1
00:00:02,520 --> 00:00:13,760
Speaker 1: Bloomberg Audio Studios, podcasts, radio news. This is the Bloomberg

2
00:00:13,840 --> 00:00:17,920
Speaker 1: Surveillance Podcast. Catch us live weekdays at seven am Eastern

3
00:00:18,200 --> 00:00:22,000
Speaker 1: on Apple CarPlay or Android Auto with the Bloomberg Business app.

4
00:00:22,360 --> 00:00:25,680
Speaker 1: Listen on demand wherever you get your podcasts, or watch

5
00:00:25,760 --> 00:00:28,440
Speaker 1: us live on YouTube right now.

6
00:00:28,280 --> 00:00:28,960
Speaker 2: In joining us.

7
00:00:29,000 --> 00:00:31,120
Speaker 3: And you know last night when this was blowing up,

8
00:00:31,200 --> 00:00:32,560
Speaker 3: I said get me Mandeep saying.

9
00:00:32,440 --> 00:00:33,559
Speaker 2: You'll join us in a bit here.

10
00:00:34,080 --> 00:00:37,080
Speaker 3: But aniog Rana joins us now from Chicago with a

11
00:00:37,120 --> 00:00:40,520
Speaker 3: little different take than Mandeep. More on the cloud, more

12
00:00:40,560 --> 00:00:43,760
Speaker 3: on the systems analysis, if you will. Of these big companies,

13
00:00:44,280 --> 00:00:46,760
Speaker 3: man Anerog, thank you so much for joining.

14
00:00:46,800 --> 00:00:48,120
Speaker 2: You don't do biholed sell.

15
00:00:48,800 --> 00:00:54,040
Speaker 3: But would you suggest the cell side will shift opinion today.

16
00:00:54,520 --> 00:00:55,560
Speaker 4: For the cloud providers?

17
00:00:55,600 --> 00:00:58,160
Speaker 5: I don't think so because, as you know Paul just

18
00:00:58,240 --> 00:01:01,960
Speaker 5: mentioned about the tweet from my ACOFT CEO, he's probably

19
00:01:02,000 --> 00:01:04,560
Speaker 5: a little bit happier because for him, if the cost

20
00:01:04,600 --> 00:01:07,200
Speaker 5: curve goes down in the long run, and we don't

21
00:01:07,200 --> 00:01:09,520
Speaker 5: know if that's going to happen, you know, his adoption

22
00:01:09,640 --> 00:01:11,679
Speaker 5: rate goes up and his CAPEX goes down.

23
00:01:11,880 --> 00:01:14,360
Speaker 4: So it's it's you know, in a roundabout way.

24
00:01:15,120 --> 00:01:16,560
Speaker 5: Now, we don't think there is going to be an

25
00:01:16,560 --> 00:01:18,840
Speaker 5: impact on CAPEX in the near term, but in the

26
00:01:18,920 --> 00:01:21,679
Speaker 5: long run, if that just shifts, you know, it helps

27
00:01:21,680 --> 00:01:25,880
Speaker 5: their profitability. But more important to that the enterprise you know,

28
00:01:26,160 --> 00:01:29,720
Speaker 5: you could say take create on AI adoption goes up the.

29
00:01:29,520 --> 00:01:32,200
Speaker 3: Theme this morning, Paul's been great about this is like

30
00:01:32,280 --> 00:01:36,000
Speaker 3: drunken sailors from where in Anerich you're just world class

31
00:01:36,000 --> 00:01:40,600
Speaker 3: at this. Are they drunken sailors in terms of CAPEX?

32
00:01:41,040 --> 00:01:43,840
Speaker 3: Or is there a plan? Is there a sequence? Is

33
00:01:43,840 --> 00:01:45,839
Speaker 3: there a pace to spending billions?

34
00:01:47,120 --> 00:01:48,760
Speaker 5: So one of the things, if you would say, is

35
00:01:48,840 --> 00:01:53,360
Speaker 5: if if this AI adoption curve, which is absolutely in

36
00:01:53,440 --> 00:01:55,640
Speaker 5: infancy right now, the only thing we have seen is

37
00:01:55,640 --> 00:01:58,920
Speaker 5: in the consumer world, if enterprises start to put this

38
00:01:59,080 --> 00:02:02,160
Speaker 5: in their core app locations, then they would need a

39
00:02:02,240 --> 00:02:05,880
Speaker 5: lot more you could say, firepower to run it. Most

40
00:02:05,880 --> 00:02:08,200
Speaker 5: of the things that we have discussed is people would

41
00:02:08,240 --> 00:02:10,560
Speaker 5: not want to do this on their own premise. They're

42
00:02:10,600 --> 00:02:13,520
Speaker 5: bigger companies will but most of them will go to

43
00:02:13,639 --> 00:02:17,280
Speaker 5: a handful of cloud providers, and that would be Microsoft, Aws,

44
00:02:17,360 --> 00:02:21,040
Speaker 5: Google and then Oracle. So for all these companies to

45
00:02:21,080 --> 00:02:24,600
Speaker 5: increase their capacity now they are only seeing the orders

46
00:02:24,639 --> 00:02:28,359
Speaker 5: and based on that they are expanding their data center footwork.

47
00:02:28,680 --> 00:02:30,000
Speaker 4: The question is whether do.

48
00:02:29,919 --> 00:02:32,120
Speaker 5: They need the expensive chips to do with this or

49
00:02:32,120 --> 00:02:34,959
Speaker 5: can they run this in with cheaper chips. And that's

50
00:02:35,000 --> 00:02:37,320
Speaker 5: really I think the biggest discussion going to be over

51
00:02:37,360 --> 00:02:38,480
Speaker 5: the next twelve months.

52
00:02:38,760 --> 00:02:41,760
Speaker 6: An raglick Tom has often want to say, my head

53
00:02:41,800 --> 00:02:45,160
Speaker 6: is spinning here. Just since Friday. I've got a big,

54
00:02:45,200 --> 00:02:48,320
Speaker 6: big tech company called Meta I call it Facebook given

55
00:02:48,360 --> 00:02:51,160
Speaker 6: the IPO. We spent a lot of time thinking about

56
00:02:51,200 --> 00:02:55,520
Speaker 6: that name in the tech the ticker FB. They up

57
00:02:55,520 --> 00:02:57,640
Speaker 6: their cap x from fifty billion to sixty five billion,

58
00:02:57,639 --> 00:03:00,800
Speaker 6: a large part on AI. Now I've got this Deep

59
00:03:00,840 --> 00:03:02,600
Speaker 6: Seek thinking about I can do this in.

60
00:03:02,720 --> 00:03:06,000
Speaker 7: Much lower costs. How do we square those two things?

61
00:03:06,520 --> 00:03:08,880
Speaker 5: Yeah, but one thing is for sure, this Deep Seek

62
00:03:08,919 --> 00:03:10,640
Speaker 5: news of the paper has been out for a while,

63
00:03:10,680 --> 00:03:12,799
Speaker 5: so it does got It did get publicized over the

64
00:03:12,840 --> 00:03:15,600
Speaker 5: last few days. But I'd be very surprised if Meta

65
00:03:15,639 --> 00:03:18,840
Speaker 5: AWS and Microsoft did not know about this. It has

66
00:03:18,880 --> 00:03:21,480
Speaker 5: been around for a while, so both of them, both

67
00:03:21,600 --> 00:03:25,959
Speaker 5: Meta and Microsoft talked about their capex last week after

68
00:03:26,040 --> 00:03:28,960
Speaker 5: knowing this news, so you know they have moved.

69
00:03:29,000 --> 00:03:31,079
Speaker 4: I mean, they're far smarter than I am when it

70
00:03:31,080 --> 00:03:31,520
Speaker 4: comes to this.

71
00:03:31,639 --> 00:03:34,000
Speaker 5: So if they're taking up their Capex, I mean there

72
00:03:34,040 --> 00:03:35,200
Speaker 5: must be a real need for it.

73
00:03:35,720 --> 00:03:38,960
Speaker 3: What's it mean for Coopertino there having an early morning?

74
00:03:39,160 --> 00:03:42,600
Speaker 3: They don't have lotes at Coopertino? They got something, Lisa.

75
00:03:42,640 --> 00:03:45,080
Speaker 3: What's the thing with the green powder in it?

76
00:03:45,280 --> 00:03:46,520
Speaker 2: The machi?

77
00:03:46,640 --> 00:03:50,960
Speaker 3: Yeah matter, they're having the machiat Coopertino some undrinkable bilge.

78
00:03:51,120 --> 00:03:53,720
Speaker 3: What are they saying at Apple today?

79
00:03:53,600 --> 00:03:56,320
Speaker 5: They're actually smiling and saying, you guys fight it out

80
00:03:56,360 --> 00:03:59,280
Speaker 5: and tell me what's the best large language model out there.

81
00:03:59,400 --> 00:04:01,200
Speaker 5: I'll just put it in my phone and I won't

82
00:04:01,200 --> 00:04:03,240
Speaker 5: even pay for it. So they if you just look

83
00:04:03,280 --> 00:04:05,560
Speaker 5: at you know, Apple right now, that's kind of the

84
00:04:05,600 --> 00:04:08,760
Speaker 5: safety haven for people is they're not spending that capex.

85
00:04:08,800 --> 00:04:11,080
Speaker 5: If you look at that Capex line, it hasn't moved

86
00:04:11,120 --> 00:04:13,080
Speaker 5: at all in two years, and the free cash slow

87
00:04:13,200 --> 00:04:16,200
Speaker 5: just keep on coming. Their issue, honestly, is more so

88
00:04:16,600 --> 00:04:19,360
Speaker 5: on the you know, the growth side. They are having

89
00:04:19,440 --> 00:04:23,080
Speaker 5: challenges in China that needs to you know, smoothened out

90
00:04:23,160 --> 00:04:25,440
Speaker 5: over the next twelve months. Otherwise this is going to

91
00:04:25,440 --> 00:04:28,160
Speaker 5: be another year for them. To grow only four five percent.

