:10:52,760
Speaker 5: system ever been stress tested for a scenario like this,

210
00:10:53,360 --> 00:10:55,520
Speaker 5: Because even if it takes five years, even if it

211
00:10:55,559 --> 00:10:58,800
Speaker 5: takes seven years, eventually we will get there. And that's

212
00:10:58,880 --> 00:11:00,840
Speaker 5: not a bearish take, it's a very bullish take. I

213
00:11:00,840 --> 00:11:03,840
Speaker 5: think that there will be great opportunities that arise because

214
00:11:03,840 --> 00:11:06,720
Speaker 5: of AI, but that's not to say that there won't

215
00:11:06,760 --> 00:11:09,240
Speaker 5: be a period of transition. And the faster that it comes,

216
00:11:09,280 --> 00:11:14,000
Speaker 5: the more aggressive that transition is. And I think the

217
00:11:14,040 --> 00:11:17,080
Speaker 5: point of The piece really was to get comfortable with

218
00:11:17,640 --> 00:11:20,880
Speaker 5: what monitoring that looks like. And I'll just make the

219
00:11:20,920 --> 00:11:23,880
Speaker 5: point that the piece also starts out with an SMP

220
00:11:24,040 --> 00:11:27,480
Speaker 5: that goes to eight thousand, because AI infrastructure is a

221
00:11:27,559 --> 00:11:30,400
Speaker 5: very bullish trade that makes up a lot of the index,

222
00:11:30,440 --> 00:11:33,800
Speaker 5: and that's a very strong and very momentum having trade

223
00:11:33,880 --> 00:11:37,440
Speaker 5: right now. And it ends with the reminder that it's

224
00:11:37,440 --> 00:11:40,679
Speaker 5: still February twenty twenty six. But in the middle of it,

225
00:11:40,679 --> 00:11:44,000
Speaker 5: it says, how do we kind of get comfortable with

226
00:11:44,520 --> 00:11:49,880
Speaker 5: the non immediacy of the replacement If a company decides

227
00:11:50,120 --> 00:11:53,040
Speaker 5: whether they're doing it because AI's gotten better or because

228
00:11:53,679 --> 00:11:56,040
Speaker 5: the market likes it when they cut jobs.

229
00:11:55,920 --> 00:11:59,120
Speaker 4: What is You're already through saw with the blog last night.

230
00:11:59,000 --> 00:12:01,280
Speaker 5: And you can argue whether that's because of AI or

231
00:12:01,280 --> 00:12:05,080
Speaker 5: whether that's because of over hiring during COVID. But Caines

232
00:12:05,160 --> 00:12:07,319
Speaker 5: said that by the end of the century we'd have

233
00:12:07,440 --> 00:12:10,760
Speaker 5: a fifteen hour work week, and he was wrong, and

234
00:12:11,040 --> 00:12:12,480
Speaker 5: there's a lot of exit you have to kind of

235
00:12:12,480 --> 00:12:15,839
Speaker 5: look at why he was wrong. There are a few explanations.

236
00:12:15,920 --> 00:12:18,320
Speaker 5: David Graeber says that we just kind of created all

237
00:12:18,360 --> 00:12:20,680
Speaker 5: these bulk jobs. This is the title of the book.

238
00:12:20,679 --> 00:12:21,720
Speaker 5: I'm not cursing.

239
00:12:21,800 --> 00:12:23,360
Speaker 3: People have said worse on this podcast.

240
00:12:24,200 --> 00:12:27,200
Speaker 5: The other explanation is that human wants and desires you

241
00:12:27,200 --> 00:12:30,760
Speaker 5: can't really model for, and we will create whatever we

242
00:12:30,840 --> 00:12:34,720
Speaker 5: need to fill that. At the same time, that required

243
00:12:35,320 --> 00:12:38,520
Speaker 5: mechanisms by which humans kind of are involved in the

244
00:12:38,559 --> 00:12:41,960
Speaker 5: process of making those machines better. It's kind of not

245
00:12:42,000 --> 00:12:47,640
Speaker 5: necessarily in every scenario concurrent with the idea of a

246
00:12:47,720 --> 00:12:51,360
Speaker 5: piece of software that has the ability for a recursive improvement.

247
00:12:52,400 --> 00:12:54,959
Speaker 5: This isn't to say that tomorrow every single company in

248
00:12:55,040 --> 00:12:58,480
Speaker 5: large enterprise goes out and replaces half their workforce, but

249
00:12:59,200 --> 00:13:01,680
Speaker 5: you do have to take a holistic picture, which is

250
00:13:02,520 --> 00:13:05,240
Speaker 5: everybody in venture capital has been talking about who's going

251
00:13:05,280 --> 00:13:07,400
Speaker 5: to be the first one person unicorn because of a

252
00:13:07,440 --> 00:13:09,960
Speaker 5: gendic AI. I don't know if we're there yet. I

253
00:13:09,960 --> 00:13:12,319
Speaker 5: haven't really kept on top of that, but that does

254
00:13:12,360 --> 00:13:16,360
Speaker 5: seem like something plausible to me, and I think one

255
00:13:16,360 --> 00:13:20,439
Speaker 5: of the better lines of the Citadel Securities counter argument, yeah,

256
00:13:20,760 --> 00:13:25,199
Speaker 5: was recursive capability doesn't imply recursive adoption. That's extremely true.

257
00:13:25,679 --> 00:13:28,840
Speaker 5: The S curve framework, though, is kind of describing the

258
00:13:28,840 --> 00:13:32,240
Speaker 5: wrong variable, and it's a variable that's really important when

259
00:13:32,280 --> 00:13:35,920
Speaker 5: you don't just have incumbents adopting, but you have startups

260
00:13:35,920 --> 00:13:39,760
Speaker 5: threatening and that variable is not necessarily breadth of adoption.

261
00:13:39,920 --> 00:13:43,720
Speaker 5: It's intensity of adoption and capability of adoption. So you

262
00:13:43,840 --> 00:13:48,400
Speaker 5: might have a flattening out S curve, and the seats

263
00:13:48,400 --> 00:13:50,840
Speaker 5: that you've already enabled with these AI tools are just

264
00:13:50,920 --> 00:13:55,480
Speaker 5: constantly getting better, and so that is you know. The

265
00:13:55,760 --> 00:13:57,920
Speaker 5: other thing is the S curve is very kind of

266
00:13:58,880 --> 00:14:02,199
Speaker 5: related to consumer adoption of new technologies. And what I

267
00:14:02,240 --> 00:14:04,920
Speaker 5: would ask is was there an S curve for the

268
00:14:05,000 --> 00:14:09,880
Speaker 5: adoption of spell check? Everybody already had a PC, everybody

269
00:14:09,920 --> 00:14:13,160
Speaker 5: already had, you know, word processing software. It was kind

270
00:14:13,160 --> 00:14:14,839
Speaker 5: of added as a feature. There are a lot of

271
00:14:14,920 --> 00:14:17,280
Speaker 5: people in the world today that have no clue how

272
00:14:17,280 --> 00:14:20,640
Speaker 5: to use SHATGPT, that are using AI every single day.

273
00:14:20,720 --> 00:14:23,360
Speaker 5: It's probably what is going to recommend you this podcast.

274
00:14:23,640 --> 00:14:27,280
Speaker 5: It's probably what is making these decisions of what items

275
00:14:27,280 --> 00:14:30,400
Speaker 5: you see when you go on Amazon. So if these

276
00:14:30,520 --> 00:14:33,840
Speaker 5: gentic capabilities are introduced as features to a technology that

277
00:14:33,880 --> 00:14:37,280
Speaker 5: everyone has already adopted, you have to adjust your model

278
00:14:37,280 --> 00:14:37,480
Speaker 5: for that.

279
00:14:53,560 --> 00:14:55,320
Speaker 2: I have so many things to say about this, but

280
00:14:55,440 --> 00:14:58,560
Speaker 2: first of all, there's something very dystopian about living in

281
00:14:58,600 --> 00:15:01,880
Speaker 2: a world where like the upside is, well, we have

282
00:15:01,920 --> 00:15:04,480
Speaker 2: a lot of bulk jobs in existence already, and so

283
00:15:04,760 --> 00:15:07,520
Speaker 2: maybe some of those bold jobs will continue to exist

284
00:15:07,560 --> 00:15:10,400
Speaker 2: even with AI. But the other thing is, like the

285
00:15:10,520 --> 00:15:14,960
Speaker 2: self reinforcing nature of AI seems really important to me

286
00:15:15,120 --> 00:15:18,240
Speaker 2: in the sense that, as you pointed out, James, like,

287
00:15:18,280 --> 00:15:20,560
Speaker 2: it's not necessarily that people have to go out and

288
00:15:20,680 --> 00:15:25,040
Speaker 2: find these new capabilities themselves. It's that the technology itself

289
00:15:25,080 --> 00:15:28,320
Speaker 2: that they're may be already using just to substitute search

290
00:15:28,440 --> 00:15:32,040
Speaker 2: or something like that, can do it on their behalf.

