1
00:00:02,520 --> 00:00:13,399
Speaker 1: Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is alive

2
00:00:13,440 --> 00:00:17,239
Speaker 1: from coast to coast with Caroline Hide in New York

3
00:00:17,520 --> 00:00:19,480
Speaker 1: and Eva Low in Sent Francisco.

4
00:00:22,200 --> 00:00:25,160
Speaker 2: This is Bloomberg Tech coming up. The US combust Department

5
00:00:25,200 --> 00:00:28,200
Speaker 2: moves closer to allowing Nvidia to sell its h two

6
00:00:28,280 --> 00:00:29,960
Speaker 2: hundred AI chips to China.

7
00:00:30,120 --> 00:00:33,480
Speaker 3: Plus robotics startup Skilled Ai just closed a one point

8
00:00:33,479 --> 00:00:37,080
Speaker 3: four billion dollar funding round, tripling its valuation in seven months.

9
00:00:37,320 --> 00:00:39,360
Speaker 4: We'll discuss with the CEO Depact PAPAC.

10
00:00:40,040 --> 00:00:42,840
Speaker 2: And Netflix is working to revise its bid for Warner

11
00:00:42,880 --> 00:00:45,960
Speaker 2: Brothers Discovery, shifting to an all cash offer.

12
00:00:46,360 --> 00:00:48,559
Speaker 3: As you see though those names in the red and

13
00:00:48,600 --> 00:00:51,320
Speaker 3: in general the market on the downside, we still got

14
00:00:51,360 --> 00:00:54,280
Speaker 3: geopolitical anxiety, We've still got earnings coming thick and fast,

15
00:00:54,320 --> 00:00:56,200
Speaker 3: and we've still got and as that one hundred off

16
00:00:56,360 --> 00:00:59,280
Speaker 3: by one and a half percent, not so in the

17
00:00:59,280 --> 00:01:01,440
Speaker 3: world of come on to a search for safety a

18
00:01:01,520 --> 00:01:03,920
Speaker 3: bid as the dollar falls. We see two point eight

19
00:01:03,920 --> 00:01:06,600
Speaker 3: percent higher on the bitcoin rally, so maybe a bit

20
00:01:06,640 --> 00:01:09,200
Speaker 3: of catch up with digital gold where we see spot

21
00:01:09,240 --> 00:01:11,760
Speaker 3: gold and a new record high up seven tens percent.

22
00:01:12,000 --> 00:01:15,440
Speaker 3: Copper all important commodity to our world of AI and

23
00:01:15,560 --> 00:01:17,920
Speaker 3: data center zed we ran a new record high on

24
00:01:17,959 --> 00:01:20,199
Speaker 3: an interday basis. We're up five tens percent just coming

25
00:01:20,200 --> 00:01:22,520
Speaker 3: off of that high on the day. But en it's

26
00:01:22,600 --> 00:01:24,640
Speaker 3: notable and shift of this year.

27
00:01:25,840 --> 00:01:29,200
Speaker 2: Right, rules and requirements are changing in the world of

28
00:01:29,240 --> 00:01:33,199
Speaker 2: AI chips. The US Commerce Department out with clarification which

29
00:01:33,240 --> 00:01:35,400
Speaker 2: impacts in video and it's H two hundred that it

30
00:01:35,440 --> 00:01:38,360
Speaker 2: wants to shift to China, but also impacts AMD as well.

31
00:01:38,400 --> 00:01:41,119
Speaker 2: AMD has been trading all over the place this morning,

32
00:01:41,160 --> 00:01:43,640
Speaker 2: at one point notably higher in the session, but in

33
00:01:43,720 --> 00:01:46,720
Speaker 2: Vidia markedly down now trading at its lowest level in

34
00:01:46,760 --> 00:01:49,080
Speaker 2: around a month. What are these new rules and what

35
00:01:49,120 --> 00:01:51,920
Speaker 2: do they mean and why is this significant? Bloomberg Senior

36
00:01:51,960 --> 00:01:54,840
Speaker 2: Tech editor Mike Shepherd joins us out of DC lay

37
00:01:54,840 --> 00:01:57,160
Speaker 2: out the new rules and requirements for US. But what

38
00:01:57,200 --> 00:01:59,200
Speaker 2: we're saying in the Bloomberg story is this marks a

39
00:01:59,280 --> 00:02:02,000
Speaker 2: significant moment for the companies in their efforts to set

40
00:02:02,000 --> 00:02:02,600
Speaker 2: into China.

41
00:02:02,640 --> 00:02:07,280
Speaker 5: Why well, this is a small but critical bureaucratic step

42
00:02:07,600 --> 00:02:10,320
Speaker 5: that paves the way for Nvidia and and D to

43
00:02:10,360 --> 00:02:13,920
Speaker 5: be able to sell H two hundred incomparable AI chips

44
00:02:13,960 --> 00:02:16,840
Speaker 5: to China. What it does is it now says that

45
00:02:16,880 --> 00:02:20,680
Speaker 5: the US government will review requests for export licenses on

46
00:02:20,720 --> 00:02:23,680
Speaker 5: a case by case basis and move away from the

47
00:02:23,720 --> 00:02:26,960
Speaker 5: past policy of a presumption of denial of any such

48
00:02:27,000 --> 00:02:30,040
Speaker 5: request and that had in the past been really an

49
00:02:30,080 --> 00:02:34,400
Speaker 5: effective export ban. Now these companies can't be there expecting

50
00:02:34,480 --> 00:02:37,520
Speaker 5: a rubber stamp for their requests. The US government is

51
00:02:37,560 --> 00:02:41,680
Speaker 5: laying out some pretty specific requirements. One that the companies

52
00:02:41,760 --> 00:02:44,799
Speaker 5: must show that the exports to China will not create

53
00:02:44,840 --> 00:02:47,920
Speaker 5: a supply shortage here. This is addressing, in a way,

54
00:02:48,040 --> 00:02:50,560
Speaker 5: some of the concerns that lawmakers here in Washington have

55
00:02:50,639 --> 00:02:53,360
Speaker 5: been raising with attempts to pass their own version of

56
00:02:53,360 --> 00:02:57,640
Speaker 5: export controls. Two, they must show that any manufacturing done

57
00:02:57,680 --> 00:03:02,040
Speaker 5: for China must not displace any production capacity intended for

58
00:03:02,280 --> 00:03:06,600
Speaker 5: US customers here. And three, the companies must implement strict

59
00:03:06,800 --> 00:03:10,080
Speaker 5: know your Customer procedures to ensure that these chips and

60
00:03:10,120 --> 00:03:13,200
Speaker 5: the related technology don't get out to unauthorized users.