92
00:04:28,360 --> 00:04:30,760
Speaker 5: That's the big issue for them. Nothing to do with AI.

93
00:04:30,920 --> 00:04:32,760
Speaker 6: All right, Well, let's just stay on the Apple story.

94
00:04:32,800 --> 00:04:35,800
Speaker 6: It stocks down eleven percent year to date this year,

95
00:04:35,800 --> 00:04:38,080
Speaker 6: it's only a fifteen percent over the trailing twelve months,

96
00:04:38,120 --> 00:04:41,200
Speaker 6: which is radical on the performance. And for a company

97
00:04:41,240 --> 00:04:45,640
Speaker 6: like Apple, is there a bulkcase here short term?

98
00:04:45,640 --> 00:04:46,080
Speaker 7: For Apple?

99
00:04:47,040 --> 00:04:48,880
Speaker 5: In the short term, it's very difficult to come up

100
00:04:48,880 --> 00:04:51,680
Speaker 5: with the bulkcase because, as we talked about it, even

101
00:04:51,800 --> 00:04:54,600
Speaker 5: you know a few months ago, Apple Intelligence is not

102
00:04:54,640 --> 00:04:56,440
Speaker 5: going to be out in other parts of the world.

103
00:04:56,480 --> 00:04:58,680
Speaker 5: It's not out in China, and China is where they

104
00:04:58,720 --> 00:05:01,520
Speaker 5: are struggling honestly at this and they need to come

105
00:05:01,560 --> 00:05:04,279
Speaker 5: out and pacify investors that you know, we have a

106
00:05:04,279 --> 00:05:08,040
Speaker 5: strategy in China, we will gain market share back. Otherwise

107
00:05:08,040 --> 00:05:10,080
Speaker 5: this is going to be another mut the year for them.

108
00:05:10,640 --> 00:05:12,680
Speaker 5: We think the next phone, which is the one that's

109
00:05:12,680 --> 00:05:15,360
Speaker 5: coming out in September, I know that's fa I think

110
00:05:15,400 --> 00:05:18,039
Speaker 5: that's the big catalyst, but that's like, you know, that's

111
00:05:18,080 --> 00:05:19,839
Speaker 5: eight months out, nine months out.

112
00:05:19,880 --> 00:05:23,280
Speaker 3: Right, and Agrana with us in Chicago right now. We

113
00:05:23,279 --> 00:05:25,760
Speaker 3: welcome all of you on your commute across the Nation.

114
00:05:26,200 --> 00:05:29,240
Speaker 3: For those of you in the office at home on YouTube,

115
00:05:29,320 --> 00:05:31,880
Speaker 3: thank you for being on youtubday. Just brilliant live chat

116
00:05:31,960 --> 00:05:34,760
Speaker 3: of learning a lot from people smarter than me on

117
00:05:34,800 --> 00:05:39,480
Speaker 3: the YouTube live chat this morning. Subscribe at Bloomberg Podcast.

118
00:05:39,600 --> 00:05:42,520
Speaker 3: That's the best way to get to the show out there.

119
00:05:42,720 --> 00:05:45,880
Speaker 3: This is a rare privilege. And for senior management listening

120
00:05:45,880 --> 00:05:49,200
Speaker 3: on radio right now. We adhere to their policy which

121
00:05:49,200 --> 00:05:51,760
Speaker 3: is Ana Agarana and man Deep Sing can never be

122
00:05:51,800 --> 00:05:54,719
Speaker 3: in the same ruct together. It's too much worse to

123
00:05:54,760 --> 00:05:57,800
Speaker 3: the to the franchise. But Paul, why don't you bring

124
00:05:57,800 --> 00:05:58,560
Speaker 3: in man Deep Sing?

125
00:05:58,600 --> 00:06:00,800
Speaker 2: As we talked to both and Ruana and.

126
00:06:00,800 --> 00:06:03,760
Speaker 6: Mandy man Deep Seeing covers the technology space along with

127
00:06:03,800 --> 00:06:07,440
Speaker 6: Ana rog Rana, there are two leaders globally running our

128
00:06:07,480 --> 00:06:11,440
Speaker 6: technology franchise and it is a global franchise analyst in Europe,

129
00:06:11,520 --> 00:06:15,719
Speaker 6: Asia and North America. Man Deep, But I guess what

130
00:06:15,760 --> 00:06:20,520
Speaker 6: we're hearing such Antela from Microsoft put out something that hey,

131
00:06:20,839 --> 00:06:24,520
Speaker 6: cost coming down is pretty natural for the technology space.

132
00:06:24,560 --> 00:06:27,839
Speaker 6: That happens, and it's a good thing because it helps adoption.

133
00:06:28,800 --> 00:06:31,880
Speaker 8: Does that make sense to you, Well, clearly, I think

134
00:06:32,400 --> 00:06:35,600
Speaker 8: the investor expectations were different, which is why you see

135
00:06:35,640 --> 00:06:37,039
Speaker 8: this sort of stock reaction.

136
00:06:37,200 --> 00:06:38,320
Speaker 7: So yes, we.

137
00:06:38,360 --> 00:06:41,080
Speaker 8: Have Moore's law and we have other laws in terms

138
00:06:41,080 --> 00:06:44,200
Speaker 8: of figuring out how the cost curve may look like

139
00:06:44,320 --> 00:06:47,200
Speaker 8: over time. But clearly this is a step change in

140
00:06:47,279 --> 00:06:49,800
Speaker 8: terms of, you know, what you can do with the

141
00:06:50,640 --> 00:06:53,240
Speaker 8: existing chips that are out there, as well as what

142
00:06:53,640 --> 00:06:56,919
Speaker 8: you will need going forward. So I think this is

143
00:06:57,000 --> 00:07:00,960
Speaker 8: a big moment in terms of just that we can

144
00:07:01,000 --> 00:07:03,640
Speaker 8: see with these chips and the kind of scale that

145
00:07:03,760 --> 00:07:07,760
Speaker 8: is required for the next versions of these foundational models.

146
00:07:08,000 --> 00:07:11,200
Speaker 8: But there is no doubt that this makes generative AI

147
00:07:11,440 --> 00:07:14,320
Speaker 8: more accessible to software companies as well as you know,

148
00:07:14,360 --> 00:07:15,200
Speaker 8: overall users.

149
00:07:15,480 --> 00:07:18,440
Speaker 6: Anrak, what do you think this means to you know,

150
00:07:18,480 --> 00:07:21,320
Speaker 6: I know you guys at Bloomberg Intelligence you have access

151
00:07:21,360 --> 00:07:23,200
Speaker 6: to this I d C data, which is some of

152
00:07:23,240 --> 00:07:26,600
Speaker 6: the best data on technology spending in the marketplace, and

153
00:07:26,640 --> 00:07:29,280
Speaker 6: you guys feature that in your research. Is this going

154
00:07:29,320 --> 00:07:32,920
Speaker 6: to change i DC's forecast of tech spending?

155
00:07:34,600 --> 00:07:37,080
Speaker 7: AI spending? Do you think so?

156
00:07:37,160 --> 00:07:39,880
Speaker 5: AI spending has been going up, Paul, We've talked about

157
00:07:39,880 --> 00:07:42,400
Speaker 5: this a few times, but it's the other non AI

158
00:07:42,480 --> 00:07:45,160
Speaker 5: spending that has actually not done so well in the

159
00:07:45,200 --> 00:07:48,120
Speaker 5: last two years and over the next I would say month,

160
00:07:48,200 --> 00:07:50,240
Speaker 5: month and a half, we will get a much clearer

161
00:07:50,280 --> 00:07:54,440
Speaker 5: picture if corporations or enterprises are moving up there over

162
00:07:54,480 --> 00:07:57,680
Speaker 5: our tech budgets or not, because right now AI spending

163
00:07:57,760 --> 00:08:00,880
Speaker 5: is coming at the expense of areas you know, normal

164
00:08:00,960 --> 00:08:05,360
Speaker 5: software or you know, just bording hardware upgrades or consulting

165
00:08:06,240 --> 00:08:08,640
Speaker 5: that really needs to pick up because a large portion

166
00:08:08,760 --> 00:08:11,000
Speaker 5: of the tech ecosystem is some of those companies.

167
00:08:12,000 --> 00:08:12,280
Speaker 7: Mande.

168
00:08:12,280 --> 00:08:14,680
Speaker 6: If I think when I think of this news today again,

169
00:08:14,720 --> 00:08:17,120
Speaker 6: I'm like everybody else. Our listeners and viewers are learning

170
00:08:17,120 --> 00:08:19,800
Speaker 6: about deep Seek and what it means. I can't help

171
00:08:19,840 --> 00:08:23,480
Speaker 6: but think about TikTok at some point. Again, I feel

172
00:08:23,480 --> 00:08:25,280
Speaker 6: like by the end of business today, we're gonna have

173
00:08:25,280 --> 00:08:27,240
Speaker 6: a story coming into Washington. There's gonna be an opinion

174
00:08:27,280 --> 00:08:31,080
Speaker 6: coming from Washington as it relates to deep Seek. Does

175
00:08:31,120 --> 00:08:33,800
Speaker 6: oppose a similar risks? Do you think to TikTok from

176
00:08:33,800 --> 00:08:36,080
Speaker 6: a regulator's perspective, Well.

177
00:08:35,840 --> 00:08:39,600
Speaker 8: It's still the app is still you know, early in

178
00:08:39,679 --> 00:08:43,720
Speaker 8: terms of adoption, and look, I'm sure it's got millions

179
00:08:43,760 --> 00:08:49,000
Speaker 8: of users in East Asia and you know, China and whatnot.

180
00:08:49,080 --> 00:08:54,120
Speaker 8: But I think when it comes to AI consumers, just

181
00:08:54,240 --> 00:08:56,960
Speaker 8: look at the functionality, and you know the fact that

182
00:08:57,000 --> 00:09:00,800
Speaker 8: they have an app wage is accessible for free as

183
00:09:00,800 --> 00:09:03,320
Speaker 8: opposed to paying two hundred dollars a month, they'll end

184
00:09:03,400 --> 00:09:05,959
Speaker 8: up using it. So it's the same argument with TikTok.