291
00:15:32,120 --> 00:15:34,480
Speaker 2: And so you just get this feedback cycle where like

292
00:15:35,200 --> 00:15:38,760
Speaker 2: one AI thing creates new AI things, and it just

293
00:15:38,880 --> 00:15:40,440
Speaker 2: builds and builds on itself.

294
00:15:40,720 --> 00:15:42,800
Speaker 5: I really would be remiss if I didn't say this again,

295
00:15:42,920 --> 00:15:44,840
Speaker 5: which is a lesson that I've learned over the past

296
00:15:45,160 --> 00:15:48,000
Speaker 5: five days, that you can put something in all caps,

297
00:15:48,080 --> 00:15:50,200
Speaker 5: you can bould it, and people will still not read it.

298
00:15:50,400 --> 00:15:53,440
Speaker 5: But maybe this is different because I'm speaking my base

299
00:15:53,520 --> 00:15:55,480
Speaker 5: case is probably a lot closer to a lot of

300
00:15:55,520 --> 00:16:00,200
Speaker 5: the people rebutting this article than the article itself. The

301
00:16:00,240 --> 00:16:04,560
Speaker 5: point of this really was to explore what the bear

302
00:16:04,680 --> 00:16:07,640
Speaker 5: case is if we continue to have a very bullish

303
00:16:07,960 --> 00:16:11,560
Speaker 5: world in AI infrastructure. I think that any investor that

304
00:16:11,640 --> 00:16:15,200
Speaker 5: reads it and thinks, and you know, disagrees with half

305
00:16:15,240 --> 00:16:17,680
Speaker 5: of the things that we say, maybe agrees with half

306
00:16:17,720 --> 00:16:21,320
Speaker 5: of it and forms a more nuanced understanding of what

307
00:16:21,440 --> 00:16:23,960
Speaker 5: to watch out for. That's kind of our job.

308
00:16:24,240 --> 00:16:26,600
Speaker 4: So this is important, and people who haven't read the

309
00:16:26,600 --> 00:16:29,480
Speaker 4: piece should know that, Like right up front, you do

310
00:16:29,560 --> 00:16:31,440
Speaker 4: say this, You say this piece is not a forecast.

311
00:16:31,840 --> 00:16:35,040
Speaker 4: This is a possible scenario and how it could go,

312
00:16:35,320 --> 00:16:37,840
Speaker 4: and we want to get into some of the details.

313
00:16:38,240 --> 00:16:41,320
Speaker 4: But you know, one counter argument to sort of the

314
00:16:41,560 --> 00:16:45,880
Speaker 4: idea of macro economic doom or financial crisis or whatever,

315
00:16:46,080 --> 00:16:50,080
Speaker 4: is okay if you have AI and it's driving incredible

316
00:16:50,080 --> 00:16:54,480
Speaker 4: productivity gains, If it's very disinflationary and so forth, if

317
00:16:54,600 --> 00:16:57,560
Speaker 4: some people are becoming fabulously wealthy and part of this

318
00:16:57,560 --> 00:17:01,360
Speaker 4: big redistribution that would happen, well, then the government has

319
00:17:01,400 --> 00:17:04,000
Speaker 4: a lot more fiscal capacity to stabilize this. Right, then

320
00:17:04,000 --> 00:17:05,760
Speaker 4: the government can spend a lot of money. Rates have

321
00:17:05,800 --> 00:17:09,800
Speaker 4: come down, they can counteract the disinflation, not totally, unlike

322
00:17:09,840 --> 00:17:13,119
Speaker 4: perhaps COVID would be like a great example, but it

323
00:17:13,160 --> 00:17:15,320
Speaker 4: strikes me as like, well, if we're ever going to

324
00:17:15,359 --> 00:17:18,520
Speaker 4: have a government that's thinking about these things proactively, that

325
00:17:18,720 --> 00:17:21,080
Speaker 4: strikes me as a good reason to write them out.

326
00:17:21,119 --> 00:17:24,399
Speaker 4: And it's notable like many of the executives at the

327
00:17:24,400 --> 00:17:28,200
Speaker 4: top AI labs they talk about exactly this. In fact,

328
00:17:28,240 --> 00:17:31,040
Speaker 4: it seems like they're pleading almost with the government to

329
00:17:31,480 --> 00:17:34,040
Speaker 4: take this war seriously, because if we're going to have

330
00:17:34,119 --> 00:17:37,840
Speaker 4: this big disruption and redistribution, we're going to have to

331
00:17:37,880 --> 00:17:41,080
Speaker 4: start thinking about what are the fiscal mechanisms to counter

332
00:17:41,119 --> 00:17:42,040
Speaker 4: it out.

333
00:17:41,640 --> 00:17:45,520
Speaker 5: One hundred percent, I think that it's something where it's

334
00:17:45,640 --> 00:17:48,280
Speaker 5: perfectly fine and good to say that the government will

335
00:17:48,359 --> 00:17:51,879
Speaker 5: be able to deal with it, but it's probably better

336
00:17:51,960 --> 00:17:55,520
Speaker 5: to formulate a framework in which the government is more

337
00:17:55,600 --> 00:17:57,879
Speaker 5: able to do that. And in order to do that,

338
00:17:57,920 --> 00:17:59,639
Speaker 5: you kind of have to have an idea of what

339
00:17:59,680 --> 00:18:02,880
Speaker 5: to keep track of, and I can say that in

340
00:18:03,720 --> 00:18:07,720
Speaker 5: the discourse that I've seen, I don't think that there's

341
00:18:07,840 --> 00:18:12,080
Speaker 5: a very strong kind of data collection on this. Specifically,

342
00:18:12,440 --> 00:18:15,640
Speaker 5: one of the big rebuttals has been that software job

343
00:18:15,680 --> 00:18:18,640
Speaker 5: postings have gone up eleven percent year over year. Those

344
00:18:18,720 --> 00:18:22,640
Speaker 5: job postings include AI and machine learning engineers, so you're

345
00:18:22,680 --> 00:18:26,960
Speaker 5: really seeing a composition shift where these new you know,

346
00:18:27,040 --> 00:18:31,119
Speaker 5: AI engineers are coming in and they're creating software that

347
00:18:31,240 --> 00:18:35,120
Speaker 5: will improve itself. And when it comes to the government response,

348
00:18:35,480 --> 00:18:38,840
Speaker 5: Jolts doesn't really speak about composition. In my opinion, there's

349
00:18:38,840 --> 00:18:41,760
Speaker 5: not a great amount of data on white collar specifically,

350
00:18:42,200 --> 00:18:45,400
Speaker 5: And yeah, it was, it was. It was almost worrying

351
00:18:45,440 --> 00:18:49,199
Speaker 5: in itself to see this reaction where we write this

352
00:18:49,320 --> 00:18:52,160
Speaker 5: article that's kind of saying what I think most people

353
00:18:52,200 --> 00:18:55,760
Speaker 5: are thinking. We're putting trillions of dollars at the white

354
00:18:55,760 --> 00:18:58,760
Speaker 5: collar productivity machine, and oh that might, you know, have

355
00:18:58,840 --> 00:19:01,600
Speaker 5: some level of disruption, and I get it. The thing

356
00:19:01,600 --> 00:19:03,800
Speaker 5: that I'm very thankful to a lot of the rebuttals

357
00:19:03,800 --> 00:19:06,200
Speaker 5: for is that they've reminded people that it's twenty twenty six,

358
00:19:06,440 --> 00:19:08,280
Speaker 5: which we tried to do three times in the piece,

359
00:19:08,320 --> 00:19:11,760
Speaker 5: but apparently we're not successful. Thank you to everyone that

360
00:19:11,600 --> 00:19:15,560
Speaker 5: that made sure that this isn't like a spinout, crazy whatever.