61
00:03:13,560 --> 00:03:15,720
Speaker 3: It's interesting, Mike, that we are seeing Nvidia down on

62
00:03:15,760 --> 00:03:19,200
Speaker 3: the day AIMD. Do we get any sense of how

63
00:03:19,280 --> 00:03:21,440
Speaker 3: many they might be able to ship, Because it's not

64
00:03:21,480 --> 00:03:23,760
Speaker 3: without its intricacies as you say fifty percent of the

65
00:03:23,760 --> 00:03:25,079
Speaker 3: overall made in the US.

66
00:03:26,200 --> 00:03:29,640
Speaker 5: Well, it is the question of the moment surrounding this

67
00:03:29,760 --> 00:03:34,720
Speaker 5: rule just how this cap that it lays out would function.

68
00:03:35,280 --> 00:03:39,400
Speaker 5: The rule says that the exports to China can go

69
00:03:39,520 --> 00:03:43,560
Speaker 5: more than no more than fifty percent of total product

70
00:03:43,600 --> 00:03:46,360
Speaker 5: here in the US. But does that mean H two

71
00:03:46,440 --> 00:03:50,000
Speaker 5: hundreds produced from this day going forward. If that's the case,

72
00:03:50,040 --> 00:03:52,360
Speaker 5: it would not be a very big set and it

73
00:03:52,400 --> 00:03:56,640
Speaker 5: would mean a far smaller number of exports to China,

74
00:03:56,720 --> 00:03:59,680
Speaker 5: and that would certainly not go anywhere near meeting the

75
00:04:00,200 --> 00:04:03,840
Speaker 5: huge demand that Jensen Wang has said that he is

76
00:04:03,880 --> 00:04:07,200
Speaker 5: getting from Chinese customers. And we have reported that Ali

77
00:04:07,240 --> 00:04:09,680
Speaker 5: Baba and Bite the Answer interested in buying as many

78
00:04:09,720 --> 00:04:13,000
Speaker 5: as two hundred thousand chips each. Now, if we are

79
00:04:13,040 --> 00:04:17,200
Speaker 5: talking about all the H two hundreds produce over all time,

80
00:04:17,400 --> 00:04:19,840
Speaker 5: that would be a much bigger universe and that would

81
00:04:19,880 --> 00:04:24,360
Speaker 5: be a much happier outcome for Nvidia, AMD and their investors.

82
00:04:24,400 --> 00:04:27,880
Speaker 5: So we will have to see exactly how that is interpreted.

83
00:04:28,040 --> 00:04:29,240
Speaker 6: If that gets spelled out.

84
00:04:29,400 --> 00:04:32,520
Speaker 5: Another big question Caro is answered, how quickly the government

85
00:04:32,600 --> 00:04:36,400
Speaker 5: can move to actually process these export requests. Because we

86
00:04:36,480 --> 00:04:39,520
Speaker 5: have been hearing over the past year and longer that

87
00:04:39,760 --> 00:04:43,760
Speaker 5: they are simply running into capacity problems in the bureaucracy

88
00:04:43,800 --> 00:04:46,680
Speaker 5: at commerce and being able to turn these around.

89
00:04:46,839 --> 00:04:50,840
Speaker 3: Bottlenecks throughout the world of AI. Mike Sheperd, thanks so much.

90
00:04:51,080 --> 00:04:53,440
Speaker 3: Let's stick with what this means for Nvidia and more

91
00:04:53,480 --> 00:04:56,559
Speaker 3: broadly for semiconductors. Beth Kennick's with us IO fun lead

92
00:04:56,600 --> 00:04:59,679
Speaker 3: tech analyst. So, is this a fifty billion dollar annual

93
00:04:59,680 --> 00:05:00,680
Speaker 3: revenue opportunity?

94
00:05:00,839 --> 00:05:02,239
Speaker 4: Is China that for Nvidia?

95
00:05:03,000 --> 00:05:06,719
Speaker 7: Hey, Caroline jensen Wok has stated it's a fifty billion

96
00:05:07,000 --> 00:05:10,360
Speaker 7: per year annual opportunity. When we left off, it was

97
00:05:10,400 --> 00:05:12,880
Speaker 7: about five billion a quarter. That's twenty five billion, so

98
00:05:12,960 --> 00:05:16,520
Speaker 7: somewhere between twenty five billion and fifty billion, which really

99
00:05:16,560 --> 00:05:19,960
Speaker 7: means that as we look into analyst estimates, they are

100
00:05:20,000 --> 00:05:23,000
Speaker 7: probably too low. And that has been something from the

101
00:05:23,080 --> 00:05:26,320
Speaker 7: very beginning that we have found Alpha in is that

102
00:05:26,360 --> 00:05:29,159
Speaker 7: analysts are really struggling to wrap their head around the

103
00:05:29,200 --> 00:05:33,000
Speaker 7: opportunity of Nvidia. And now here comes back the aged

104
00:05:33,080 --> 00:05:36,919
Speaker 7: two hundreds, and where my firm is very focused is

105
00:05:37,040 --> 00:05:40,400
Speaker 7: calendar year twenty twenty seven. If you look at these estimates,

106
00:05:40,600 --> 00:05:43,480
Speaker 7: they are too low, as it is let alone if

107
00:05:43,560 --> 00:05:44,960
Speaker 7: China revenue resumes.

108
00:05:45,320 --> 00:05:49,320
Speaker 3: What's really interesting, though, is China's response to all of this.

109
00:05:49,960 --> 00:05:51,240
Speaker 4: You say it's too low.

110
00:05:51,320 --> 00:05:54,680
Speaker 3: How much might be the revenue for Nvidia in China.

111
00:05:54,720 --> 00:05:56,760
Speaker 4: But at the same time we understand.