185
00:09:06,000 --> 00:09:07,840
Speaker 8: You know they will go with the functionality. They don't

186
00:09:07,880 --> 00:09:09,120
Speaker 8: care about the privacy and.

187
00:09:09,160 --> 00:09:11,920
Speaker 3: The data man keep saying with us and ana ug Runa,

188
00:09:11,960 --> 00:09:15,640
Speaker 3: we continue anerog, let's go to earnings. Let's like remove

189
00:09:15,679 --> 00:09:19,800
Speaker 3: ourselves from deep seek. What is the character of the

190
00:09:19,880 --> 00:09:23,679
Speaker 3: cloud mystery that you see in the earning season, Aniog.

191
00:09:24,800 --> 00:09:27,760
Speaker 5: Yeah, we are anticipating in improvement in growth rates for

192
00:09:27,800 --> 00:09:31,040
Speaker 5: both Microsoft and AWS going in. In fact, Microsoft, we

193
00:09:31,080 --> 00:09:34,040
Speaker 5: are hoping they come out and increase their guidance for

194
00:09:34,200 --> 00:09:36,920
Speaker 5: Azure numbers for the next two quarters. They have been

195
00:09:36,960 --> 00:09:40,560
Speaker 5: struggling with supply challenges where they are not able to

196
00:09:40,600 --> 00:09:43,679
Speaker 5: fulfill the demand that's coming in in their data centers,

197
00:09:43,679 --> 00:09:46,760
Speaker 5: and if they give that signal that you know, that

198
00:09:47,000 --> 00:09:50,760
Speaker 5: should revive or pacify some of the fears that we

199
00:09:50,800 --> 00:09:54,200
Speaker 5: are seeing right now in the market. Same thing for

200
00:09:54,240 --> 00:09:57,160
Speaker 5: a WS AWS has been talking about a recovery in

201
00:09:57,200 --> 00:10:01,319
Speaker 5: consumption for their core business looking forward to some comments

202
00:10:01,360 --> 00:10:03,120
Speaker 5: sometimes as well from them.

203
00:10:03,240 --> 00:10:04,319
Speaker 4: So I'm hopeful that.

204
00:10:04,280 --> 00:10:06,760
Speaker 5: Both companies will come up with good numbers over the

205
00:10:06,760 --> 00:10:08,120
Speaker 5: next you know, ten days.

206
00:10:08,200 --> 00:10:10,560
Speaker 3: But let's say it's not like you know, Procter and Gamble.

207
00:10:10,600 --> 00:10:13,840
Speaker 3: It's not a unit and price store hit the revenue line.

208
00:10:14,080 --> 00:10:17,600
Speaker 3: But where is the fear in the earning season analog

209
00:10:18,080 --> 00:10:20,920
Speaker 3: on the income statement. Is it a margin fear, a

210
00:10:21,040 --> 00:10:23,599
Speaker 3: revenue fear, or something I don't understand.

211
00:10:24,679 --> 00:10:26,840
Speaker 5: I think it's going to be a backlock slash revenue

212
00:10:26,840 --> 00:10:30,320
Speaker 5: fear more than the margins, because everybody understands this is

213
00:10:30,360 --> 00:10:32,760
Speaker 5: a very you know, unique business. When you have a

214
00:10:32,760 --> 00:10:35,560
Speaker 5: fixed cost business like this, you can scale it up

215
00:10:35,559 --> 00:10:38,400
Speaker 5: in the near town, but once you reach a mature face,

216
00:10:38,640 --> 00:10:40,760
Speaker 5: you know, all the revenue comes back, or you can

217
00:10:40,760 --> 00:10:42,920
Speaker 5: distribute that over a very large platform.

218
00:10:43,080 --> 00:10:45,040
Speaker 4: That has been the case for Cloud over the last

219
00:10:45,160 --> 00:10:45,960
Speaker 4: you know, ten years.

220
00:10:46,520 --> 00:10:49,320
Speaker 5: So in an expansion of capacity, people don't really you know,

221
00:10:49,480 --> 00:10:51,240
Speaker 5: see that as a big issue in the near town

222
00:10:51,920 --> 00:10:52,400
Speaker 5: man Deep.

223
00:10:52,520 --> 00:10:56,760
Speaker 6: Last Friday, Meta took their capex guidance from fifty billion,

224
00:10:57,000 --> 00:11:00,560
Speaker 6: which was consensus, which was kind of guidance, to sixty

225
00:11:00,600 --> 00:11:04,960
Speaker 6: five billion. I've never seen that before. In order of magnitude,

226
00:11:05,400 --> 00:11:06,679
Speaker 6: what's going on there.

227
00:11:06,840 --> 00:11:11,199
Speaker 8: Well, and I guess they probably knew this deep Seek advancement,

228
00:11:11,320 --> 00:11:13,080
Speaker 8: so it's not as if, you know, they were cut

229
00:11:13,080 --> 00:11:16,000
Speaker 8: off guard with the deep Seek release, so they did

230
00:11:16,000 --> 00:11:20,000
Speaker 8: it despite the you know, the whole thing around deep Seek.

231
00:11:20,080 --> 00:11:23,040
Speaker 8: So to my mind, you know, every company is doubling

232
00:11:23,080 --> 00:11:26,160
Speaker 8: down on AI right now, even though we are talking

233
00:11:26,160 --> 00:11:29,440
Speaker 8: about pricing, compression and whatnot. But this is the moment

234
00:11:29,520 --> 00:11:34,480
Speaker 8: where generative AI will be incorporated in all software, all applications,

235
00:11:34,520 --> 00:11:36,600
Speaker 8: and every company wants a piece of it, so they

236
00:11:36,600 --> 00:11:38,839
Speaker 8: don't mind spending more on County in.

237
00:11:38,760 --> 00:11:41,480
Speaker 3: The room a ROD does a public one AI. Is

238
00:11:41,480 --> 00:11:45,480
Speaker 3: there any interest that Lisa Mateo is waiting and waiting

239
00:11:45,559 --> 00:11:46,800
Speaker 3: and waiting for AI.

240
00:11:48,320 --> 00:11:51,000
Speaker 5: See, the thing is that the consumers have already embraced

241
00:11:51,040 --> 00:11:55,640
Speaker 5: it through chat, GPT, through Gemini, through Perplexity. But when

242
00:11:55,640 --> 00:11:58,040
Speaker 5: you look at the enterprise use case, we are still

243
00:11:58,040 --> 00:12:00,880
Speaker 5: in the building phase. And that's probably this year is

244
00:12:00,920 --> 00:12:03,360
Speaker 5: going to be like next year, we should see a

245
00:12:03,440 --> 00:12:06,880
Speaker 5: much fostered adoption by companies in terms of and that

246
00:12:07,160 --> 00:12:09,839
Speaker 5: you said will show up in cloud numbers right now.

247
00:12:09,880 --> 00:12:12,480
Speaker 5: Even when you look for somebody like a Microsoft, most

248
00:12:12,480 --> 00:12:15,080
Speaker 5: of their AI revenue is from consumer app, which is

249
00:12:15,320 --> 00:12:17,000
Speaker 5: basically open AI is shock GPT.

250
00:12:18,000 --> 00:12:20,240
Speaker 7: So all right, Tom, we like to talk to anog.

251
00:12:20,480 --> 00:12:22,440
Speaker 2: Do you like Ludlow this morning?

252
00:12:22,640 --> 00:12:22,720
Speaker 8: No?

253
00:12:22,800 --> 00:12:26,120
Speaker 6: You just I mean we're likes will be Bloomberg Tech exactly.

254
00:12:26,200 --> 00:12:28,960
Speaker 2: I mean, can you get the shoes with the White Souls?

255
00:12:29,040 --> 00:12:32,040
Speaker 7: Yes, that's what David shoes.

256
00:12:33,040 --> 00:12:36,000
Speaker 2: No, you weren't shoes of the White Souls.

257
00:12:36,520 --> 00:12:39,600
Speaker 5: No, No, old school.

258
00:12:39,400 --> 00:12:41,880
Speaker 6: And Tom, you know Man Deep and Hon they're pretty good.

259
00:12:41,920 --> 00:12:45,319
Speaker 6: But the research, the real investment research on this whole

260
00:12:46,200 --> 00:12:48,520
Speaker 6: Deep seek Chinese AI.

261
00:12:48,559 --> 00:12:51,760
Speaker 7: Robert Lee. These are Bloomberg intelligence analysts in Hong Kong.

262
00:12:51,800 --> 00:12:53,680
Speaker 6: So folks who have access to the Bloomberg terminal go

263
00:12:53,720 --> 00:12:57,679
Speaker 6: there because Man Deep and they're just he was out

264
00:12:57,679 --> 00:12:59,440
Speaker 6: with research a couple of months ago on this and

265
00:12:59,480 --> 00:13:01,000
Speaker 6: he says, hey, this stuff is there.

266
00:13:01,080 --> 00:13:02,160
Speaker 2: Okay.

267
00:13:02,440 --> 00:13:06,840
Speaker 3: You know Lea in Uh in Hong Kong was way

268
00:13:06,840 --> 00:13:08,960
Speaker 3: out front of this. Why are we stunned this morning?

269
00:13:09,520 --> 00:13:12,199
Speaker 8: Well because he didn't found the table in terms of saying,

270
00:13:12,240 --> 00:13:15,560
Speaker 8: my god, the table we're negative?

271
00:13:15,679 --> 00:13:18,240
Speaker 3: Where are we negative? Three point five percent? In YAC

272
00:13:18,360 --> 00:13:21,880
Speaker 3: right now? What I mean, let's bring it back to

273
00:13:25,080 --> 00:13:28,280
Speaker 3: what happened. Why were we surprised.