361
00:19:16,119 --> 00:19:19,240
Speaker 5: But the worrying side is, well, everyone seems very very

362
00:19:19,240 --> 00:19:22,080
Speaker 5: comfortable that this is all going to be okay, and

363
00:19:22,160 --> 00:19:24,879
Speaker 5: I think that that reasonably. I'm also a student of

364
00:19:24,920 --> 00:19:28,000
Speaker 5: financial history. That reasonably comes from when you look back

365
00:19:28,040 --> 00:19:29,480
Speaker 5: at the past and you say, well, we had this

366
00:19:29,520 --> 00:19:32,919
Speaker 5: industrial revolution and it was amazing, and we've had mechanization

367
00:19:32,960 --> 00:19:34,960
Speaker 5: and it was amazing, and we've had the Internet and

368
00:19:35,000 --> 00:19:37,359
Speaker 5: it was amazing, and it created all these jobs that

369
00:19:37,359 --> 00:19:40,880
Speaker 5: we couldn't have possibly foreseen beforehand. And you're looking at

370
00:19:40,880 --> 00:19:44,600
Speaker 5: that from one hundred or more years in the future.

371
00:19:45,200 --> 00:19:48,040
Speaker 5: We have the term Luddite because of the fact that

372
00:19:48,720 --> 00:19:55,440
Speaker 5: the transition was so abrupt and marked that people were

373
00:19:55,440 --> 00:19:58,160
Speaker 5: moved to physical violence. Right, we don't want that to happen.

374
00:19:58,400 --> 00:20:01,440
Speaker 5: The transitions do occur in the faster that this happens.

375
00:20:01,760 --> 00:20:03,600
Speaker 5: If this were going to happen over the next twenty

376
00:20:03,960 --> 00:20:06,560
Speaker 5: or thirty years, fine, you know that that's going to

377
00:20:06,560 --> 00:20:09,280
Speaker 5: be great. Everything's going to be awesome. I think that

378
00:20:09,320 --> 00:20:13,080
Speaker 5: the real time frame is closer to five to fifteen,

379
00:20:13,560 --> 00:20:17,199
Speaker 5: and obviously this piece extrapolates where it's three years. We

380
00:20:17,240 --> 00:20:20,520
Speaker 5: should be prepared for anything, because the government isn't going

381
00:20:20,560 --> 00:20:24,520
Speaker 5: to accurately forecast technological advancement, but they can accurately forecast

382
00:20:24,560 --> 00:20:26,439
Speaker 5: what they should watch and what the best policy response

383
00:20:26,440 --> 00:20:26,800
Speaker 5: would be.

384
00:20:27,040 --> 00:20:27,679
Speaker 3: Yeah, this's the thing.

385
00:20:27,680 --> 00:20:30,040
Speaker 2: The Luddites were like ultimately on the wrong side of

386
00:20:30,119 --> 00:20:33,440
Speaker 2: history in terms of thinking that resistance to new technology

387
00:20:33,480 --> 00:20:35,680
Speaker 2: would actually matter. But that doesn't mean that there wasn't

388
00:20:35,960 --> 00:20:39,680
Speaker 2: major resistance and disruption on the way, that it.

389
00:20:39,520 --> 00:20:43,240
Speaker 4: Wasn't absolutely awful. Yeah, no, exactly right from their perspective,

390
00:20:43,240 --> 00:20:44,400
Speaker 4: from their lives exactly.

391
00:20:44,680 --> 00:20:48,159
Speaker 2: You know, you mentioned software job openings still rising, and

392
00:20:48,160 --> 00:20:51,280
Speaker 2: one of the reasons that's able to happen is because

393
00:20:51,320 --> 00:20:55,160
Speaker 2: we still have a financial system that up until relatively recently,

394
00:20:55,200 --> 00:20:59,879
Speaker 2: has been very comfortable with extending credit to software companies.

395
00:21:00,040 --> 00:21:04,040
Speaker 2: And there's obviously a reflexivity between the financial system, the market,

396
00:21:04,320 --> 00:21:07,080
Speaker 2: and the real economy. And you dig into that in

397
00:21:07,119 --> 00:21:09,240
Speaker 2: your piece as well, And this is the part of

398
00:21:09,280 --> 00:21:11,960
Speaker 2: it that I actually found the most interesting, where you

399
00:21:12,040 --> 00:21:16,879
Speaker 2: describe how AI could actually and the disruptive effects of

400
00:21:16,920 --> 00:21:21,200
Speaker 2: AI could actually end up becoming problematic, especially for private capital.

401
00:21:21,240 --> 00:21:23,840
Speaker 2: And this again is something that is very much in

402
00:21:23,880 --> 00:21:27,200
Speaker 2: the public slash market psyche this week because we've had

403
00:21:27,200 --> 00:21:30,919
Speaker 2: a number of private credit blowups starting to become public.

404
00:21:31,200 --> 00:21:33,719
Speaker 2: Talk a little bit more about how you see that

405
00:21:33,840 --> 00:21:38,040
Speaker 2: kind of private credit AI disruption now insurance as well,

406
00:21:38,240 --> 00:21:39,439
Speaker 2: nexus unfolding.

407
00:21:40,720 --> 00:21:45,400
Speaker 5: Just to reiterate, I don't see it, but I think

408
00:21:45,440 --> 00:21:47,720
Speaker 5: this wasn't like a singling out of private credit was

409
00:21:47,800 --> 00:21:50,880
Speaker 5: very much a response to the price action of the market,

410
00:21:51,440 --> 00:21:55,200
Speaker 5: but it is something worth considering that it's a relatively

411
00:21:55,280 --> 00:21:58,960
Speaker 5: new in the grand scheme of things, and there's a

412
00:21:59,000 --> 00:22:02,480
Speaker 5: system that's built upon the assumption that things stay relatively stable,

413
00:22:02,840 --> 00:22:06,840
Speaker 5: and if things aren't relatively stable, then what could possibly happen.

414
00:22:06,920 --> 00:22:10,120
Speaker 5: We're not really private credit analysts, right Worreth thematic equity

415
00:22:10,440 --> 00:22:14,480
Speaker 5: and macro research. This was something where we presented kind

416
00:22:14,480 --> 00:22:17,160
Speaker 5: of if you were to have a wave of defaults

417
00:22:17,240 --> 00:22:20,199
Speaker 5: in one of these disrupted industries, what would happen? And

418
00:22:20,240 --> 00:22:23,800
Speaker 5: then the other thing is maybe the job losses are fine,

419
00:22:23,960 --> 00:22:26,439
Speaker 5: and we go back to a economy like the nineteen

420
00:22:26,440 --> 00:22:28,920
Speaker 5: fifties where the participation rate is much lower but productivity

421
00:22:28,960 --> 00:22:33,880
Speaker 5: is much higher. That's great too. In the transition, the

422
00:22:33,920 --> 00:22:36,719
Speaker 5: people that are at the highest risk of being replaced

423
00:22:36,720 --> 00:22:40,400
Speaker 5: by AI have like seven eighty FICO scores and they're

424
00:22:40,440 --> 00:22:44,040
Speaker 5: not classically what gets modeled as a risk in terms

425
00:22:44,080 --> 00:22:46,840
Speaker 5: of a default. So these are all things where it's

426
00:22:46,880 --> 00:22:49,040
Speaker 5: not saying that this is going to happen. It's saying,

427
00:22:49,520 --> 00:22:51,920
Speaker 5: has a private credit lending and you know, to their

428
00:22:51,920 --> 00:22:55,960
Speaker 5: credit I will say Apollo much earlier to the software

429
00:22:56,000 --> 00:22:59,040
Speaker 5: thing than even I was or the market was right.

430
00:22:59,080 --> 00:23:02,359
Speaker 5: Apollo reduced their software lending pretty early on. I think

431
00:23:02,440 --> 00:23:06,440
Speaker 5: it was in early twenty twenty five. For the rest

432
00:23:06,480 --> 00:23:10,240
Speaker 5: of it, you know, like, has there been enough changes

433
00:23:10,240 --> 00:23:13,560
Speaker 5: to the assumptions about the income and about you know,

434
00:23:13,680 --> 00:23:16,600
Speaker 5: does arr stay recurring? That's just something to consider.

435
00:23:16,640 --> 00:23:19,040
Speaker 3: I think, what's your base case on private credit then?

436
00:23:19,160 --> 00:23:22,040
Speaker 2: Is it the sort of Jamie Diamond cockroach scenario?

437
00:23:22,520 --> 00:23:26,359
Speaker 5: So I think that private credit isn't banking right like

438
00:23:26,440 --> 00:23:29,800
Speaker 5: run on the bank dynamic doesn't necessarily play out. They

439
00:23:30,359 --> 00:23:33,720
Speaker 5: are in possession of permanent capital to a certain degree,

440
00:23:33,760 --> 00:23:37,639
Speaker 5: and that's through in a lot of areas the acquisition

441
00:23:37,680 --> 00:23:41,120
Speaker 5: of these life insurers. So I think you could definitely

442
00:23:41,160 --> 00:23:43,960
Speaker 5: see the contagion being very minimized if there were to be.