112
00:05:56,320 --> 00:05:59,720
Speaker 3: That deep seats driving new models with seeing them use

113
00:06:00,360 --> 00:06:02,880
Speaker 3: what they have in the different method that maybe Ali

114
00:06:02,960 --> 00:06:04,560
Speaker 3: Barbara and why do you don't need those two hundred

115
00:06:04,560 --> 00:06:06,919
Speaker 3: thousand that they've already put off efforts and offers in

116
00:06:07,000 --> 00:06:08,800
Speaker 3: for what do you make of the way in which

117
00:06:08,880 --> 00:06:11,240
Speaker 3: China is going to pivot to domestic chips?

118
00:06:12,000 --> 00:06:14,839
Speaker 7: I would argue that right now the number I had

119
00:06:14,920 --> 00:06:17,400
Speaker 7: was about two million, h two hundreds and there's only

120
00:06:17,440 --> 00:06:22,040
Speaker 7: about seven hundred thousand available, which means that once again

121
00:06:22,080 --> 00:06:24,920
Speaker 7: in videos oversubscribed. I have been asked before, who is

122
00:06:24,960 --> 00:06:27,800
Speaker 7: the next in Vidia to something like Huawei, you know,

123
00:06:28,000 --> 00:06:30,919
Speaker 7: is it Google's TPUs? And the answer is the next

124
00:06:30,960 --> 00:06:35,320
Speaker 7: in Vidia is in video. They are extremely hard to disrupt,

125
00:06:35,720 --> 00:06:40,800
Speaker 7: and that goes for China or even United States competitors BEV.

126
00:06:40,880 --> 00:06:43,359
Speaker 2: Taking into account that the news story that Mike Shephard

127
00:06:43,440 --> 00:06:46,640
Speaker 2: just brought us, Jensen Wong told us at CS that

128
00:06:46,720 --> 00:06:50,039
Speaker 2: the five hundred billion dollar forecast for five fiscal quarters.

129
00:06:50,080 --> 00:06:53,760
Speaker 2: It includes this one for Blackwell Rubin could get bigger

130
00:06:54,320 --> 00:06:56,919
Speaker 2: because they could factor into H two hundred sales in

131
00:06:56,960 --> 00:06:59,400
Speaker 2: twenty six. I know twenty seven to a year, but

132
00:06:59,440 --> 00:07:02,200
Speaker 2: if you model for that little update that the Jensen

133
00:07:02,240 --> 00:07:02,880
Speaker 2: one gave.

134
00:07:04,080 --> 00:07:07,920
Speaker 7: Yes, so right now the analyst estimates did actually catch up.

135
00:07:07,920 --> 00:07:11,840
Speaker 7: They're about three to twenty billion this calendar year. Let's

136
00:07:11,880 --> 00:07:14,840
Speaker 7: just go with twenty twenty six. I want to really

137
00:07:14,880 --> 00:07:18,200
Speaker 7: emphasize that its next calendar year that we're going to

138
00:07:18,240 --> 00:07:21,480
Speaker 7: start to see analyst estimates too low. That is where

139
00:07:21,480 --> 00:07:24,320
Speaker 7: the alpha is at right now, and it only takes

140
00:07:24,360 --> 00:07:27,280
Speaker 7: one or two quarters for us to really see Nvidia's

141
00:07:27,320 --> 00:07:32,480
Speaker 7: price movement from all of the all of the momentum

142
00:07:32,520 --> 00:07:35,320
Speaker 7: we're seeing not only from Blackwell Ultra, but then we've

143
00:07:35,320 --> 00:07:37,960
Speaker 7: got Vera Rubin second half of this year. Now we

144
00:07:38,040 --> 00:07:41,840
Speaker 7: have the H two hundreds coming back. Where that falls

145
00:07:41,880 --> 00:07:45,240
Speaker 7: specifically in what quarter matters less to me, and what

146
00:07:45,320 --> 00:07:47,920
Speaker 7: matters more is that those analyst estimates are far too low.

147
00:07:48,840 --> 00:07:52,000
Speaker 2: Bes The big story in global markets this morning is

148
00:07:52,320 --> 00:07:56,960
Speaker 2: a blistering rally in metals, particularly copper copper, as S

149
00:07:57,000 --> 00:08:01,320
Speaker 2: and P Global wrote January eight foundational to the expanding

150
00:08:01,440 --> 00:08:04,840
Speaker 2: electricity supply that will power the data center build out,

151
00:08:05,240 --> 00:08:07,760
Speaker 2: and it's now a big growth area in that market.

152
00:08:08,200 --> 00:08:10,840
Speaker 2: How worried are you about that copper in the supply

153
00:08:11,000 --> 00:08:13,720
Speaker 2: chain and how it relates to the power needs of

154
00:08:13,760 --> 00:08:16,400
Speaker 2: the data centers that Nvidia's chips go into.

155
00:08:17,040 --> 00:08:21,240
Speaker 7: These are excellent questions, and across the board resources when

156
00:08:21,280 --> 00:08:24,520
Speaker 7: it comes to this data center expansion are extremely stretched,

157
00:08:25,040 --> 00:08:29,040
Speaker 7: especially anything that supports the electrical grids, such as copper.

158
00:08:30,000 --> 00:08:33,000
Speaker 7: What I would really emphasize is those alternative sources of

159
00:08:33,280 --> 00:08:36,440
Speaker 7: energy is a great way to position going into twenty

160
00:08:36,520 --> 00:08:39,880
Speaker 7: twenty six. One thing to my firm has dug up

161
00:08:40,000 --> 00:08:41,840
Speaker 7: is that in order for us to get to VERA,

162
00:08:42,040 --> 00:08:45,440
Speaker 7: Ruben and then Reuben Ultra power has to precede that

163
00:08:45,840 --> 00:08:48,439
Speaker 7: we have to really get this power to these data

164
00:08:48,440 --> 00:08:52,000
Speaker 7: centers before those systems can really ship in volume and

165
00:08:52,040 --> 00:08:57,360
Speaker 7: be installed. Things like net gas, even solid oxide fuel cells,

166
00:08:57,720 --> 00:09:01,680
Speaker 7: those that do not need great dependency on the electrical

167
00:09:01,720 --> 00:09:04,679
Speaker 7: grid is a wonderful way to position, and copper surging

168
00:09:04,720 --> 00:09:06,160
Speaker 7: today is a reminder of that.

169
00:09:06,679 --> 00:09:09,120
Speaker 4: What's interesting is we've hit in video.