274
00:13:27,720 --> 00:13:30,160
Speaker 8: By this, No, I don't think you are surprised because

275
00:13:30,240 --> 00:13:32,760
Speaker 8: open source has been talked about a lot. We have

276
00:13:32,880 --> 00:13:35,679
Speaker 8: been talking about how companies are looking to you know,

277
00:13:35,840 --> 00:13:40,280
Speaker 8: drive scaling uh at inference time, and how pricing was

278
00:13:40,320 --> 00:13:43,520
Speaker 8: a constraint, and so look, this is the way the

279
00:13:43,559 --> 00:13:46,440
Speaker 8: industry was headed. It's just that Deep Sea came out

280
00:13:46,440 --> 00:13:49,839
Speaker 8: with a you know, a long paper that describes what

281
00:13:49,880 --> 00:13:52,480
Speaker 8: they did in detail, and everyone else is kind of

282
00:13:52,679 --> 00:13:56,160
Speaker 8: closed source. So being open source has its advantages, and

283
00:13:56,200 --> 00:13:57,840
Speaker 8: that's what we're seeing with deep Seat here.

284
00:13:58,600 --> 00:14:02,440
Speaker 6: So the baniers to entry here on ROD just in general,

285
00:14:02,840 --> 00:14:06,240
Speaker 6: on AI are the banished entry low for AI.

286
00:14:08,120 --> 00:14:11,079
Speaker 5: I think that's exactly what this paper that Mandep is

287
00:14:11,080 --> 00:14:13,360
Speaker 5: alluding to is talking about. That you and I can

288
00:14:13,400 --> 00:14:15,880
Speaker 5: go out and build some of this thing without much capital.

289
00:14:16,120 --> 00:14:18,160
Speaker 5: Now the question is how are you going to distribute it?

290
00:14:18,400 --> 00:14:19,960
Speaker 5: And which is why I go back to the cloud

291
00:14:19,960 --> 00:14:22,400
Speaker 5: providers that those are the guys who are going to

292
00:14:22,440 --> 00:14:25,600
Speaker 5: work with enterprises to embed some of this technology. If

293
00:14:25,640 --> 00:14:28,000
Speaker 5: you look at somebody like an AWS, they work with

294
00:14:28,040 --> 00:14:31,080
Speaker 5: all these providers, they work with Meta Islama model or

295
00:14:31,080 --> 00:14:35,600
Speaker 5: Anthropics cloud models. So it's really up to the customer

296
00:14:35,680 --> 00:14:39,200
Speaker 5: which model they choose or sometimes they would write the

297
00:14:39,240 --> 00:14:43,160
Speaker 5: reroute the inquiry based on what's better for that particular question.

298
00:14:43,680 --> 00:14:45,520
Speaker 4: So it's really the distribution.

299
00:14:45,880 --> 00:14:48,360
Speaker 5: What I'm saying is is the value here which takes

300
00:14:48,400 --> 00:14:50,600
Speaker 5: us back to the cloud providers and also do Apple.

301
00:14:50,840 --> 00:14:52,960
Speaker 3: This has been fun erragran and your trooper to get

302
00:14:53,040 --> 00:14:56,560
Speaker 3: up this early in Chicago aner Agroana with Bloomberg Intelligence

303
00:14:56,560 --> 00:14:58,680
Speaker 3: at Man Deep saying, can we say on behalf of

304
00:14:58,680 --> 00:14:59,160
Speaker 3: all of us?

305
00:14:59,160 --> 00:15:00,440
Speaker 2: Mandep, thank you much.

306
00:15:01,040 --> 00:15:03,680
Speaker 3: Okay, we'll try not to do it three times tomorrow,

307
00:15:04,080 --> 00:15:05,680
Speaker 3: maybe once, or maybe we'll see.

308
00:15:05,480 --> 00:15:06,480
Speaker 2: A Wednesday or whatever.

309
00:15:06,840 --> 00:15:09,920
Speaker 3: Man Deep sing and enter Agrana. Is what the show

310
00:15:09,960 --> 00:15:13,480
Speaker 3: is about, just world class on tech analogy.

311
00:15:18,360 --> 00:15:21,960
Speaker 1: You're listening to the Bloomberg Surveillance Podcast. Catch us live

312
00:15:22,000 --> 00:15:25,200
Speaker 1: weekday afternoons from seven to ten am Eastern Listen on

313
00:15:25,240 --> 00:15:28,920
Speaker 1: Applecarplay and Android Otto with the Bloomberg Business app, or

314
00:15:29,080 --> 00:15:30,720
Speaker 1: watch us live on YouTube.

315
00:15:31,200 --> 00:15:31,880
Speaker 2: We Get Lucky.

316
00:15:31,960 --> 00:15:35,200
Speaker 3: Joe Wisenthal has done so much for Bloomberg, I should say,

317
00:15:35,200 --> 00:15:37,560
Speaker 3: for business journalism, and of course it's work here at

318
00:15:37,560 --> 00:15:41,520
Speaker 3: Bloomberg with Tracy Alloway. But one of the great things

319
00:15:41,840 --> 00:15:44,400
Speaker 3: is he did Paul, he you know, I'm like a

320
00:15:44,640 --> 00:15:48,400
Speaker 3: I E I E I O. Joe's actually doing it,

321
00:15:48,480 --> 00:15:51,320
Speaker 3: he said, no, but but nobody. He looked at the

322
00:15:51,360 --> 00:15:54,480
Speaker 3: summer tour of Light Sweet Crude and said, I can't

323
00:15:54,560 --> 00:15:59,240
Speaker 3: do the Light Sweet Crude analysis without AI. You three

324
00:15:59,320 --> 00:16:04,680
Speaker 3: days ago you nailed deep Seek. You're using it at home. Yeah,

325
00:16:04,920 --> 00:16:07,160
Speaker 3: what's the experience of deep Sea?

326
00:16:07,240 --> 00:16:09,600
Speaker 9: So I'll just see the main There are two main

327
00:16:09,640 --> 00:16:14,560
Speaker 9: takeaways for me. One is I have never experienced any

328
00:16:14,600 --> 00:16:18,160
Speaker 9: technology in the history of any tech, computers, apps, whatever,

329
00:16:18,400 --> 00:16:20,040
Speaker 9: that have zero switching.

330
00:16:19,720 --> 00:16:21,400
Speaker 7: Costs the way AI chatbots have.

331
00:16:21,840 --> 00:16:25,720
Speaker 9: I used Chad GBT for a long time. Claude came out,

332
00:16:26,360 --> 00:16:29,280
Speaker 9: the one from Anthropic, I was like, oh, this works Gemini.

333
00:16:29,600 --> 00:16:31,600
Speaker 9: I used Gemini now that for some reason, that one

334
00:16:31,600 --> 00:16:33,600
Speaker 9: hasn't sucking as much then I. But then deep Sea

335
00:16:33,680 --> 00:16:35,320
Speaker 9: came out and I said, or I saw the chap

336
00:16:35,400 --> 00:16:37,760
Speaker 9: out last week online and I was like, I'm gonna

337
00:16:37,800 --> 00:16:40,360
Speaker 9: try this out. It took me thirty seconds to register,

338
00:16:40,720 --> 00:16:42,600
Speaker 9: and I was like, this is just as good as

339
00:16:42,680 --> 00:16:44,800
Speaker 9: what I'm getting from Chad. You know, I ask at

340
00:16:44,840 --> 00:16:46,480
Speaker 9: random things during the day and.

341
00:16:47,680 --> 00:16:52,400
Speaker 2: It's just zerop radio. It's all, you know.

342
00:16:52,520 --> 00:16:55,440
Speaker 9: I asked a random historical questions or linguistic I'm interested

343
00:16:55,440 --> 00:16:56,040
Speaker 9: in Lenco, like.

344
00:16:55,960 --> 00:16:58,680
Speaker 2: Why did Texas lose to Ohio? No, I can't. That's

345
00:16:58,720 --> 00:17:00,640
Speaker 2: too painful. I don't ask that.

346
00:17:00,920 --> 00:17:04,240
Speaker 9: But then the other thing is I made a twenty

347
00:17:04,280 --> 00:17:07,760
Speaker 9: twenty five one of my New Year's resolutions. Yep, I've

348
00:17:07,800 --> 00:17:09,800
Speaker 9: never I said I'm going to try and code something

349
00:17:09,920 --> 00:17:12,159
Speaker 9: with AI one of my goals. Yeah, And so I

350
00:17:12,200 --> 00:17:14,280
Speaker 9: started like building a little software. But I just asking

351
00:17:14,320 --> 00:17:17,480
Speaker 9: to create this code, and uh, I was able to

352
00:17:17,520 --> 00:17:19,840
Speaker 9: plug my app that I'm building on my home. It's

353
00:17:19,840 --> 00:17:24,920
Speaker 9: a little linguistics app into the Deepseek API, and it's

354
00:17:24,920 --> 00:17:26,840
Speaker 9: like a fraction of the cost of what I use

355
00:17:26,880 --> 00:17:28,920
Speaker 9: for the open AI API. So, like I just played

356
00:17:28,920 --> 00:17:30,920
Speaker 9: around with it. I was like, Wow, here's this thing

357
00:17:30,960 --> 00:17:35,520
Speaker 9: that seems just as good as changchipyt. Sorry, here's this thing.

358
00:17:35,359 --> 00:17:37,879
Speaker 2: That seems just good changing formosa.

359
00:17:39,040 --> 00:17:41,960
Speaker 6: So yes, I'm not surprised that, Like, all right, the

360
00:17:42,080 --> 00:17:44,639
Speaker 6: dumb question there, Yeah, are we not using Google anymore

361
00:17:44,640 --> 00:17:45,080
Speaker 6: because of this?

362
00:17:45,200 --> 00:17:45,280
Speaker 10: No?

363
00:17:45,320 --> 00:17:47,719
Speaker 9: I still well, look it is you know, like I

364
00:17:47,760 --> 00:17:49,800
Speaker 9: still do. I still use Google all the time. I

365
00:17:49,840 --> 00:17:53,000
Speaker 9: still use Wikipedia all the time. They're just different.

366
00:17:53,240 --> 00:17:55,679
Speaker 7: But what's your initial go to, oh, I need to

367
00:17:55,760 --> 00:17:56,440
Speaker 7: know about this.