443
00:23:44,520 --> 00:23:46,320
Speaker 5: I don't think there have been any, like very high

444
00:23:46,359 --> 00:23:47,200
Speaker 5: profile blow ups.

445
00:23:47,240 --> 00:23:47,400
Speaker 2: Yet.

446
00:23:47,440 --> 00:23:50,639
Speaker 5: Everything's pretty much fine right now as I understand it.

447
00:23:50,680 --> 00:23:53,960
Speaker 5: The progression of it, though, I don't think that you're

448
00:23:53,960 --> 00:23:56,000
Speaker 5: at a very high risk. My base case would be

449
00:23:56,480 --> 00:24:00,480
Speaker 5: just like that, And the only kind of add risk

450
00:24:00,720 --> 00:24:04,080
Speaker 5: is if you were to have some sort of change

451
00:24:04,440 --> 00:24:07,639
Speaker 5: to how private credit is treated. From a regulatory perspective

452
00:24:07,960 --> 00:24:10,160
Speaker 5: on the balance sheet of these life insurans.

453
00:24:10,680 --> 00:24:14,960
Speaker 4: So there's sort of two major components to the piece

454
00:24:15,000 --> 00:24:18,520
Speaker 4: that you wrote, and one is obviously the macro scenario,

455
00:24:18,880 --> 00:24:20,960
Speaker 4: and the way it's framed it is like, okay, years

456
00:24:21,000 --> 00:24:24,920
Speaker 4: twenty twenty eight, unemployment is above ten percent, the stock

457
00:24:24,960 --> 00:24:27,320
Speaker 4: market has falling forty percent. So there's the macro story,

458
00:24:27,320 --> 00:24:30,800
Speaker 4: but then there's also the sort of secular microstory. And

459
00:24:30,840 --> 00:24:32,679
Speaker 4: I think this is really interesting, and this is the

460
00:24:32,720 --> 00:24:34,760
Speaker 4: part that I've been like trying to work out and

461
00:24:34,760 --> 00:24:37,560
Speaker 4: trying to understand better. This idea that like, there are

462
00:24:37,560 --> 00:24:41,040
Speaker 4: all these businesses that have essentially been built up around

463
00:24:41,520 --> 00:24:45,000
Speaker 4: building a mote based on network effects, you know, payments

464
00:24:45,040 --> 00:24:48,040
Speaker 4: platforms and so forth and whatever, and so this idea

465
00:24:48,080 --> 00:24:52,640
Speaker 4: that AI and agentic commerce will fundamentally change the way

466
00:24:52,680 --> 00:24:55,480
Speaker 4: a lot of these businesses operate and these motes will

467
00:24:55,480 --> 00:24:59,280
Speaker 4: disappear and talk to us about that, because I have

468
00:24:59,320 --> 00:25:03,399
Speaker 4: a harder time I'm wrapping my head around what is

469
00:25:03,480 --> 00:25:05,800
Speaker 4: it about AI per se that it's like, here you

470
00:25:05,880 --> 00:25:11,840
Speaker 4: have these legacy networks, delivery drivers, payment companies with whatever

471
00:25:12,000 --> 00:25:13,720
Speaker 4: they have on the desk, and you swipe your cards

472
00:25:13,720 --> 00:25:16,520
Speaker 4: and stuff like that. What are those called little.

473
00:25:16,400 --> 00:25:16,920
Speaker 5: Point of sale?

474
00:25:16,960 --> 00:25:19,920
Speaker 4: What there's a little point of sale machines? But talk

475
00:25:19,960 --> 00:25:22,679
Speaker 4: to us about, like, from a pure tech standpoint, what

476
00:25:22,800 --> 00:25:26,480
Speaker 4: is it about agentic AI that can sort of evaporate

477
00:25:26,520 --> 00:25:27,040
Speaker 4: this mode?

478
00:25:27,400 --> 00:25:29,880
Speaker 5: So I will say, if I had to go back

479
00:25:29,920 --> 00:25:32,280
Speaker 5: in time and write the piece differently, okay, I would

480
00:25:32,359 --> 00:25:34,280
Speaker 5: not have singled that. I would have just kept it

481
00:25:34,320 --> 00:25:38,160
Speaker 5: on a sector basis, right, And I think that if

482
00:25:38,200 --> 00:25:39,880
Speaker 5: I knew that it was going to get thirty million views,

483
00:25:39,920 --> 00:25:42,240
Speaker 5: I would not have mentioned single stocks at all. So

484
00:25:42,320 --> 00:25:46,120
Speaker 5: I won't do that here. But what I will say is,

485
00:25:47,000 --> 00:25:49,119
Speaker 5: and this future could be wrong, but if you envision

486
00:25:49,160 --> 00:25:52,280
Speaker 5: a future where I remember talking to you guys about

487
00:25:52,280 --> 00:25:55,560
Speaker 5: this in twenty twenty four when I was using it

488
00:25:55,600 --> 00:25:58,520
Speaker 5: as a bow case for Apple, which didn't end up coming.

489
00:25:58,600 --> 00:26:01,240
Speaker 5: You know, the Apple was kind of of let the

490
00:26:01,320 --> 00:26:03,080
Speaker 5: chips fall where they may and then we'll come in afterwards,

491
00:26:03,119 --> 00:26:05,359
Speaker 5: which they've done a lot in the past ten years.

492
00:26:05,400 --> 00:26:08,800
Speaker 5: But the idea is you have this agentic assistant and

493
00:26:09,000 --> 00:26:11,159
Speaker 5: it's in your phone and it knows everything about you,

494
00:26:11,560 --> 00:26:14,680
Speaker 5: and then you kind of extrapolate that to a lot

495
00:26:14,720 --> 00:26:17,320
Speaker 5: of people spend a decent amount of time shopping. What

496
00:26:17,400 --> 00:26:19,399
Speaker 5: they don't spend a lot of time doing is price matching.

497
00:26:19,640 --> 00:26:21,280
Speaker 5: If you're going to buy a box of protein bars,

498
00:26:21,320 --> 00:26:24,840
Speaker 5: you don't really check five different vendors because it's tedious.

499
00:26:24,960 --> 00:26:29,359
Speaker 5: AI agents do not experience tedium, right, So the kind

500
00:26:29,359 --> 00:26:34,119
Speaker 5: of way that there are a lot of layered intermediation

501
00:26:34,840 --> 00:26:38,560
Speaker 5: and rent kind of extraction layer in the economy, and

502
00:26:38,600 --> 00:26:41,160
Speaker 5: then there are a lot of places where having a

503
00:26:41,440 --> 00:26:45,600
Speaker 5: like an oligopoly essentially has allowed margins to really be

504
00:26:46,240 --> 00:26:50,920
Speaker 5: artificially increased. So just to address I don't think that

505
00:26:51,000 --> 00:26:53,280
Speaker 5: code is the moat on a delivery network for like

506
00:26:53,520 --> 00:26:58,200
Speaker 5: like like that's you have the drivers, you have the customers.

507
00:26:58,680 --> 00:27:02,840
Speaker 5: I get that I could see happening is something that's

508
00:27:02,880 --> 00:27:06,760
Speaker 5: already kind of happening where these startups are enabled to

509
00:27:07,320 --> 00:27:10,879
Speaker 5: create something that's similar and well, you don't have the

510
00:27:10,880 --> 00:27:13,520
Speaker 5: network effect, okay, But if you have an AI agent

511
00:27:14,040 --> 00:27:16,359
Speaker 5: that has the explicit instructions to go out and find

512
00:27:16,359 --> 00:27:20,479
Speaker 5: the cheapest option, then it doesn't really care about using

513
00:27:20,720 --> 00:27:22,840
Speaker 5: this thing that has a network effect. It cares about

514
00:27:22,920 --> 00:27:25,159
Speaker 5: using the thing that's the cheapest. So if you have

515
00:27:25,200 --> 00:27:28,240
Speaker 5: an order aggregator that's an agentic kind of aggregator on

516
00:27:28,280 --> 00:27:31,760
Speaker 5: the driver side and the customer side. Then the customer

517
00:27:31,800 --> 00:27:34,600
Speaker 5: says to the agent, hey, I want this burrito from Chipotle.