170
00:09:09,200 --> 00:09:12,600
Speaker 3: We're talking about the supply chain therein we're also getting

171
00:09:12,640 --> 00:09:14,040
Speaker 3: the idea that CPUs.

172
00:09:13,600 --> 00:09:14,560
Speaker 4: As well at s US as well.

173
00:09:14,600 --> 00:09:16,160
Speaker 3: And you just think about the moves that we've seen

174
00:09:16,240 --> 00:09:19,760
Speaker 3: yesterday today in Intel, I mean, where do you think

175
00:09:20,240 --> 00:09:22,280
Speaker 3: the glutts or the bottlenecks are really going to be

176
00:09:22,320 --> 00:09:22,679
Speaker 3: hitting hard?

177
00:09:22,720 --> 00:09:22,920
Speaker 4: We are?

178
00:09:23,559 --> 00:09:25,920
Speaker 3: I mean, analysts keep Bank saying Intel and AMDA basically

179
00:09:25,960 --> 00:09:28,080
Speaker 3: sold out for twenty twenty six on the CPU service

180
00:09:28,080 --> 00:09:30,040
Speaker 3: side exactly.

181
00:09:30,080 --> 00:09:32,719
Speaker 7: And what's even more interesting is that AMD was up

182
00:09:32,800 --> 00:09:35,640
Speaker 7: more or last year than Nvidia, and they are such

183
00:09:35,640 --> 00:09:39,040
Speaker 7: a strong CPU player as well as Intel. There are

184
00:09:39,040 --> 00:09:42,960
Speaker 7: many bottlenecks. Caroline networking is another one that comes to mind.

185
00:09:43,559 --> 00:09:46,720
Speaker 7: Of course, energy we just spoke about. Just keep in

186
00:09:46,760 --> 00:09:50,160
Speaker 7: mind that it's not just GPUs, but there's almost an

187
00:09:50,200 --> 00:09:53,160
Speaker 7: equal amount of money being spent on all of the components,

188
00:09:53,320 --> 00:09:56,800
Speaker 7: all of the networking, all of the power, and we

189
00:09:56,920 --> 00:10:00,640
Speaker 7: are no longer compute constraint, We're no longer GPU constraint.

190
00:10:00,920 --> 00:10:03,400
Speaker 7: It really is becoming a power problem and a networking

191
00:10:03,400 --> 00:10:06,040
Speaker 7: problem that needs to be solved, especially as we move

192
00:10:06,080 --> 00:10:08,960
Speaker 7: into the inference market, which is only going to require

193
00:10:09,280 --> 00:10:13,880
Speaker 7: more memory, more bandwidth, less latency, and more and more energy.

194
00:10:14,920 --> 00:10:18,200
Speaker 2: Beth do AMD and Intel just accept that it's the

195
00:10:18,240 --> 00:10:21,319
Speaker 2: CPU market where they can do battle and just let

196
00:10:21,360 --> 00:10:24,600
Speaker 2: the accelertor market go to Nvidia and other specialists.

197
00:10:25,400 --> 00:10:27,120
Speaker 4: No, I do not believe that will be the outcome.

198
00:10:27,240 --> 00:10:31,640
Speaker 7: I think AMD is a strong contender, a strong number two,

199
00:10:32,240 --> 00:10:35,240
Speaker 7: especially in the second half of this year. They've really

200
00:10:35,280 --> 00:10:38,679
Speaker 7: been preparing to bring the rack scale architecture to market,

201
00:10:38,880 --> 00:10:42,640
Speaker 7: which is called Helios and ED what was I've covered

202
00:10:42,679 --> 00:10:45,439
Speaker 7: AMD very thoroughly and we could get into the specs

203
00:10:45,480 --> 00:10:46,120
Speaker 7: all day long.

204
00:10:46,400 --> 00:10:48,600
Speaker 4: But open AI giving them.

205
00:10:48,480 --> 00:10:51,640
Speaker 7: That six gigawad deal is a massive nod. I mean

206
00:10:51,640 --> 00:10:54,520
Speaker 7: that is a big nod that AMD is a contender

207
00:10:54,800 --> 00:10:56,680
Speaker 7: from the leading R and D firm.

208
00:10:57,480 --> 00:11:00,280
Speaker 2: Helios, AMD's first rack scale solution, but then in the

209
00:11:00,280 --> 00:11:03,120
Speaker 2: world's first two nanimeter chip inside, which they talked up

210
00:11:03,120 --> 00:11:05,800
Speaker 2: in our interview last week. Be f Kindig the IO Fund.

211
00:11:05,800 --> 00:11:08,120
Speaker 2: Great to have you back on Bloomberg Tech. Thank you

212
00:11:08,240 --> 00:11:12,280
Speaker 2: very much. Now coming up, anger grows over sexually explicit,

213
00:11:12,480 --> 00:11:17,480
Speaker 2: non consensual AI generated images on x We have the latest.

214
00:11:17,960 --> 00:11:18,920
Speaker 6: This is Bloomberg Tech.

215
00:11:30,080 --> 00:11:33,480
Speaker 3: The Supreme Court still hasn't ruled on challenges to President

216
00:11:33,480 --> 00:11:36,560
Speaker 3: Trump's tariffs, leaving the world to wait until at least

217
00:11:36,640 --> 00:11:39,280
Speaker 3: next week to learn the fate of visits signature economic

218
00:11:39,280 --> 00:11:42,160
Speaker 3: policy from a massurveillance. Co host Amriy Horden joins us

219
00:11:42,200 --> 00:11:45,240
Speaker 3: right here for more. So we delay once again, and

220
00:11:45,280 --> 00:11:47,720
Speaker 3: that timing could be awkward. It could be awkward. So

221
00:11:47,760 --> 00:11:49,160
Speaker 3: Supreme Court punts it once again.