368
00:17:56,720 --> 00:18:01,000
Speaker 9: I think it really depends on the context. I say,

369
00:18:01,080 --> 00:18:05,399
Speaker 9: I really like using AI for things like understanding a

370
00:18:05,520 --> 00:18:08,639
Speaker 9: term within a specific context or something like that or something.

371
00:18:08,880 --> 00:18:10,480
Speaker 9: But you know, like you don't know if it's true,

372
00:18:10,520 --> 00:18:12,640
Speaker 9: So you know, I always do a backup. I see

373
00:18:12,640 --> 00:18:14,720
Speaker 9: something and then I say, oh, look at Google is

374
00:18:14,760 --> 00:18:15,600
Speaker 9: like is this actually true?

375
00:18:16,160 --> 00:18:21,120
Speaker 3: Morgan Brown, Dropbox Imperial w He's out with the smartest

376
00:18:21,160 --> 00:18:23,359
Speaker 3: note I've seen where he's talking about decimal points. I

377
00:18:23,359 --> 00:18:25,880
Speaker 3: don't want to go into that. You and Tracy at

378
00:18:25,880 --> 00:18:30,520
Speaker 3: audience talk a lot more to techy people about this.

379
00:18:31,000 --> 00:18:36,600
Speaker 3: What's their message on the future of American AI versus

380
00:18:36,640 --> 00:18:37,720
Speaker 3: these shock threats.

381
00:18:39,400 --> 00:18:41,840
Speaker 9: I mean, you know what, here's what my my take,

382
00:18:42,600 --> 00:18:46,000
Speaker 9: which is that they're already I mean, we saw the

383
00:18:46,000 --> 00:18:49,000
Speaker 9: Trump announcement last week with Stargate right half a trillion dollar.

384
00:18:49,200 --> 00:18:51,800
Speaker 3: I thought that was I thought that was the movie

385
00:18:51,880 --> 00:18:54,320
Speaker 3: with what's happening now?

386
00:18:54,440 --> 00:18:56,680
Speaker 9: I think because of the arrival of this very compelling

387
00:18:56,920 --> 00:18:59,680
Speaker 9: Chinese model, is it just cements the notion that there

388
00:18:59,720 --> 00:19:02,119
Speaker 9: is a global arms race allah the bomb And I

389
00:19:02,119 --> 00:19:04,920
Speaker 9: don't know if achieving AGI or whatever people talk about

390
00:19:05,080 --> 00:19:08,760
Speaker 9: is a breakthrough on par or is strategic as a

391
00:19:08,960 --> 00:19:12,159
Speaker 9: nuclear bomb. But now everyone is just going to be

392
00:19:12,480 --> 00:19:13,959
Speaker 9: we have to beat them, we have to beat them,

393
00:19:13,960 --> 00:19:16,480
Speaker 9: and there's gonna be so much money flowing to tech

394
00:19:16,520 --> 00:19:19,760
Speaker 9: companies now in this idea that there is a global race, Joe.

395
00:19:19,560 --> 00:19:21,840
Speaker 6: I feel like by the end of business today, the

396
00:19:21,840 --> 00:19:24,719
Speaker 6: top star on the Bloomberg terournal will be US government

397
00:19:24,760 --> 00:19:28,840
Speaker 6: looking at deep Seek as it's looking at TikTok TikTok.

398
00:19:29,119 --> 00:19:30,560
Speaker 7: Well, you know what, so this is interesting.

399
00:19:30,560 --> 00:19:32,240
Speaker 9: The other thing, and thanks for reminding me of this,

400
00:19:32,440 --> 00:19:35,520
Speaker 9: which is that deep Seek is a piece of open,

401
00:19:35,920 --> 00:19:38,960
Speaker 9: open source software. It is it's unbannable. Now, there is

402
00:19:39,000 --> 00:19:42,000
Speaker 9: an app that theoretically could be banned from the app store,

403
00:19:42,000 --> 00:19:43,720
Speaker 9: but I think it'll be high. But it's a piece

404
00:19:43,760 --> 00:19:44,639
Speaker 9: of open source software.

405
00:19:44,880 --> 00:19:45,480
Speaker 3: I have a friend.

406
00:19:45,560 --> 00:19:47,320
Speaker 9: You know, you can run it on your own computer

407
00:19:47,480 --> 00:19:50,040
Speaker 9: if you have the And so this is fundamentally different

408
00:19:50,040 --> 00:19:52,639
Speaker 9: from TikTok in the sense that anyone can run this

409
00:19:52,680 --> 00:19:56,160
Speaker 9: piece of code, just download it, and so it's fundamentally unbannable.

410
00:19:56,240 --> 00:19:59,680
Speaker 3: You've been on fire, Adam Tooms this weekend. Very nice,

411
00:19:59,720 --> 00:20:04,280
Speaker 3: come your euro dollar three part series. You have Howard

412
00:20:04,320 --> 00:20:06,800
Speaker 3: Marx on Good morning, mister Mark, thank you so much

413
00:20:06,840 --> 00:20:09,840
Speaker 3: for your support over the years. What's your message from

414
00:20:09,880 --> 00:20:11,600
Speaker 3: Howard Marx about this?

415
00:20:11,640 --> 00:20:14,080
Speaker 9: Odd He doesn't think we're in an AYE bubble yet,

416
00:20:14,240 --> 00:20:17,600
Speaker 9: he said, you know, he like doesn't zuberans he doesn't see,

417
00:20:17,680 --> 00:20:19,480
Speaker 9: which I thought was really interesting because to me, it

418
00:20:19,480 --> 00:20:22,440
Speaker 9: feels like exuberance. But the idea of me questioning Howard

419
00:20:22,480 --> 00:20:25,760
Speaker 9: marks on questions like what is exuberance really feels like

420
00:20:25,960 --> 00:20:28,720
Speaker 9: someone who has experienced multiple cycles over the years, right,

421
00:20:28,920 --> 00:20:31,120
Speaker 9: I don't think I would go with my take versus take.

422
00:20:31,160 --> 00:20:32,639
Speaker 2: Okay, did you go to NAM were you out?

423
00:20:32,720 --> 00:20:34,840
Speaker 4: And well?

424
00:20:35,520 --> 00:20:39,119
Speaker 3: Well, I mean the endorsements like sweet crudez. Should I

425
00:20:39,240 --> 00:20:41,880
Speaker 3: go with the Gratch, pink Rander Penguin?

426
00:20:42,359 --> 00:20:45,119
Speaker 2: I mean, that's what's got the fishman pick you got it?

427
00:20:45,200 --> 00:20:47,639
Speaker 2: You got it? I mean it's five forty nine and Sweetwater.

428
00:20:47,720 --> 00:20:52,000
Speaker 2: I'll tell missus that's like you got to it's free

429
00:20:52,200 --> 00:20:52,840
Speaker 2: and it's pink.

430
00:20:54,480 --> 00:20:55,360
Speaker 7: Is that done with the Jeff?

431
00:20:56,640 --> 00:20:56,760
Speaker 4: Oh?

432
00:20:56,800 --> 00:21:01,000
Speaker 3: Yeah, they did, Joe bid But I'm sorry, Wretches on fire,

433
00:21:01,680 --> 00:21:04,959
Speaker 3: Retches on fire, that's the real break. Fender bought Gretch

434
00:21:05,000 --> 00:21:07,520
Speaker 3: and you know what's light sweet crew?

435
00:21:07,520 --> 00:21:08,160
Speaker 2: Doing you playing.

436
00:21:08,200 --> 00:21:11,760
Speaker 9: We have a show this Wednesday the eleventh Street Bar

437
00:21:11,920 --> 00:21:16,800
Speaker 9: in Manhattan. It's free, eight o'clock East Village. Everyone should

438
00:21:16,800 --> 00:21:18,399
Speaker 9: come out. We're gonna play some new songs.

439
00:21:18,400 --> 00:21:19,439
Speaker 2: Are you really for?

440
00:21:19,640 --> 00:21:22,080
Speaker 3: And I like, you know, I'm in bed at a thirty?

441
00:21:22,920 --> 00:21:24,760
Speaker 3: Does it really start at eight? Or are you doing

442
00:21:24,760 --> 00:21:25,280
Speaker 3: the rock star?

443
00:21:25,640 --> 00:21:27,359
Speaker 9: Do we have an opening act at eight? We're going

444
00:21:27,440 --> 00:21:29,360
Speaker 9: to be on just after nine? But you know after

445
00:21:29,960 --> 00:21:32,640
Speaker 9: come out one night. It's one night, yeah.

446
00:21:32,440 --> 00:21:35,440
Speaker 7: But the next morning have one.

447
00:21:35,680 --> 00:21:37,040
Speaker 9: You have one off day the next day.

448
00:21:37,240 --> 00:21:40,400
Speaker 2: Okay. So you got Howard Marks coming up online out today.

449
00:21:40,480 --> 00:21:40,920
Speaker 2: Check it out?

450
00:21:41,080 --> 00:21:44,280
Speaker 3: Okay, Joe Wisenthal, thank you so much and really folks

451
00:21:44,320 --> 00:21:48,760
Speaker 3: followed Joe the Stalwart and uh on Twitter, and also

452
00:21:48,960 --> 00:21:51,920
Speaker 3: followed Joe on LinkedIn where he's really going step by

453
00:21:51,920 --> 00:21:55,359
Speaker 3: step through his experience with AI, which I think is

454
00:21:55,440 --> 00:21:56,119
Speaker 3: really cool.

455
00:22:00,880 --> 00:22:04,760
Speaker 1: This is the Bloomberg Surveillance Podcast. Listen live each weekday

456
00:22:04,800 --> 00:22:07,840
Speaker 1: starting at seven am Eastern on Apple Corplay and Android

457
00:22:07,840 --> 00:22:10,879
Speaker 1: Auto with the Bloomberg Business app. You can also listen

458
00:22:10,960 --> 00:22:14,240
Speaker 1: live on Amazon Alexa from our flagship New York station,

459
00:22:14,760 --> 00:22:17,440
Speaker 1: Just Say Alexa play Bloomberg eleven thirty.