518
00:27:35,040 --> 00:27:38,080
Speaker 5: And then there's a bunch of different platforms that the

519
00:27:38,119 --> 00:27:40,359
Speaker 5: listing is on because the restaurant has used one of

520
00:27:40,359 --> 00:27:42,760
Speaker 5: these agentic aggregators to go on every single one and

521
00:27:42,800 --> 00:27:45,120
Speaker 5: put their thing, and the driver also has the one

522
00:27:45,119 --> 00:27:48,120
Speaker 5: that will get them paid the most. So the idea

523
00:27:48,200 --> 00:27:52,240
Speaker 5: of you know, taking half of the delivery fee as

524
00:27:52,480 --> 00:27:55,760
Speaker 5: the company kind of goes away because your margin is

525
00:27:55,760 --> 00:28:00,040
Speaker 5: my opportunity and someone that's five people that's kind of

526
00:28:00,080 --> 00:28:03,720
Speaker 5: cutting up this maybe shoddy replacement, is very happy to

527
00:28:04,000 --> 00:28:07,800
Speaker 5: you know. Obviously there are other modes here, but that's

528
00:28:07,920 --> 00:28:11,160
Speaker 5: just one example of how you might see a world

529
00:28:11,160 --> 00:28:13,919
Speaker 5: in which agenta commerce and the It's very similar to

530
00:28:13,960 --> 00:28:16,520
Speaker 5: like the paper clip problem. If you tell a machine

531
00:28:16,520 --> 00:28:18,600
Speaker 5: and to do something, it's just trying to get you

532
00:28:18,640 --> 00:28:21,880
Speaker 5: the best price, and maybe that includes finding way around

533
00:28:21,880 --> 00:28:22,600
Speaker 5: interchange just.

534
00:28:22,560 --> 00:28:24,800
Speaker 4: To push back of this or just a pressure. I mean,

535
00:28:24,960 --> 00:28:28,640
Speaker 4: like comparison shopping websites have existed for a long time

536
00:28:28,800 --> 00:28:32,119
Speaker 4: almost it's the beginning of the Internet, right, and you know,

537
00:28:32,560 --> 00:28:33,400
Speaker 4: you could Google.

538
00:28:33,480 --> 00:28:33,840
Speaker 1: I don't know.

539
00:28:33,880 --> 00:28:36,160
Speaker 4: It's just like Google Shop had a thing for a while.

540
00:28:36,160 --> 00:28:38,720
Speaker 4: I don't think people ever that took off, but you know,

541
00:28:38,760 --> 00:28:41,400
Speaker 4: it would show you like here's the price of a

542
00:28:41,400 --> 00:28:45,680
Speaker 4: computer monitor on Amazon and Walmart dot com and new

543
00:28:45,720 --> 00:28:48,040
Speaker 4: egg dot com and a few of these sites that

544
00:28:48,120 --> 00:28:51,880
Speaker 4: like don't exist anymore, et cetera. Like in theory, like,

545
00:28:52,000 --> 00:28:54,440
Speaker 4: isn't it describing the same thing that like from the

546
00:28:54,480 --> 00:28:57,240
Speaker 4: customer's perspective, It's like, Okay, I'll just they're all the same.

547
00:28:57,480 --> 00:28:58,960
Speaker 4: I'm going to click the cheap it totally.

548
00:28:59,000 --> 00:29:02,719
Speaker 5: I get that, And that's an entirely possible case. What

549
00:29:02,800 --> 00:29:05,480
Speaker 5: I will say is there's a big difference between actively

550
00:29:05,600 --> 00:29:09,640
Speaker 5: going and taking the effort and taking the time to

551
00:29:09,960 --> 00:29:11,960
Speaker 5: go to one of these comparison shopping sites to get

552
00:29:11,960 --> 00:29:14,719
Speaker 5: the best price versus just telling your phone, get me

553
00:29:14,720 --> 00:29:18,560
Speaker 5: a burrito, get me the best price. Right, those are

554
00:29:18,000 --> 00:29:21,640
Speaker 5: They're two kind of fundamentally different things. This will play

555
00:29:21,640 --> 00:29:24,000
Speaker 5: out over the next five or ten years, and we'll see.

556
00:29:24,000 --> 00:29:27,000
Speaker 5: And also I'm sure that we're not going to just

557
00:29:27,040 --> 00:29:30,760
Speaker 5: delete friction overnight, right, So that's why it was so

558
00:29:30,880 --> 00:29:33,920
Speaker 5: shocking to see this kind of like media reaction it's

559
00:29:33,960 --> 00:29:36,280
Speaker 5: like this stuff hasn't happened yet, and we don't know

560
00:29:36,320 --> 00:29:38,800
Speaker 5: exactly how it's going to happen. It's just a future

561
00:29:39,080 --> 00:29:40,920
Speaker 5: scenario where things happen a certain way.

562
00:29:40,960 --> 00:29:58,680
Speaker 2: So can you talk to us for a second just

563
00:29:58,760 --> 00:30:02,280
Speaker 2: where you see AI valuations at the moment, because I

564
00:30:02,280 --> 00:30:04,360
Speaker 2: think this is also part of the reason that people

565
00:30:04,480 --> 00:30:07,680
Speaker 2: are very nervous at the moment, which is like, Okay,

566
00:30:07,960 --> 00:30:09,800
Speaker 2: on the one hand, we think AI is going to

567
00:30:09,840 --> 00:30:11,720
Speaker 2: eat the world, but on the other hand, it's not

568
00:30:11,920 --> 00:30:14,280
Speaker 2: entirely clear that a lot of AI is going to

569
00:30:14,320 --> 00:30:17,200
Speaker 2: make money in doing so. And if you look at

570
00:30:17,280 --> 00:30:19,560
Speaker 2: you know, some of the big hyperscalers at the moment,

571
00:30:19,640 --> 00:30:23,280
Speaker 2: they're still losing money on certain power users.

572
00:30:23,360 --> 00:30:24,760
Speaker 3: So how do we.

573
00:30:24,760 --> 00:30:28,400
Speaker 2: Think that AI is actually going to make money as

574
00:30:28,440 --> 00:30:30,320
Speaker 2: it sort of eats the world.

575
00:30:30,640 --> 00:30:33,480
Speaker 5: I think that that's the other thing that's important here

576
00:30:33,640 --> 00:30:36,000
Speaker 5: is these companies need to go out and search for

577
00:30:36,120 --> 00:30:40,080
Speaker 5: roy and there are a lot of threats. You saw

578
00:30:40,080 --> 00:30:44,520
Speaker 5: Anthropic respond to the Chinese distillation of models, and you know,

579
00:30:44,560 --> 00:30:48,120
Speaker 5: if you go when you use Mini Max, it's relatively comparable,

580
00:30:48,560 --> 00:30:52,880
Speaker 5: but it's also ninety percent cheaper. So this is like

581
00:30:53,280 --> 00:30:56,600
Speaker 5: that there is a race happening right now, and the

582
00:30:56,680 --> 00:31:00,320
Speaker 5: economics are they span the gamut, right, the good and

583
00:31:00,360 --> 00:31:03,360
Speaker 5: bad on both sides. The thing that drives this kind

584
00:31:03,400 --> 00:31:08,200
Speaker 5: of capability improvement is you do need customers to pay

585
00:31:08,240 --> 00:31:12,040
Speaker 5: for these things that you have spent so much money on,

586
00:31:12,480 --> 00:31:15,040
Speaker 5: and that means making it capable in a way that's

587
00:31:15,120 --> 00:31:17,560
Speaker 5: useful to your customers, or integrating it in a way

588
00:31:17,560 --> 00:31:21,200
Speaker 5: that's useful to your customers. So I personally think that

589
00:31:21,200 --> 00:31:25,040
Speaker 5: that will happen. How quickly it happens is anybody's guess.

590
00:31:25,160 --> 00:31:27,560
Speaker 5: But I think valuations right now are reflective of this

591
00:31:27,680 --> 00:31:31,920
Speaker 5: expectation that we are going to continue adding compute capacity

592
00:31:31,920 --> 00:31:34,239
Speaker 5: to be able to handle this. And I think that

593
00:31:34,360 --> 00:31:36,640
Speaker 5: if you spend eight hours just thinking about it, you

594
00:31:36,680 --> 00:31:39,520
Speaker 5: can see a lot of places where AI is pretty valuable.

595
00:31:40,000 --> 00:31:42,320
Speaker 5: But a lot of those places are places where you

596
00:31:42,400 --> 00:31:45,800
Speaker 5: might otherwise pay a human right now. So yeah, it's

597
00:31:45,880 --> 00:31:48,160
Speaker 5: just you just have to balance it, and there's a

598
00:31:48,160 --> 00:31:50,200
Speaker 5: lot of ways that it can go well, and then

599
00:31:50,240 --> 00:31:52,200
Speaker 5: there's a couple of ways that it doesn't.