222
00:11:49,200 --> 00:11:50,760
Speaker 8: They do not come out with the decision on whether

223
00:11:50,840 --> 00:11:53,280
Speaker 8: or not the majority of Actually a lot of this

224
00:11:53,400 --> 00:11:55,960
Speaker 8: terriphone venue we're bringing in into the United States, more

225
00:11:56,000 --> 00:12:00,679
Speaker 8: than seventies percent, is tethered to the president using aiber flexible,

226
00:12:00,720 --> 00:12:02,920
Speaker 8: a blunt tool that he has used since he came

227
00:12:02,960 --> 00:12:05,360
Speaker 8: back into office. And why it could be so awkward

228
00:12:05,400 --> 00:12:07,280
Speaker 8: is because the President will be in Davos next week

229
00:12:07,280 --> 00:12:09,520
Speaker 8: and he's going to be meeting with global leaders, the

230
00:12:09,559 --> 00:12:15,240
Speaker 8: world elite. Last year he joined on remote really Lambasstard,

231
00:12:15,360 --> 00:12:19,000
Speaker 8: the global elite. And at this point, potentially we could

232
00:12:19,000 --> 00:12:22,199
Speaker 8: get the decision next Tuesday or Wednesday, because that's when the.

233
00:12:22,240 --> 00:12:23,280
Speaker 4: Justices will be meeting again.

234
00:12:23,440 --> 00:12:25,920
Speaker 8: That's when potentially we can get the decision and he

235
00:12:26,200 --> 00:12:29,760
Speaker 8: could potentially be facing a Supreme Court saying you cannot

236
00:12:29,840 --> 00:12:33,200
Speaker 8: use that legal authority while he's talking to all these

237
00:12:33,240 --> 00:12:35,199
Speaker 8: countries that he already did trade deals with.

238
00:12:36,480 --> 00:12:39,880
Speaker 2: Bloomberg zamree, Howden, thank you very much. Another top story,

239
00:12:39,960 --> 00:12:43,720
Speaker 2: the US Senate unanimously passed a bill allowing victims of

240
00:12:43,800 --> 00:12:48,160
Speaker 2: sexually explicit AI images to sue the creators of that content.

241
00:12:48,520 --> 00:12:51,040
Speaker 2: The move is in response to growing public anger after

242
00:12:51,040 --> 00:12:54,640
Speaker 2: Bloomberg reported the Lelon musk x has become a top

243
00:12:54,720 --> 00:12:58,000
Speaker 2: site for generating pictures of people who have been non

244
00:12:58,040 --> 00:13:01,680
Speaker 2: consensually undressed by a must post it on x that

245
00:13:01,720 --> 00:13:05,960
Speaker 2: he wasn't aware of any naked underage images generated by Grock.

246
00:13:06,320 --> 00:13:10,640
Speaker 2: That's Xai's generative AI tool. Let's bring in Bloomberg's Emily Burnbaum,

247
00:13:10,800 --> 00:13:15,840
Speaker 2: who covers corporate lobbying government. The Defiance Act confers on

248
00:13:16,000 --> 00:13:21,719
Speaker 2: American citizens a civil right to sue, take that information

249
00:13:21,960 --> 00:13:23,920
Speaker 2: and give the rest of the details. Because that's the

250
00:13:23,960 --> 00:13:25,679
Speaker 2: bit that's new here.

251
00:13:26,520 --> 00:13:31,120
Speaker 9: Yes, that's exactly what it does. It essentially empowers victims

252
00:13:31,320 --> 00:13:34,960
Speaker 9: of this non consensual deep fake technology, people who have

253
00:13:35,040 --> 00:13:38,480
Speaker 9: been non consensually undressed by this AI tool, and it

254
00:13:38,520 --> 00:13:42,080
Speaker 9: allows them to sue the people responsible for posting that.

255
00:13:42,480 --> 00:13:45,920
Speaker 9: And it's a companion to a bill, to a bill

256
00:13:45,960 --> 00:13:48,640
Speaker 9: that was signed into law last year by President Trump

257
00:13:49,400 --> 00:13:53,679
Speaker 9: that creates criminal penalties for distributing that kind of content.

258
00:13:54,240 --> 00:13:58,440
Speaker 3: Ah, it didn't go through the House last time, the

259
00:13:58,480 --> 00:14:02,160
Speaker 3: Defiance Act, whereas take Down did, And why would this

260
00:14:02,240 --> 00:14:03,439
Speaker 3: time be any different?

261
00:14:03,640 --> 00:14:06,040
Speaker 4: Is the pushback becoming enough.

262
00:14:05,840 --> 00:14:08,959
Speaker 3: Because thus far, all we've really seen is the Department

263
00:14:09,000 --> 00:14:12,960
Speaker 3: of Defense put groquai within its military system instead of

264
00:14:12,960 --> 00:14:16,320
Speaker 3: pushing back on this being something that Grocco X needs

265
00:14:16,360 --> 00:14:19,680
Speaker 3: to camp down in some way, that's true.

266
00:14:20,240 --> 00:14:23,280
Speaker 9: There definitely is more of an appetite in Congress to

267
00:14:23,520 --> 00:14:28,560
Speaker 9: pass this legislation. There have been six new co sponsors

268
00:14:28,560 --> 00:14:30,760
Speaker 9: that have come on board since the beginning of the year.

269
00:14:31,560 --> 00:14:36,440
Speaker 9: I think that the issue of non consensual AI generated

270
00:14:36,520 --> 00:14:41,720
Speaker 9: images has become so vivid to lawmakers. Many of them,

271
00:14:42,200 --> 00:14:45,280
Speaker 9: like Alexandria Ocasio Cortez, are themselves the victims of this.

272
00:14:45,520 --> 00:14:48,040
Speaker 9: So I think the appetite has really broadened this year.

273
00:14:49,120 --> 00:14:53,480
Speaker 2: Emily, where does the company X and the platform X

274
00:14:53,520 --> 00:14:57,840
Speaker 2: fit into this in a citizen's right to sue and

275
00:14:57,880 --> 00:14:59,480
Speaker 2: those that distribute the content.

276
00:15:01,080 --> 00:15:05,000
Speaker 9: So the Defiance Act, which passed the Sunate this week,

277
00:15:05,160 --> 00:15:09,440
Speaker 9: that probably would not impact GROC, that impacts the people

278
00:15:09,480 --> 00:15:12,920
Speaker 9: who are responsible for distributing the images. But there is

279
00:15:12,960 --> 00:15:15,000
Speaker 9: a lot of conversation right now about whether the Take

280
00:15:15,040 --> 00:15:19,840
Speaker 9: It Down Act would empower Attorney's General to sue Grock

281
00:15:20,120 --> 00:15:26,520
Speaker 9: Xai Elon Musk over some level of responsibility for you know,

282
00:15:26,680 --> 00:15:29,160
Speaker 9: thousands of these images proliferating every hour.