460
00:22:18,200 --> 00:22:23,320
Speaker 3: Stacey Raskin is double chemical engineering definitive out of UCLA

461
00:22:23,960 --> 00:22:26,639
Speaker 3: and mit he wandered back to the Avenue of the

462
00:22:26,720 --> 00:22:30,399
Speaker 3: Stars where he is a star for Bernstein. What you

463
00:22:30,520 --> 00:22:33,480
Speaker 3: need to know is in the old days, you had

464
00:22:33,480 --> 00:22:37,719
Speaker 3: a black covered book from Sanford Bernstein and you knew

465
00:22:37,920 --> 00:22:40,720
Speaker 3: good morning Brad Hints that you were dealing with the

466
00:22:40,800 --> 00:22:43,840
Speaker 3: best Stacy honored that you could join us this morning.

467
00:22:44,080 --> 00:22:48,520
Speaker 3: Will you change by hold sell on the American tech

468
00:22:48,560 --> 00:22:52,760
Speaker 3: companies because of this Deep Seak announcement, I mean.

469
00:22:52,800 --> 00:22:54,120
Speaker 2: Lok, so we're not change.

470
00:22:54,119 --> 00:22:58,439
Speaker 10: We haven't changed anything. We get killed ourselves over the

471
00:22:58,440 --> 00:23:00,880
Speaker 10: weekend to go through some of this stuff and put

472
00:23:00,920 --> 00:23:04,320
Speaker 10: some views in places like I had three kind of

473
00:23:04,359 --> 00:23:08,520
Speaker 10: broad takeaways please from yeah, from from from the work.

474
00:23:08,560 --> 00:23:10,680
Speaker 10: But number one is I think there's been a headline.

475
00:23:11,080 --> 00:23:13,520
Speaker 10: But I think this is why the circus. There was

476
00:23:13,560 --> 00:23:16,200
Speaker 10: a headline that kind of came out all is that

477
00:23:16,320 --> 00:23:18,840
Speaker 10: was kind of like, hey, deep Sea duplicated open AI

478
00:23:19,000 --> 00:23:21,320
Speaker 10: for five million dollars, and then went so I do

479
00:23:21,400 --> 00:23:24,639
Speaker 10: not think they duplicated open AI for five million dollars.

480
00:23:26,119 --> 00:23:28,440
Speaker 10: Number two, Look, I think the models that they built

481
00:23:28,440 --> 00:23:30,919
Speaker 10: are fantastic. They're really really good. But I don't think

482
00:23:30,960 --> 00:23:35,080
Speaker 10: they're miracles. We can talk about why and and number

483
00:23:35,080 --> 00:23:37,520
Speaker 10: three finally, like, I think the panic and there's clearly

484
00:23:37,720 --> 00:23:40,639
Speaker 10: a full blown on panic going on over the weekend

485
00:23:40,720 --> 00:23:44,720
Speaker 10: and this morning, I think it's overblown. I think, you know,

486
00:23:44,800 --> 00:23:46,840
Speaker 10: I do not think that deep seek is the doomsday

487
00:23:46,880 --> 00:23:48,080
Speaker 10: for for AI infrastructure.

488
00:23:48,119 --> 00:23:50,240
Speaker 2: For those of you are for those of you are

489
00:23:50,320 --> 00:23:50,840
Speaker 2: on radio.

490
00:23:51,160 --> 00:23:54,240
Speaker 3: On YouTube, you can see a space shuttle behind a

491
00:23:54,320 --> 00:23:56,440
Speaker 3: gentleman from m I T I.

492
00:23:56,920 --> 00:23:58,440
Speaker 10: I didn't even know that I lived.

493
00:23:58,200 --> 00:24:02,720
Speaker 3: That Stacey, and we were certain it worked until the

494
00:24:02,760 --> 00:24:06,840
Speaker 3: heat panels didn't work in January of nineteen eighty six.

495
00:24:07,320 --> 00:24:09,399
Speaker 3: Do we have a certitude that the stuff is going

496
00:24:09,480 --> 00:24:13,600
Speaker 3: to work for Nvidia or Microsoft? Are the other fourteen

497
00:24:13,680 --> 00:24:14,520
Speaker 3: names you follow?

498
00:24:15,640 --> 00:24:18,320
Speaker 10: I cover ten names Microsoft and some of the others

499
00:24:18,359 --> 00:24:22,000
Speaker 10: are covered by my college cover semicaracters. When you talk

500
00:24:22,040 --> 00:24:24,879
Speaker 10: about work, like, I don't understand what you mean. But

501
00:24:24,960 --> 00:24:29,440
Speaker 10: the question here is these models look really efficient, and

502
00:24:29,560 --> 00:24:32,159
Speaker 10: the resources that were required to train them, you know,

503
00:24:32,320 --> 00:24:34,640
Speaker 10: look lower than what we've seen from some other models

504
00:24:34,680 --> 00:24:35,880
Speaker 10: in the past. And then even you look at where

505
00:24:35,880 --> 00:24:40,000
Speaker 10: they're actually pricing the stuff, and they're pricing and they're

506
00:24:40,000 --> 00:24:42,400
Speaker 10: pricing it at levels that are quite a bit lower

507
00:24:42,440 --> 00:24:44,760
Speaker 10: where say open ay makes their models available on if

508
00:24:44,760 --> 00:24:46,879
Speaker 10: you want to go use them. And so the question is,

509
00:24:46,880 --> 00:24:48,680
Speaker 10: given all of that that you need less, you need

510
00:24:48,720 --> 00:24:52,639
Speaker 10: fewer GPUs, that's that's the big one, and fewer infrastructure

511
00:24:52,640 --> 00:24:55,320
Speaker 10: buildouts and everything else. And so I mean, just just

512
00:24:55,359 --> 00:24:58,760
Speaker 10: to talk about it's like, what have they done. They've

513
00:24:58,800 --> 00:25:01,280
Speaker 10: got two flavors of models. They've got something called V three,

514
00:25:01,359 --> 00:25:04,400
Speaker 10: which is sort of like a chat model. It uses

515
00:25:04,400 --> 00:25:06,680
Speaker 10: something called a mixture of experts that makes it very efficient.

516
00:25:06,960 --> 00:25:09,480
Speaker 10: And they've got a reasoning model called one where they

517
00:25:09,560 --> 00:25:13,840
Speaker 10: used reinforcement learning and some other techniques from the V

518
00:25:13,880 --> 00:25:18,400
Speaker 10: three base model to build rivals to say to open

519
00:25:18,480 --> 00:25:22,639
Speaker 10: a eyes one reasoning model. If you look at the

520
00:25:23,880 --> 00:25:27,639
Speaker 10: why they're so efficient in terms of the training that

521
00:25:27,680 --> 00:25:29,960
Speaker 10: they did on and the compute resources they used to

522
00:25:29,960 --> 00:25:33,280
Speaker 10: train the V three models, it's very efficient. They compared

523
00:25:33,280 --> 00:25:35,159
Speaker 10: it to say like LAMA four or five B, and

524
00:25:35,200 --> 00:25:37,800
Speaker 10: it took you know, roughly ten percent of the computer

525
00:25:37,880 --> 00:25:41,119
Speaker 10: resource to train it then metas a LAMA four or

526
00:25:41,119 --> 00:25:43,840
Speaker 10: five B did. However, the structure of the model, would

527
00:25:43,880 --> 00:25:46,120
Speaker 10: that make sense? It uses something called a mixture of experts.

528
00:25:46,160 --> 00:25:49,840
Speaker 10: That's the whole point. Other mixture of experts, models that

529
00:25:49,880 --> 00:25:51,920
Speaker 10: are out there, you know, take anywhere from you know,

530
00:25:52,520 --> 00:25:55,080
Speaker 10: a six to one seventh as mount compute resources.

531
00:25:55,119 --> 00:25:55,760
Speaker 2: This is a good one.

532
00:25:55,760 --> 00:25:58,000
Speaker 10: It was like one tent or even even better than that.

533
00:25:58,080 --> 00:26:01,680
Speaker 10: But I don't think it's a miracle. And you got

534
00:26:01,720 --> 00:26:04,840
Speaker 10: to remember, like remember I'm a semiconductor analyst. Like for

535
00:26:05,920 --> 00:26:09,760
Speaker 10: fifty years, costs and semiconductors got cut in half, like

536
00:26:09,800 --> 00:26:11,960
Speaker 10: for every every cheers for fifty years. That was not

537
00:26:12,080 --> 00:26:14,560
Speaker 10: a bad thing for semiconductor man, that was a good thing.

538
00:26:14,600 --> 00:26:19,080
Speaker 10: We all want adoption. To get adoption, we need lower costs.

539
00:26:19,080 --> 00:26:23,119
Speaker 10: This is something called Jevans paradox, right, like increasing efficiency

540
00:26:23,200 --> 00:26:25,760
Speaker 10: drives actually more demand, more net demand, not less demand.

541
00:26:25,840 --> 00:26:28,000
Speaker 10: I'm a firm believer in that. I have a strong

542
00:26:28,080 --> 00:26:30,520
Speaker 10: suspicion that if we can free up more compute resources,

543
00:26:30,520 --> 00:26:32,080
Speaker 10: it will get it.

544
00:26:32,000 --> 00:26:32,800
Speaker 2: Will be used.

545
00:26:33,080 --> 00:26:36,120
Speaker 10: I don't think we're anywhere close to the end of

546
00:26:36,200 --> 00:26:38,720
Speaker 10: compute needs for artificial intelligence. We're gonna need this just

547
00:26:38,720 --> 00:26:39,320
Speaker 10: given I mean.

548
00:26:39,200 --> 00:26:41,760
Speaker 2: The costs have been going up. Right next year, we

549
00:26:41,800 --> 00:26:42,119
Speaker 2: need this.

550
00:26:43,040 --> 00:26:46,200
Speaker 3: But anytime anybody coast William Stanley, Jefvins on the show,

551
00:26:46,240 --> 00:26:47,159
Speaker 3: they get to come back.