600
00:31:52,320 --> 00:31:55,160
Speaker 4: Let's talk about enterprise software for a second, because okay,

601
00:31:55,720 --> 00:31:58,600
Speaker 4: the public facing, these modes, these network effects, et cetera,

602
00:31:58,840 --> 00:32:01,800
Speaker 4: maybe AI agent slow us to get the best price forever.

603
00:32:02,320 --> 00:32:05,320
Speaker 4: Is it different economics if we're saying the enterprise, we

604
00:32:05,360 --> 00:32:07,640
Speaker 4: know about the enterprise, the SaaS sell off, et cetera.

605
00:32:07,920 --> 00:32:10,760
Speaker 4: What is the scenario. How would you articulate the fear

606
00:32:10,920 --> 00:32:13,680
Speaker 4: in the market right now that all of these incumbent

607
00:32:13,800 --> 00:32:18,040
Speaker 4: software companies could theoretically get ripped out because something something

608
00:32:18,160 --> 00:32:20,760
Speaker 4: AI will make it so that customers don't need them.

609
00:32:20,800 --> 00:32:23,560
Speaker 5: So you can separate software. You have kind of like

610
00:32:23,600 --> 00:32:26,800
Speaker 5: this long tail of SaaS that includes these you know,

611
00:32:26,880 --> 00:32:29,600
Speaker 5: workflow automation tools, and then you have like the systems

612
00:32:29,640 --> 00:32:32,800
Speaker 5: of record. I think that it's very likely that the

613
00:32:32,840 --> 00:32:35,240
Speaker 5: at least the systems of record have like a short

614
00:32:35,280 --> 00:32:39,000
Speaker 5: squeeze in the sense that right now they kind of

615
00:32:39,120 --> 00:32:42,200
Speaker 5: just have upside and that they are they're most situated

616
00:32:42,200 --> 00:32:44,520
Speaker 5: to be able to improve their margins because of AI.

617
00:32:44,480 --> 00:32:47,320
Speaker 4: Right, because coding is a cost for them, right, they

618
00:32:47,000 --> 00:32:49,960
Speaker 4: can theoretically maintain these things much cheaper than they is.

619
00:32:50,080 --> 00:32:52,960
Speaker 5: Yeah, one hundred percent. And what we said in the piece,

620
00:32:52,960 --> 00:32:54,920
Speaker 5: which will be you know, interesting to see in real life,

621
00:32:54,920 --> 00:32:57,840
Speaker 5: and I don't necessarily it's a good point that enterprises

622
00:32:57,880 --> 00:33:00,000
Speaker 5: don't really react as quickly as this, so the time

623
00:33:00,000 --> 00:33:03,360
Speaker 5: line is probably aggressive. But the way that these kind

624
00:33:03,360 --> 00:33:07,000
Speaker 5: of contracts are negotiated. Last year, when you had the

625
00:33:07,000 --> 00:33:10,640
Speaker 5: first half, the kind of budget resetting these CIOs and

626
00:33:10,640 --> 00:33:14,600
Speaker 5: procurement teams they agentic AI was still kind of a buzzword, right.

627
00:33:14,880 --> 00:33:18,520
Speaker 5: It wasn't until the end of November that it became insane.

628
00:33:18,560 --> 00:33:21,000
Speaker 5: You know, I saw you have vibe coded a couple

629
00:33:20,960 --> 00:33:23,360
Speaker 5: of things yourself, So there was a.

630
00:33:23,600 --> 00:33:24,560
Speaker 4: Cool coming out next week.

631
00:33:24,640 --> 00:33:28,719
Speaker 5: Nice, there was a great kind of jumping capability. What

632
00:33:28,800 --> 00:33:30,600
Speaker 5: is it? By the way, are you can to speak

633
00:33:30,640 --> 00:33:30,920
Speaker 5: about it?

634
00:33:31,040 --> 00:33:34,920
Speaker 3: Or oh he can't it requires some finesse.

635
00:33:35,040 --> 00:33:38,120
Speaker 2: I think, well, thing, this is the thing like I

636
00:33:38,200 --> 00:33:40,719
Speaker 2: used to blame Joe for the SaaS sell off, right

637
00:33:40,760 --> 00:33:43,800
Speaker 2: because he was the one vibe coding and publicizing vibe coding,

638
00:33:43,840 --> 00:33:45,400
Speaker 2: But now we can all blame situation.

639
00:33:46,360 --> 00:33:52,479
Speaker 5: Yeah, you're welcome. But the strategy that's been adopted by

640
00:33:52,480 --> 00:33:54,840
Speaker 5: Opening Eye is very similar to Talent here, where they

641
00:33:54,840 --> 00:33:57,280
Speaker 5: say we have these forward deployed engineers and we're just

642
00:33:57,320 --> 00:34:02,120
Speaker 5: gonna install them at your place. And so maybe you know,

643
00:34:02,160 --> 00:34:04,360
Speaker 5: I don't necessarily think the enterprises are going to jump

644
00:34:04,400 --> 00:34:07,040
Speaker 5: to vibe code their own system of record. But what

645
00:34:07,080 --> 00:34:10,880
Speaker 5: I do think is that when you have these sales

646
00:34:10,920 --> 00:34:14,240
Speaker 5: teams that call up their customers and say, hey, remember

647
00:34:14,360 --> 00:34:16,239
Speaker 5: last year we said this was what inflation was, and

648
00:34:16,239 --> 00:34:17,520
Speaker 5: then we added a couple percent on top of that.

649
00:34:17,520 --> 00:34:20,280
Speaker 5: So you're getting a five percent price increase. All good, Okay,

650
00:34:20,360 --> 00:34:22,000
Speaker 5: you're not going anywhere because you don't have anywhere else

651
00:34:22,040 --> 00:34:24,560
Speaker 5: to go. Done. Now the person on the other side

652
00:34:24,600 --> 00:34:27,799
Speaker 5: of the phone can say, you know, open AI called

653
00:34:27,800 --> 00:34:30,200
Speaker 5: me the other day, even if they're bluffing, right, So

654
00:34:30,880 --> 00:34:35,240
Speaker 5: you do see like some potential downside to pricing power,

655
00:34:35,680 --> 00:34:38,839
Speaker 5: and that's in the places where it's very unlikely that

656
00:34:38,880 --> 00:34:42,200
Speaker 5: these vibe coded alternatives actually pose a threat. And then

657
00:34:42,239 --> 00:34:45,520
Speaker 5: you see it's been interesting how Anthropic is handled it

658
00:34:45,560 --> 00:34:50,120
Speaker 5: where they've recognized this capability gap where they say, oh,

659
00:34:50,239 --> 00:34:54,000
Speaker 5: the people don't really understand what these tools can do,

660
00:34:54,280 --> 00:34:57,880
Speaker 5: so they've started releasing like suites of AI tools. I

661
00:34:57,920 --> 00:34:59,759
Speaker 5: don't know if you saw the wealth management one, right,

662
00:35:00,120 --> 00:35:02,680
Speaker 5: it's they released the wealth management when I think a

663
00:35:02,719 --> 00:35:05,000
Speaker 5: couple of days ago. It's like you could have done

664
00:35:05,000 --> 00:35:06,400
Speaker 5: this yourself with claud customs.

665
00:35:06,600 --> 00:35:08,600
Speaker 4: This is a really good point, and I hadn't really

666
00:35:08,600 --> 00:35:10,680
Speaker 4: thought of it in that term, because these things that

667
00:35:10,880 --> 00:35:13,799
Speaker 4: like Claude announced or entropic releases something, they're not that

668
00:35:14,000 --> 00:35:17,760
Speaker 4: incredible in some sense. Well, they're essentially just very simple

669
00:35:17,840 --> 00:35:21,400
Speaker 4: reminders you hadn't thought to use this for, you know,

670
00:35:21,520 --> 00:35:24,680
Speaker 4: modeling various retirement scenarios. Actually it's very simple you could

671
00:35:24,680 --> 00:35:27,160
Speaker 4: do that you hadn't thought to use this. So because

672
00:35:27,200 --> 00:35:29,120
Speaker 4: they're simple, they're like marked on files. They're not like

673
00:35:29,160 --> 00:35:32,640
Speaker 4: particularly exotic pieces of software, but they are reminders that

674
00:35:32,880 --> 00:35:35,280
Speaker 4: this thing you didn't think of, Yeah, just do it.