283
00:15:29,960 --> 00:15:31,840
Speaker 4: So far, no Attorney's.

284
00:15:31,360 --> 00:15:35,560
Speaker 9: General have taken that up, but it's definitely under serious consideration.

285
00:15:35,640 --> 00:15:38,360
Speaker 3: Certainly something that maybe Malaysia has been putting forward about

286
00:15:38,360 --> 00:15:41,400
Speaker 3: whether or not it's Xai's responsibility in some way. Blueing

287
00:15:41,400 --> 00:15:44,640
Speaker 3: Magzemine van Bound, fantastic reporting, thank you very much. Indeed,

288
00:15:45,080 --> 00:15:48,920
Speaker 3: not coming up. ABMB has a new chief technology officer.

289
00:15:49,040 --> 00:15:50,840
Speaker 3: We'll hear about his plans for the short term mental

290
00:15:50,840 --> 00:15:52,080
Speaker 3: company that's next is.

291
00:15:52,000 --> 00:15:52,720
Speaker 4: A body meg Tech.

292
00:16:05,240 --> 00:16:08,480
Speaker 10: I'm extremely excited, like Brian said, to bring craft to

293
00:16:08,960 --> 00:16:12,480
Speaker 10: the art of deploying AI first features and products. So

294
00:16:13,320 --> 00:16:15,400
Speaker 10: I hope to be able to surprise the world with

295
00:16:15,520 --> 00:16:18,240
Speaker 10: really beautiful features that they're going to love using every day.

296
00:16:19,640 --> 00:16:23,720
Speaker 2: That was Airbnb's new CTO, Armadal Dala, weighing in on

297
00:16:23,760 --> 00:16:26,720
Speaker 2: the short term rental platforms AI efforts. Let's get more

298
00:16:26,880 --> 00:16:29,560
Speaker 2: wherever and Latterly Lung, who spoke with our Dala and

299
00:16:29,640 --> 00:16:32,960
Speaker 2: Airbnb CEO Brian Chesky. I think an interesting base start

300
00:16:33,000 --> 00:16:35,400
Speaker 2: is what is Airbnb trying to solve for here by

301
00:16:35,400 --> 00:16:37,200
Speaker 2: bringing in a new CTO. What is it they want

302
00:16:37,200 --> 00:16:38,680
Speaker 2: to change about the platform.

303
00:16:39,080 --> 00:16:41,600
Speaker 11: So after a few years of sort of working on

304
00:16:41,680 --> 00:16:45,000
Speaker 11: the plumbing of Airmb's app, like vamping the app, they

305
00:16:45,000 --> 00:16:48,480
Speaker 11: are ready to take it forward, adding AI across the app,

306
00:16:48,560 --> 00:16:51,520
Speaker 11: including customer service with which they did last year, and

307
00:16:51,560 --> 00:16:53,360
Speaker 11: now they're going to do more with search as well,

308
00:16:53,400 --> 00:16:56,840
Speaker 11: like letting people discover listings and also cross markets who

309
00:16:56,840 --> 00:17:00,240
Speaker 11: have their new services like tours and experiences going forward.

310
00:17:00,720 --> 00:17:02,760
Speaker 3: And I think you sort of put that exact question

311
00:17:02,800 --> 00:17:04,800
Speaker 3: to Brancheski himself. Let's just take a listen to how

312
00:17:04,840 --> 00:17:08,120
Speaker 3: he responded as to why they're pivoting and focusing all

313
00:17:08,119 --> 00:17:09,760
Speaker 3: in on AI at this moment, and take a listen.

314
00:17:10,720 --> 00:17:13,760
Speaker 12: We are at the beginning of this incredibly exciting technological

315
00:17:13,800 --> 00:17:16,600
Speaker 12: transformation with AI. It's going to be a journey. And

316
00:17:16,680 --> 00:17:19,320
Speaker 12: I couldn't be more excited about Ahmed because he's one

317
00:17:19,320 --> 00:17:21,480
Speaker 12: of the leading AI experts in the world, but he

318
00:17:21,560 --> 00:17:24,240
Speaker 12: also brings a sense of craft Airbnb, something we're really

319
00:17:24,280 --> 00:17:26,120
Speaker 12: known for. And I think what you're going to see

320
00:17:26,119 --> 00:17:29,640
Speaker 12: in Airbnb is not only AI search, but a very

321
00:17:29,680 --> 00:17:32,520
Speaker 12: personalized AI experience where I think we're going to bring

322
00:17:32,600 --> 00:17:36,400
Speaker 12: great design sensibility into the experience of Airbnb.

323
00:17:38,000 --> 00:17:40,879
Speaker 3: Let's talk a little bit about how they're talking of

324
00:17:40,880 --> 00:17:44,159
Speaker 3: the narrative of change because they've said farewell to the

325
00:17:44,160 --> 00:17:47,160
Speaker 3: previous CTO who's there for about seven years, and brought

326
00:17:47,200 --> 00:17:49,920
Speaker 3: in this expert from Meta. So who do they say

327
00:17:49,920 --> 00:17:52,320
Speaker 3: you farewell to and why is this Meta executed the

328
00:17:52,359 --> 00:17:52,879
Speaker 3: best place?

329
00:17:54,280 --> 00:17:57,760
Speaker 11: So Aribalo, the previous CTO came in during our time

330
00:17:57,800 --> 00:18:01,560
Speaker 11: when Airbnb was going through hypergrowth. There was seven more

331
00:18:01,560 --> 00:18:04,320
Speaker 11: than seven years ago. And now Airbnb is sort of,

332
00:18:04,760 --> 00:18:07,919
Speaker 11: as Cheskey said, going through a technological revolution of you

333
00:18:07,960 --> 00:18:11,480
Speaker 11: know lms and chat agents, and now Airbnb sort of

334
00:18:11,520 --> 00:18:13,840
Speaker 11: wants to use sort of the unique data they have

335
00:18:13,960 --> 00:18:17,399
Speaker 11: on the platform to differentiate itself from other you know,

336
00:18:17,560 --> 00:18:21,880
Speaker 11: chatbots to provide better recommendations to customers on their platform.