552
00:26:47,040 --> 00:26:48,040
Speaker 7: They get to continue.

553
00:26:48,280 --> 00:26:51,000
Speaker 6: Stacey, what do you expect to hear from the semiconductor

554
00:26:51,000 --> 00:26:53,800
Speaker 6: companies that you cover in terms of their response to

555
00:26:53,840 --> 00:26:55,879
Speaker 6: this news, whether it's and video or others.

556
00:26:56,200 --> 00:26:57,440
Speaker 10: Yeah, well, I mean, so we were in the middle

557
00:26:57,440 --> 00:26:59,159
Speaker 10: of earning season right now, so we'll see as they

558
00:26:59,680 --> 00:27:02,520
Speaker 10: come out, right, but we forget the semiconductor comings from it.

559
00:27:02,520 --> 00:27:06,680
Speaker 10: We've already seen, like like some of the other like responses,

560
00:27:06,800 --> 00:27:09,040
Speaker 10: and you have to remember everything that that that we've

561
00:27:09,080 --> 00:27:10,520
Speaker 10: seen deep seaking like this. I don't want to I

562
00:27:10,560 --> 00:27:12,520
Speaker 10: don't want to dismiss it like they're they're really good

563
00:27:12,520 --> 00:27:15,560
Speaker 10: and very efficient, they're very smart engineers. But this is

564
00:27:15,600 --> 00:27:17,560
Speaker 10: these are not things that are not known, and you know,

565
00:27:17,600 --> 00:27:20,160
Speaker 10: these are not things that like the the other top

566
00:27:20,240 --> 00:27:22,840
Speaker 10: tier AI researchers and AI labs are not aware of

567
00:27:23,040 --> 00:27:26,680
Speaker 10: and potentially already using themselves. Right, And if we look

568
00:27:26,720 --> 00:27:28,840
Speaker 10: at what we saw last week on top of all

569
00:27:28,880 --> 00:27:32,199
Speaker 10: of this, clearly spending is going up, not down. So

570
00:27:32,280 --> 00:27:36,159
Speaker 10: Meta just just significantly increased their capex. I think on Friday,

571
00:27:36,480 --> 00:27:39,119
Speaker 10: right we had the whole Stargate announce and five hundred

572
00:27:39,119 --> 00:27:42,320
Speaker 10: billion dollars whatever that is, and even China a few

573
00:27:42,359 --> 00:27:45,280
Speaker 10: days ago announced a one trillion yuan which is about

574
00:27:45,320 --> 00:27:48,520
Speaker 10: one hundred and forty billion dollar US sort of like

575
00:27:48,560 --> 00:27:52,119
Speaker 10: sovereign like AI effort right, so like clearly spending is

576
00:27:52,119 --> 00:27:54,800
Speaker 10: still going up, not down. And and these models, I said,

577
00:27:55,160 --> 00:27:57,560
Speaker 10: they didn't just just dump on us. Yesterday, the Deep

578
00:27:57,560 --> 00:28:00,040
Speaker 10: Seat V three was released the end after just.

579
00:28:00,080 --> 00:28:02,720
Speaker 3: Well, Tracey Stacey, I got to get this in. It's

580
00:28:02,720 --> 00:28:06,840
Speaker 3: too important. At UCLA. You had to survive Physics one

581
00:28:06,960 --> 00:28:10,240
Speaker 3: A a few years ago. The smartest thing I've seen

582
00:28:10,760 --> 00:28:13,080
Speaker 3: is that American AI is trying to go out to

583
00:28:13,160 --> 00:28:16,800
Speaker 3: thirty two decimal points, where the Chinese are saying, you

584
00:28:16,840 --> 00:28:19,000
Speaker 3: know what eight decimal points is?

585
00:28:19,080 --> 00:28:19,560
Speaker 2: Okay?

586
00:28:19,960 --> 00:28:23,520
Speaker 3: Is this a question of granularity or we're being too

587
00:28:23,600 --> 00:28:26,440
Speaker 3: perfect in AI? And the Chinese are saying, you don't

588
00:28:26,480 --> 00:28:28,120
Speaker 3: need to be too perfect.

589
00:28:28,600 --> 00:28:31,160
Speaker 10: Now, so they use what you're trying. They use mixed precision.

590
00:28:35,720 --> 00:28:37,479
Speaker 10: They use mixed precision numbers to train this that they

591
00:28:37,520 --> 00:28:40,080
Speaker 10: used to have floating point eight eight bit numbers you

592
00:28:40,160 --> 00:28:42,720
Speaker 10: typically use floating point sixty or floating point thirty two.

593
00:28:43,280 --> 00:28:46,920
Speaker 10: It's actually really interesting though, in Vidia with their Hopper generation,

594
00:28:47,000 --> 00:28:49,720
Speaker 10: which is the H one D eight hundreds. It has

595
00:28:49,760 --> 00:28:52,600
Speaker 10: something called on it. It's actually called a transformer engine.

596
00:28:53,680 --> 00:28:55,960
Speaker 10: What that actually did was it enabled you to train

597
00:28:56,120 --> 00:28:59,320
Speaker 10: at eight bit precision. And in fact Blackwell, which is

598
00:28:59,360 --> 00:29:02,320
Speaker 10: their their their next generation stuff, this spending on now

599
00:29:02,560 --> 00:29:07,480
Speaker 10: actually enables four bit precision. So no, these are things

600
00:29:07,520 --> 00:29:09,920
Speaker 10: that everybody is looking at. It may it may be

601
00:29:10,080 --> 00:29:12,040
Speaker 10: that that deep Seek is the first ones to actually

602
00:29:12,080 --> 00:29:14,960
Speaker 10: deploy this in a very large model, but these are.

603
00:29:14,960 --> 00:29:15,680
Speaker 2: Things they're looking at it.

604
00:29:15,640 --> 00:29:18,720
Speaker 10: And if you actually look over time, we've seen the

605
00:29:18,800 --> 00:29:21,240
Speaker 10: numerical precisions that people are using coming in and other

606
00:29:21,360 --> 00:29:22,880
Speaker 10: the reason you do this, did you get more compute?

607
00:29:22,880 --> 00:29:24,560
Speaker 10: If I go from sixteen bit to eight in theory,

608
00:29:24,720 --> 00:29:27,480
Speaker 10: I get twice as many flops, twice as many operations

609
00:29:27,480 --> 00:29:29,640
Speaker 10: per second. But no, these are things that have been

610
00:29:29,680 --> 00:29:32,040
Speaker 10: happening all along. This is just the next step. And

611
00:29:32,240 --> 00:29:34,640
Speaker 10: they're again I don't want to dismiss them, Like I said,

612
00:29:34,640 --> 00:29:37,640
Speaker 10: they've done some really clever things and but but none

613
00:29:37,680 --> 00:29:39,480
Speaker 10: of this is a miracle or and none of this

614
00:29:39,760 --> 00:29:42,040
Speaker 10: I think is an unknown or a slap in the

615
00:29:42,040 --> 00:29:45,040
Speaker 10: face to the other AI labs.

616
00:29:45,240 --> 00:29:47,640
Speaker 3: We're out of time. Don't be a stranger thing. You're

617
00:29:47,760 --> 00:29:51,040
Speaker 3: trooper to get up this early. Really really appreciate these.

618
00:29:51,040 --> 00:29:54,200
Speaker 3: In Los Angeles daily lunches at Michael's. They're closed now

619
00:29:54,240 --> 00:29:57,000
Speaker 3: because of the fire, but I'm sure when Michael's reopened,

620
00:29:57,200 --> 00:29:59,959
Speaker 3: you know Stacy will be there. Stacy redskin with burns,

621
00:30:00,280 --> 00:30:03,600
Speaker 3: definitive there.

622
00:30:05,920 --> 00:30:09,840
Speaker 1: This is the Bloomberg Surveillance Podcast. Listen live each weekday

623
00:30:09,880 --> 00:30:13,200
Speaker 1: starting at seven am Eastern on Applecarplay and Android Auto

624
00:30:13,320 --> 00:30:16,120
Speaker 1: with the Bloomberg Business app. You can also watch us

625
00:30:16,160 --> 00:30:20,080
Speaker 1: live every weekday on YouTube and always on the Bloomberg terminal.

626
00:30:20,720 --> 00:30:22,480
Speaker 3: The daily look at the front pages. It's brought to

627
00:30:22,480 --> 00:30:26,640
Speaker 3: you by IBKR. With the average US temperature in January

628
00:30:26,680 --> 00:30:30,760
Speaker 3: twenty twenty five be greater than thirty four degrees. Trade

629
00:30:30,800 --> 00:30:35,000
Speaker 3: your prediction with IBKR forecast Trader that yes was recently

630
00:30:35,040 --> 00:30:39,440
Speaker 3: at fourteen percent. Makes sense by yes or no anterner

631
00:30:39,520 --> 00:30:43,040
Speaker 3: dollar if you're right. Go to ibkr dot com slash

632
00:30:43,440 --> 00:30:50,160
Speaker 3: forecast Interactive Brokers ibkr dot com slash Forecasts.

633
00:30:50,480 --> 00:30:53,520
Speaker 2: We dashed to the newspapers. Lisa tell me we're deep,

634
00:30:53,880 --> 00:30:55,040
Speaker 2: deep seek free.

635
00:30:55,280 --> 00:30:56,920
Speaker 7: We are newspaper free.

636
00:30:56,960 --> 00:30:59,600
Speaker 11: I know, trying to do something a little different. We

637
00:30:59,720 --> 00:31:02,360
Speaker 11: talked about more companies pushing workers back to the office, right,

638
00:31:02,400 --> 00:31:05,560
Speaker 11: but not everyone at the company is being treated the same.

639
00:31:05,640 --> 00:31:07,800
Speaker 11: They're saying. It's causing a lot of tension now in

640
00:31:07,840 --> 00:31:10,480
Speaker 11: the workplace because you have some people in a company,

641
00:31:10,560 --> 00:31:13,760
Speaker 11: like the top performers, the employees with those certain skill sets.