675
00:35:35,280 --> 00:35:37,240
Speaker 2: It's like a thing that you can use to hammer

676
00:35:37,320 --> 00:35:38,719
Speaker 2: your supplier over the head with.

677
00:35:38,960 --> 00:35:41,640
Speaker 5: Right. Yeah, I don't know exactly what the timeline that

678
00:35:41,640 --> 00:35:44,800
Speaker 5: that happens on, but there are going to be adjustments

679
00:35:44,840 --> 00:35:47,360
Speaker 5: to pricing power because of it. And yeah, it seems

680
00:35:47,400 --> 00:35:50,000
Speaker 5: that this is kind of the reason why in the

681
00:35:50,040 --> 00:35:51,880
Speaker 5: beginning I thought that framing the piece this way was

682
00:35:51,960 --> 00:35:55,480
Speaker 5: valuable to our client base and reader base was because

683
00:35:56,160 --> 00:35:58,960
Speaker 5: as an investor, you don't really care if you're presented

684
00:35:59,040 --> 00:36:01,759
Speaker 5: with ten scenario and nine of them are wrong if

685
00:36:01,760 --> 00:36:04,560
Speaker 5: one of them makes you money, right, So I I

686
00:36:04,600 --> 00:36:07,319
Speaker 5: obviously knew that that some people who had already bought

687
00:36:07,360 --> 00:36:09,680
Speaker 5: the dippin software would disagree with the software part, but

688
00:36:09,719 --> 00:36:11,799
Speaker 5: maybe they would agree with the you know, uh with

689
00:36:11,880 --> 00:36:15,920
Speaker 5: the disintermediation part. But then it kind of escaped containment

690
00:36:16,160 --> 00:36:18,160
Speaker 5: and in retrospect, if I was going to write a

691
00:36:18,200 --> 00:36:21,719
Speaker 5: piece for broad distribution, it would probably be pretty optimistic,

692
00:36:22,160 --> 00:36:26,160
Speaker 5: because I'm a pretty optimistic guy. Like, uh so, yeah,

693
00:36:26,160 --> 00:36:27,560
Speaker 5: that's been an interesting experience.

694
00:36:27,760 --> 00:36:30,960
Speaker 3: What was the most surprising thing from this week for you?

695
00:36:32,000 --> 00:36:35,440
Speaker 5: Well, I had someone that that really strongly disagreed with me,

696
00:36:35,480 --> 00:36:37,759
Speaker 5: and then when I asked why, sent me a claud

697
00:36:37,800 --> 00:36:38,360
Speaker 5: read out.

698
00:36:39,920 --> 00:36:43,839
Speaker 6: The kelchy is cool that there's a you can use

699
00:36:43,880 --> 00:36:46,840
Speaker 6: this as a hedge for like, there's.

700
00:36:46,719 --> 00:36:49,839
Speaker 4: Now an instrument which wait, let's see kelshy. I'm gonna

701
00:36:49,840 --> 00:36:54,000
Speaker 4: look it up Kelshi Sactrini scenario. Look if you start

702
00:36:54,080 --> 00:36:57,120
Speaker 4: typing in kelshy and then start see it auto fill

703
00:36:57,200 --> 00:37:00,960
Speaker 4: Satrini scenario, will I love that? Will the Satrini scenario happened?

704
00:37:01,040 --> 00:37:02,480
Speaker 4: It's an eleven point six percent?

705
00:37:02,719 --> 00:37:04,319
Speaker 5: Is that basically the right that you would get if

706
00:37:04,320 --> 00:37:05,200
Speaker 5: you put it in the money market.

707
00:37:05,920 --> 00:37:08,680
Speaker 4: So just trying to read these specifications of the contract

708
00:37:08,760 --> 00:37:11,240
Speaker 4: fine print matter or summary. So if at least three

709
00:37:11,320 --> 00:37:14,520
Speaker 4: of colon unemployment rate exceeds ten percent for the BLS

710
00:37:14,680 --> 00:37:16,840
Speaker 4: S and P five hundred declines more than thirty percent

711
00:37:16,880 --> 00:37:20,960
Speaker 4: from its closing level of issuance. That's weird terminology. Zillow

712
00:37:21,000 --> 00:37:23,239
Speaker 4: Home Index declines more than ten percent, and then you

713
00:37:23,239 --> 00:37:26,400
Speaker 4: have New York City, La San Francisco, Chicago, Houston, Phoenix,

714
00:37:26,800 --> 00:37:31,080
Speaker 4: labor share of GDI falls below fifty percent and CPU

715
00:37:31,200 --> 00:37:33,920
Speaker 4: falls below zero percent. If any of those three things happen,

716
00:37:34,440 --> 00:37:35,840
Speaker 4: then this attornee scenario.

717
00:37:36,120 --> 00:37:37,840
Speaker 2: Is that crazy because like most of that is just

718
00:37:37,840 --> 00:37:42,120
Speaker 2: a financial crash, right, it's not even necessarily tied to AI.

719
00:37:42,480 --> 00:37:44,319
Speaker 4: It's cool, Like do you like that? That's like this

720
00:37:44,440 --> 00:37:46,280
Speaker 4: is now going to be known as just a trainee

721
00:37:46,360 --> 00:37:49,440
Speaker 4: scenario forever, like when like when we get the next

722
00:37:49,520 --> 00:37:52,279
Speaker 4: crisis whenever people it's like, oh, this is like an

723
00:37:52,280 --> 00:37:53,040
Speaker 4: omen this is.

724
00:37:53,520 --> 00:37:55,920
Speaker 5: I feel like anybody considered, like I feel like you

725
00:37:55,920 --> 00:37:58,560
Speaker 5: could make a lot more money on TLT calls if

726
00:37:58,800 --> 00:37:59,560
Speaker 5: three of these things.

727
00:38:00,160 --> 00:38:03,160
Speaker 4: But there's one hundred and twenty five thousand dollars been

728
00:38:03,200 --> 00:38:05,000
Speaker 4: traded in this market, Okay, so it's still.

729
00:38:04,840 --> 00:38:07,719
Speaker 5: Pretty minor deep liquidity. You can't, right, you.

730
00:38:07,680 --> 00:38:10,400
Speaker 4: Can't probably hedge you can't head your whole life or

731
00:38:10,760 --> 00:38:12,320
Speaker 4: your whole business.

732
00:38:12,360 --> 00:38:14,920
Speaker 5: But you know, if I was going to pick a

733
00:38:14,960 --> 00:38:16,959
Speaker 5: thing I'd been known for, it probably would have been

734
00:38:17,120 --> 00:38:19,440
Speaker 5: not this. But you know, you don't get to picked.

735
00:38:19,560 --> 00:38:22,759
Speaker 5: So I still stand by what we've written, and I

736
00:38:22,760 --> 00:38:25,960
Speaker 5: think that it's as a scenario, useful to consider.

737
00:38:26,160 --> 00:38:28,719
Speaker 2: All Right, James, thank you for coming on during a

738
00:38:29,160 --> 00:38:31,560
Speaker 2: very busy and i'm sure surreal week for you.

739
00:38:31,880 --> 00:38:45,640
Speaker 5: Thank you for having me, all.

740
00:38:45,680 --> 00:38:48,719
Speaker 2: Right, Joe, I'm very glad we got James on to

741
00:38:48,760 --> 00:38:52,000
Speaker 2: discuss that because obviously this is the talking point of

742
00:38:52,040 --> 00:38:52,399
Speaker 2: the week.

743
00:38:52,680 --> 00:38:53,320
Speaker 3: At least.

744
00:38:53,480 --> 00:38:56,840
Speaker 2: It is just fascinating from a media perspective how you

745
00:38:56,880 --> 00:38:59,160
Speaker 2: can have these viral pieces that kind of get out

746
00:38:59,160 --> 00:39:01,120
Speaker 2: into the world and develop a life of their own.

747
00:39:01,200 --> 00:39:04,600
Speaker 2: But obviously the major point of interest in all of

748
00:39:04,600 --> 00:39:07,040
Speaker 2: this is these are the things that the market seems

749
00:39:07,080 --> 00:39:09,480
Speaker 2: to be actively considering at the moment.

750
00:39:09,560 --> 00:39:11,840
Speaker 4: Right Paul Krugman wrote a good piece. He disagreed with

751
00:39:11,880 --> 00:39:13,480
Speaker 4: a lot of it, but he pointed out, you know,

752
00:39:14,040 --> 00:39:17,239
Speaker 4: when the radio broadcast of World the World's happened and

753
00:39:17,280 --> 00:39:18,840
Speaker 4: a bunch of people paniced because they thought there was

754
00:39:18,880 --> 00:39:21,880
Speaker 4: some big invasion. It occurred in the environment of a

755
00:39:22,000 --> 00:39:24,200
Speaker 4: very it was like you know, during the depression.