337
00:18:22,440 --> 00:18:26,080
Speaker 2: Abnb the company who stock growse about three percent last year,

338
00:18:26,600 --> 00:18:28,560
Speaker 2: down a couple of percent so far this year.

339
00:18:29,000 --> 00:18:30,200
Speaker 6: And Cheski was one of those.

340
00:18:30,040 --> 00:18:34,800
Speaker 2: People that was the early vocal behind founder mode it.

341
00:18:34,880 --> 00:18:36,920
Speaker 2: What's the health of Airbnb right now. Is this them

342
00:18:36,960 --> 00:18:39,280
Speaker 2: saying we've got to do something to re accelerate.

343
00:18:39,560 --> 00:18:44,040
Speaker 11: Yeah, Brian was not satisfied with the below ten percent

344
00:18:44,119 --> 00:18:46,520
Speaker 11: growth rate we've seen last year, and so he's been

345
00:18:46,640 --> 00:18:51,840
Speaker 11: very active in wanting to pilot more businesses for this year.

346
00:18:52,240 --> 00:18:55,880
Speaker 11: They're piloting boutique hotels and server cities, and they're trying

347
00:18:55,920 --> 00:18:59,359
Speaker 11: to add grocery delivery for people to pre stock their

348
00:18:59,440 --> 00:19:01,960
Speaker 11: kitchens before or are they check in. So they're trying

349
00:19:01,960 --> 00:19:04,240
Speaker 11: to get their growth free BACKRUP as sort of the

350
00:19:04,359 --> 00:19:06,200
Speaker 11: core rental service for chors.

351
00:19:06,560 --> 00:19:06,720
Speaker 4: Now.

352
00:19:06,760 --> 00:19:08,400
Speaker 3: So you've also got a story about how the chief

353
00:19:08,400 --> 00:19:11,120
Speaker 3: information officer has departed after the CTO shakeup.

354
00:19:11,359 --> 00:19:13,360
Speaker 4: Is there more executive leadership changes to come?

355
00:19:14,600 --> 00:19:16,800
Speaker 11: Yeah, so there would be. They're still looking for the

356
00:19:16,920 --> 00:19:20,960
Speaker 11: new CIO. We can expect that to come in in

357
00:19:21,000 --> 00:19:23,119
Speaker 11: the months in the year ahead.

358
00:19:23,600 --> 00:19:26,480
Speaker 3: Most Natalie Lung, thank you very much. Indeed, let's turn

359
00:19:26,480 --> 00:19:30,040
Speaker 3: out attention to entertainment now. Netflix is working to revise

360
00:19:30,160 --> 00:19:33,439
Speaker 3: it's been for Warner Brothers Discovery, potentially shifting to an

361
00:19:33,440 --> 00:19:36,240
Speaker 3: all cash offer, according to sources. Now the moves aimed

362
00:19:36,240 --> 00:19:38,640
Speaker 3: at speeding up a deal that lookers face pushback from

363
00:19:38,720 --> 00:19:43,360
Speaker 3: Washington and competition from rival, bit of Paramounts guidance for more, Let's.

364
00:19:43,200 --> 00:19:43,760
Speaker 4: Get to the new most.

365
00:19:43,800 --> 00:19:47,159
Speaker 3: Lucas Shaw, who covers media and entertainment, hit this story

366
00:19:47,160 --> 00:19:51,400
Speaker 3: hard and Lucas, I mean the all cash narrative. Why

367
00:19:51,480 --> 00:19:53,880
Speaker 3: is that important? Why would that be the winning formula

368
00:19:53,920 --> 00:19:54,520
Speaker 3: for Netflix?

369
00:19:56,080 --> 00:20:00,320
Speaker 13: Because investors like cash now versus stock, and that's that

370
00:20:00,359 --> 00:20:01,480
Speaker 13: they could sell later.

371
00:20:01,600 --> 00:20:01,760
Speaker 11: Right.

372
00:20:02,440 --> 00:20:05,440
Speaker 13: One of the knocks against the Netflix deal as perpetuated

373
00:20:05,440 --> 00:20:08,240
Speaker 13: at least by by Paramounts guidance. The rival bidder has

374
00:20:08,240 --> 00:20:11,159
Speaker 13: been that Paramount's deal is all cash, and that Netflix

375
00:20:11,240 --> 00:20:14,600
Speaker 13: deal contains you know, between four and five dollars a.

376
00:20:14,640 --> 00:20:15,440
Speaker 6: Share of stock.

377
00:20:15,520 --> 00:20:17,960
Speaker 13: Now, Netflix stock is pretty valuable, something that you think

378
00:20:18,000 --> 00:20:20,320
Speaker 13: people would want to have, but most people would rather

379
00:20:20,400 --> 00:20:23,080
Speaker 13: just take the cash now and get paid. And so

380
00:20:23,800 --> 00:20:26,720
Speaker 13: that I think Netflix hopes will accelerate on many fronts.

381
00:20:26,720 --> 00:20:26,760
Speaker 11: It.

382
00:20:27,160 --> 00:20:30,120
Speaker 13: You know, it'll assuage any shareholder concerns about their bid

383
00:20:30,200 --> 00:20:34,440
Speaker 13: and also eliminate any potential complexities as they continue to

384
00:20:34,920 --> 00:20:35,919
Speaker 13: negotiate and pursue it.

385
00:20:37,119 --> 00:20:39,359
Speaker 2: Lucas, when we were reporting this out, you know, the

386
00:20:39,440 --> 00:20:42,359
Speaker 2: Netflix response and the tit for tat, you know, Paramounts

387
00:20:42,359 --> 00:20:44,760
Speaker 2: suing earlier this week and nominated our It's to the board.

388
00:20:45,040 --> 00:20:47,600
Speaker 2: Do you get the sense that Netflix think this will work,

389
00:20:48,440 --> 00:20:51,320
Speaker 2: that they now are back in sort of the lead

390
00:20:51,359 --> 00:20:53,240
Speaker 2: in a tight race for the assets.

391
00:20:54,280 --> 00:20:54,879
Speaker 6: Well, I don't.