642
00:31:14,080 --> 00:31:18,479
Speaker 11: They're being given that flexibility, and so he comes in

643
00:31:18,520 --> 00:31:21,080
Speaker 11: once in a while because the companies don't want to

644
00:31:21,080 --> 00:31:22,960
Speaker 11: lose them to other companies. You know, it's a whole

645
00:31:22,960 --> 00:31:25,800
Speaker 11: thing of stealing each other's workers. So it's a lot

646
00:31:25,840 --> 00:31:28,040
Speaker 11: of tension in the workplace that's starting to cause attention

647
00:31:28,160 --> 00:31:30,600
Speaker 11: between the managers and the staff on top of it.

648
00:31:30,720 --> 00:31:33,600
Speaker 3: So it's truly fluid, I mean into the warmer season

649
00:31:33,680 --> 00:31:34,840
Speaker 3: into spring.

650
00:31:35,640 --> 00:31:37,840
Speaker 2: To me, it's a hugely fluid debate.

651
00:31:37,960 --> 00:31:39,880
Speaker 7: Some people take Fridays off in the summertime.

652
00:31:39,880 --> 00:31:41,040
Speaker 2: I've heard, well they do.

653
00:31:41,160 --> 00:31:43,280
Speaker 7: Yes, I've heard we missed that memo.

654
00:31:43,760 --> 00:31:44,120
Speaker 2: It works.

655
00:31:44,400 --> 00:31:46,160
Speaker 3: It works out with me in PA because he takes

656
00:31:46,160 --> 00:31:48,520
Speaker 3: every Friday off, and I'm like, you know, you're in

657
00:31:48,640 --> 00:31:49,640
Speaker 3: the fam and all that.

658
00:31:50,200 --> 00:31:50,800
Speaker 7: It works out.

659
00:31:51,240 --> 00:31:53,920
Speaker 11: Nextll does we want to go to the airline industry

660
00:31:53,960 --> 00:31:56,320
Speaker 11: because you're seeing this shift in the pricing power now

661
00:31:56,360 --> 00:31:59,360
Speaker 11: going back to the carriers. If you've been pricing tickets

662
00:31:59,400 --> 00:32:02,640
Speaker 11: out there, notice prices for the cheapest flights up twelve

663
00:32:02,720 --> 00:32:05,320
Speaker 11: percent this month from a year earlier. That's according to Hopper.

664
00:32:05,760 --> 00:32:07,520
Speaker 11: But the airlines are saying people are gonna pay it

665
00:32:07,600 --> 00:32:11,200
Speaker 11: no matter what. And they're saying people are paying it domestic, yes,

666
00:32:11,240 --> 00:32:12,680
Speaker 11: international as well.

667
00:32:13,280 --> 00:32:13,880
Speaker 8: And there.

668
00:32:16,000 --> 00:32:17,760
Speaker 11: I don't know, I don't know how much of a

669
00:32:17,800 --> 00:32:21,280
Speaker 11: deal there you got, But they're saying what's shifting to

670
00:32:21,520 --> 00:32:24,040
Speaker 11: is that the discount airlines. You know, you've seen them

671
00:32:24,600 --> 00:32:27,000
Speaker 11: who usually go to these you know consumer you know,

672
00:32:27,480 --> 00:32:30,440
Speaker 11: budget conscious people. They are starting to shift to the

673
00:32:30,680 --> 00:32:33,080
Speaker 11: to the more advanced. And then you have the luxury sector.

674
00:32:33,120 --> 00:32:35,760
Speaker 11: Now you have all these luxury items being offered out

675
00:32:36,200 --> 00:32:38,320
Speaker 11: on the plane. So if you go on a flight,

676
00:32:38,400 --> 00:32:41,360
Speaker 11: you may have a you know, a cosmetic bag filled

677
00:32:41,360 --> 00:32:46,120
Speaker 11: with all these premium you know, lotions and perfumes. You're

678
00:32:46,160 --> 00:32:50,320
Speaker 11: getting caviars, champagne, cae.

679
00:32:50,680 --> 00:32:51,520
Speaker 4: You have not seen.

680
00:32:53,600 --> 00:32:55,480
Speaker 11: You haven't gotten your crystal champagne.

681
00:32:56,040 --> 00:32:58,960
Speaker 2: What is it Hudson River, Yeah, exactly right.

682
00:32:59,240 --> 00:33:01,040
Speaker 6: I mean it's everything's a cart though you pay for

683
00:33:01,080 --> 00:33:03,320
Speaker 6: everything now, so I mean, you know the bags that

684
00:33:03,560 --> 00:33:05,120
Speaker 6: that that the better seats so.

685
00:33:05,440 --> 00:33:07,320
Speaker 3: We had an offspring at the airport with a lot

686
00:33:07,320 --> 00:33:09,720
Speaker 3: of luggage. Call up and say, I can't believe what

687
00:33:09,760 --> 00:33:13,160
Speaker 3: they're you know, yep, they want it for me right now,

688
00:33:13,280 --> 00:33:15,200
Speaker 3: for three extra bags or.

689
00:33:15,120 --> 00:33:19,120
Speaker 7: Whatever, two weeks in Europe with one check on check it. Yep,

690
00:33:19,600 --> 00:33:20,920
Speaker 7: I don't take it. You don't take a bag. I

691
00:33:20,960 --> 00:33:22,280
Speaker 7: don't take the kids.

692
00:33:22,280 --> 00:33:24,600
Speaker 2: They travels, they take too much stuff today.

693
00:33:24,720 --> 00:33:28,719
Speaker 11: Next, lastly, okay, so you heard about costco is switching

694
00:33:28,760 --> 00:33:32,840
Speaker 11: back from coke back to coke from pepps. Really, yeah,

695
00:33:32,880 --> 00:33:35,560
Speaker 11: that's that's huge for you. I know, Paul, you know

696
00:33:35,760 --> 00:33:36,440
Speaker 11: you're the Coke.

697
00:33:36,240 --> 00:33:38,640
Speaker 6: Pepsi coke person. But I won't die on that hill.

698
00:33:38,800 --> 00:33:39,960
Speaker 6: You will put PEPs in front of me.

699
00:33:40,000 --> 00:33:40,480
Speaker 7: I'll be fine.

700
00:33:40,680 --> 00:33:42,360
Speaker 11: But it's like this huge debate on.

701
00:33:42,320 --> 00:33:46,120
Speaker 3: Someone to explain a Costco on a Saturday morning Cocono Pepsi.

702
00:33:46,360 --> 00:33:46,600
Speaker 4: Yeah.

703
00:33:46,680 --> 00:33:48,920
Speaker 11: People are like Ritt. They want it with their dollar

704
00:33:49,000 --> 00:33:52,520
Speaker 11: fifty hot dog. They want to have their certain soda

705
00:33:52,560 --> 00:33:53,560
Speaker 11: because they're used to it.

706
00:33:53,600 --> 00:33:54,720
Speaker 2: Can you tell the difference?

707
00:33:55,400 --> 00:33:56,160
Speaker 7: I can't wait.

708
00:33:56,160 --> 00:33:59,680
Speaker 2: Wait, you haven't had mix?

709
00:34:00,080 --> 00:34:00,240
Speaker 3: Now.

710
00:34:00,280 --> 00:34:02,840
Speaker 6: I'm the soda person here and I actually have almost

711
00:34:02,880 --> 00:34:04,440
Speaker 6: a year and a half. I put in a ticket

712
00:34:04,520 --> 00:34:06,640
Speaker 6: every day to bring back the soda machines on the

713
00:34:06,680 --> 00:34:08,799
Speaker 6: sixth floor. It's going up to the highest level of

714
00:34:08,800 --> 00:34:09,480
Speaker 6: bloom But if you.

715
00:34:09,400 --> 00:34:11,319
Speaker 11: Did the blind taste test, could you tell you yes,

716
00:34:11,480 --> 00:34:11,759
Speaker 11: you could?

717
00:34:11,800 --> 00:34:12,080
Speaker 4: You can.

718
00:34:12,840 --> 00:34:13,040
Speaker 2: Yeah.

719
00:34:13,040 --> 00:34:15,160
Speaker 6: But again, I'm a coke person, but you put a

720
00:34:15,200 --> 00:34:16,760
Speaker 6: PEPs in front of me just as happy.

721
00:34:17,200 --> 00:34:20,640
Speaker 7: I mean, I don't know anybody else back there coke

722
00:34:20,760 --> 00:34:23,400
Speaker 7: PEPSI I don't know who cares. I don't know.

723
00:34:23,480 --> 00:34:25,680
Speaker 6: But the organic stuff they have on the sixth floor

724
00:34:26,120 --> 00:34:29,480
Speaker 6: has to go. If somebody's listening that can do something

725
00:34:29,480 --> 00:34:30,480
Speaker 6: here at bloom or get rids.

726
00:34:30,600 --> 00:34:33,960
Speaker 7: They have organic soda on the sixth floor. I mean,

727
00:34:34,040 --> 00:34:36,600
Speaker 7: that's un American. It's un America exactly.

728
00:34:36,960 --> 00:34:38,680
Speaker 6: So I'm talking to the highest people in here at

729
00:34:38,680 --> 00:34:39,960
Speaker 6: Bloomberg to get our soda.

730
00:34:39,800 --> 00:34:43,800
Speaker 2: Machines, the newspapers. Lisa Mateo, thank you so much.

731
00:34:45,200 --> 00:34:50,080
Speaker 1: This is the Bloomberg Surveillance podcast, available on Apple, Spotify,

732
00:34:50,200 --> 00:34:53,960
Speaker 1: and anywhere else you get your podcasts. Listen live each

733
00:34:54,000 --> 00:34:57,840
Speaker 1: weekday seven to ten am Easter and on Bloomberg dot com,

734
00:34:57,960 --> 00:35:01,759
Speaker 1: the iHeartRadio app, tune In, and the Bloomberg Business app.

735
00:35:02,080 --> 00:35:05,200
Speaker 1: You can also watch us live every weekday on YouTube

736
00:35:05,480 --> 00:35:07,520
Speaker 1: and always on the Bloomberg terminal