756
00:39:23,840 --> 00:39:26,680
Speaker 6: Yeah, of like existential thread and look like this is

757
00:39:26,719 --> 00:39:28,719
Speaker 6: the worry that has been people have been talking about

758
00:39:28,760 --> 00:39:32,160
Speaker 6: all year long before this piece, and so like the

759
00:39:32,239 --> 00:39:34,520
Speaker 6: whole reason people are like talking about, oh are all

760
00:39:34,560 --> 00:39:36,400
Speaker 6: these software companies that have thrived forever?

761
00:39:36,560 --> 00:39:38,640
Speaker 4: The reason whether many of them are at all time

762
00:39:38,680 --> 00:39:41,080
Speaker 4: lows is because of like, wow, people are very impressed

763
00:39:41,120 --> 00:39:43,680
Speaker 4: with the capabilities, and you have a lot of people

764
00:39:43,840 --> 00:39:47,319
Speaker 4: talking about the potential for mass white collar layoffs. And

765
00:39:47,360 --> 00:39:50,359
Speaker 4: so therefore, you know, I read it as a sort

766
00:39:50,360 --> 00:39:53,080
Speaker 4: of let's put this all together. And to the point

767
00:39:53,120 --> 00:39:54,920
Speaker 4: is that you want to be thinking about scenarios, particularly

768
00:39:54,960 --> 00:39:57,879
Speaker 4: from the public sector response, like, let's actually talk about

769
00:39:57,920 --> 00:39:59,319
Speaker 4: what this could look like. It's strike to me as

770
00:39:59,320 --> 00:40:00,520
Speaker 4: a usefultress.

771
00:40:00,320 --> 00:40:03,239
Speaker 2: Right, and the reaction itself is informative.

772
00:40:03,840 --> 00:40:04,080
Speaker 5: Right.

773
00:40:04,160 --> 00:40:07,680
Speaker 2: So again, we should not be in an environment where

774
00:40:07,880 --> 00:40:11,000
Speaker 2: you can have a think piece a single scenario that

775
00:40:11,080 --> 00:40:14,240
Speaker 2: actually causes a broad sell off that lots of people

776
00:40:14,280 --> 00:40:17,200
Speaker 2: start like pinning on this particular piece. And likewise, we

777
00:40:17,200 --> 00:40:20,120
Speaker 2: shouldn't really be in a scenario where Citadel's Securities published

778
00:40:20,120 --> 00:40:22,400
Speaker 2: as a rebuttal and then everything starts rallying. All it

779
00:40:22,560 --> 00:40:24,879
Speaker 2: underscores is that no one really knows anything at.

780
00:40:24,760 --> 00:40:27,120
Speaker 4: The moment this is ONOK. Yeah, Like there's like people

781
00:40:27,200 --> 00:40:30,520
Speaker 4: are extremely stressed and knowing it's you know, it's like

782
00:40:30,600 --> 00:40:34,680
Speaker 4: it's genuinely it's uncharted territory. It's charted to have a

783
00:40:34,719 --> 00:40:37,399
Speaker 4: technology that is improving as fast as it is. It's

784
00:40:37,520 --> 00:40:38,560
Speaker 4: uncharted to have it.

785
00:40:38,840 --> 00:40:39,080
Speaker 1: You know.

786
00:40:39,120 --> 00:40:43,799
Speaker 4: It's not like one lot, one specific industry is the threat.

787
00:40:43,920 --> 00:40:46,319
Speaker 4: It's like a broad range. No one knows where it's

788
00:40:46,320 --> 00:40:48,520
Speaker 4: going to be. So it's like people are like deeply

789
00:40:48,600 --> 00:40:51,720
Speaker 4: anxious about it, and it articulated a lot of views,

790
00:40:52,080 --> 00:40:54,560
Speaker 4: and it landed at a moment where this was just

791
00:40:54,600 --> 00:40:55,720
Speaker 4: top of mind for everyone.

792
00:40:55,880 --> 00:40:58,000
Speaker 2: The one last thing I'll say about this is I'm

793
00:40:58,040 --> 00:41:00,640
Speaker 2: really glad you asked about policy, because this also seems

794
00:41:00,640 --> 00:41:03,320
Speaker 2: to be the wild card in this entire discussion, which

795
00:41:03,360 --> 00:41:06,040
Speaker 2: is like the outcome of all of this could end

796
00:41:06,120 --> 00:41:09,880
Speaker 2: up being very different depending on what policymakers actually decide

797
00:41:09,920 --> 00:41:10,560
Speaker 2: to do about it.

798
00:41:10,800 --> 00:41:12,520
Speaker 3: And so far we.

799
00:41:12,560 --> 00:41:15,959
Speaker 2: Haven't really seen any like not even early signs of

800
00:41:16,000 --> 00:41:17,560
Speaker 2: how people are thinking about this.

801
00:41:17,640 --> 00:41:23,560
Speaker 4: There's virtually no discussion in DC about anything substantive related

802
00:41:23,600 --> 00:41:27,719
Speaker 4: to like the actual impacts of AI, there's almost none,

803
00:41:27,760 --> 00:41:31,480
Speaker 4: and there's it's this very weird chasm that's opened up

804
00:41:31,680 --> 00:41:34,520
Speaker 4: between how much of a big deal so many people

805
00:41:34,560 --> 00:41:37,680
Speaker 4: are thinking about this and how politicians like they'll talk

806
00:41:37,719 --> 00:41:40,840
Speaker 4: about anything but this. It's very it's actually it's starting

807
00:41:40,840 --> 00:41:41,960
Speaker 4: to get pretty surreal on this.

808
00:41:42,120 --> 00:41:44,000
Speaker 3: Yeah, all right, well shall we leave it there.

809
00:41:44,080 --> 00:41:44,719
Speaker 4: Let's leave it there.

810
00:41:44,760 --> 00:41:45,040
Speaker 3: Okay.

811
00:41:45,120 --> 00:41:47,640
Speaker 2: This has been another episode of the Authoughts podcast. I'm

812
00:41:47,640 --> 00:41:50,200
Speaker 2: Tracy Alloway. You can follow me at Tracy Alloway.

813
00:41:50,320 --> 00:41:52,880
Speaker 4: And I'm Jill Wisenthal. You can follow me at the Stalwart.

814
00:41:52,920 --> 00:41:55,720
Speaker 4: Follow our guest James van Giellen, He's at satriny seven.

815
00:41:56,000 --> 00:41:59,040
Speaker 4: Follow our producers Carmen Rodriguez at Carman armand dash El

816
00:41:59,040 --> 00:42:02,120
Speaker 4: Bennett at DASHBO, and kill Brooks at Kilbrooks. And for

817
00:42:02,160 --> 00:42:04,560
Speaker 4: more odd Laws content, go to Bloomberg dot com slash

818
00:42:04,560 --> 00:42:07,000
Speaker 4: odd Lots for the daily newsletter and all of our episodes.

819
00:42:07,239 --> 00:42:09,040
Speaker 4: You can chat about all of these topics twenty four

820
00:42:09,080 --> 00:42:12,279
Speaker 4: to seven in our discord Discord dot gg slash odd

821
00:42:12,320 --> 00:42:13,200
Speaker 4: lots and.

822
00:42:13,239 --> 00:42:15,200
Speaker 3: If you enjoy odd Lots, if you like it.

823
00:42:15,280 --> 00:42:17,920
Speaker 2: When we talk about the AI doom scenario of twenty

824
00:42:17,960 --> 00:42:20,520
Speaker 2: twenty eight. Then please leave us a positive review on

825
00:42:20,600 --> 00:42:23,440
Speaker 2: your favorite podcast platform. And remember, if you are a

826
00:42:23,480 --> 00:42:26,520
Speaker 2: Bloomberg subscriber, you can listen to all of our episodes

827
00:42:26,560 --> 00:42:27,560
Speaker 2: absolutely ad free.

828
00:42:27,719 --> 00:42:28,799
Speaker 3: All you need to do is.

829
00:42:28,760 --> 00:42:31,400
Speaker 2: Find the Bloomberg channel on Apple Podcasts and follow the

830
00:42:31,440 --> 00:42:32,280
Speaker 2: instructions there.

831
00:42:32,600 --> 00:42:33,400
Speaker 3: Thanks for listening.