392
00:20:55,280 --> 00:20:57,040
Speaker 13: I think if you were to talk to the people

393
00:20:57,040 --> 00:20:59,600
Speaker 13: at Netflix, they've never felt they were out of the lead,

394
00:20:59,720 --> 00:21:03,639
Speaker 13: right know, they are obviously facing a stiff challenge from Paramount,

395
00:21:04,119 --> 00:21:07,840
Speaker 13: but they're the ones who have an approved deal, and

396
00:21:07,880 --> 00:21:11,760
Speaker 13: they have yet to see kind of some they have

397
00:21:11,840 --> 00:21:14,160
Speaker 13: yet to see shareholders say and masks we don't want

398
00:21:14,160 --> 00:21:15,879
Speaker 13: the Netflix deal. They have yet to see the federal

399
00:21:15,920 --> 00:21:18,000
Speaker 13: government say we're going to block this deal. There are

400
00:21:18,119 --> 00:21:21,080
Speaker 13: certainly some kind of blinking yellow lights, but I don't

401
00:21:21,080 --> 00:21:23,880
Speaker 13: think they've they've seen anything yet that that gives makes

402
00:21:23,920 --> 00:21:26,240
Speaker 13: them too concerned. They seem still pretty confident that they'll

403
00:21:26,240 --> 00:21:26,600
Speaker 13: get this.

404
00:21:26,600 --> 00:21:29,680
Speaker 3: Done and in a week January twenty first, that's when

405
00:21:29,680 --> 00:21:33,680
Speaker 3: the tend to offer for certainly the Paramount Skuydance side

406
00:21:33,680 --> 00:21:35,960
Speaker 3: of the equation runs out, right. I mean, do you

407
00:21:35,960 --> 00:21:38,080
Speaker 3: get any signal for investors that they were tool interested

408
00:21:38,119 --> 00:21:38,800
Speaker 3: in that side?

409
00:21:39,880 --> 00:21:43,880
Speaker 13: You know, investors have been split some of the big

410
00:21:43,880 --> 00:21:46,560
Speaker 13: investor investment funds, you know, they don't say anything, but

411
00:21:46,560 --> 00:21:48,919
Speaker 13: we have spoken with a number of the major shareholders

412
00:21:48,920 --> 00:21:52,040
Speaker 13: and I'd say their reactions range from we prefer Netflix,

413
00:21:52,240 --> 00:21:54,399
Speaker 13: do we prefer Paramount, And then a lot of people

414
00:21:54,400 --> 00:21:57,120
Speaker 13: in the middle are basically saying, you know, tie right

415
00:21:57,160 --> 00:21:59,360
Speaker 13: now goes to the winner, which is Netflix. But we

416
00:21:59,480 --> 00:22:03,160
Speaker 13: hope Slash expect that Paramount will increase its offer, which

417
00:22:03,240 --> 00:22:04,760
Speaker 13: Lucas is a pretty clear Paramount move.

418
00:22:06,080 --> 00:22:08,280
Speaker 2: Just very quickly remind us of the state of play.

419
00:22:08,440 --> 00:22:10,000
Speaker 2: The difference between the two bids.

420
00:22:11,160 --> 00:22:15,000
Speaker 13: Well, Netflix's bid is for just the studio and streaming service,

421
00:22:15,160 --> 00:22:19,320
Speaker 13: so Warner Brothers and HBO. Paramount's bid is for the

422
00:22:19,359 --> 00:22:22,280
Speaker 13: whole company, so that includes all the cable networks like CNNT, N,

423
00:22:22,320 --> 00:22:23,399
Speaker 13: T TBS and the like.

424
00:22:24,560 --> 00:22:27,280
Speaker 2: Bloombos Lucas Shaw, who leads the team at screen time,

425
00:22:27,359 --> 00:22:29,080
Speaker 2: that has led the way on reporting this story.

426
00:22:29,119 --> 00:22:31,600
Speaker 6: Thank you so much. Now coming up on Bloomberg.

427
00:22:31,280 --> 00:22:34,960
Speaker 2: Tech, robotics startup Skilled Ai just closed a one point

428
00:22:34,960 --> 00:22:39,440
Speaker 2: four billion dollar funding round, a big valuation, tripling.

429
00:22:39,440 --> 00:22:43,160
Speaker 6: From a year. This is next. This is Bloomberg Tech.

430
00:23:02,920 --> 00:23:04,879
Speaker 4: Welcome back to Bloomberg Tech. Let's check in on the

431
00:23:04,960 --> 00:23:06,159
Speaker 4: markets that are under pressure.

432
00:23:06,160 --> 00:23:08,359
Speaker 3: Look we've got a wall of warriors to come to

433
00:23:08,400 --> 00:23:12,119
Speaker 3: the earning season geopolitical risks, and we see some dragged

434
00:23:12,119 --> 00:23:13,560
Speaker 3: down and big tech. We're off by one point few

435
00:23:13,600 --> 00:23:15,359
Speaker 3: percent on the NASDAK. In fact, two straight days of

436
00:23:15,359 --> 00:23:17,800
Speaker 3: losses on the SMP. We're coming off of record highs

437
00:23:17,920 --> 00:23:19,879
Speaker 3: in video, of course, confronting that news that maybe EAH

438
00:23:19,880 --> 00:23:22,639
Speaker 3: two hundreds could be getting closer to exporting to China.

439
00:23:22,880 --> 00:23:24,520
Speaker 3: But at the moment we're under pressure. It's one of

440
00:23:24,520 --> 00:23:26,800
Speaker 3: the key drags on the downside. Bitcoin though, of three

441
00:23:26,840 --> 00:23:28,920
Speaker 3: point four percent. We're seeing maybe a bit of love

442
00:23:28,960 --> 00:23:31,640
Speaker 3: for digital gold, not just gold, silver and copper. We're

443
00:23:31,680 --> 00:23:34,159
Speaker 3: also seeing maybe some shorts getting squeezed out. Move on,

444
00:23:34,359 --> 00:23:36,239
Speaker 3: have a look at what's happening in individual stocks. I'm

445
00:23:36,280 --> 00:23:38,480
Speaker 3: trying to light on Honeywell, look, we're at one point

446
00:23:38,480 --> 00:23:42,640
Speaker 3: eight percent. Maybe an asset that it controls quantinum could

447
00:23:42,680 --> 00:23:45,520
Speaker 3: be looking for an IPO. We're an understanding that they're

448
00:23:45,560 --> 00:23:49,119
Speaker 3: potentially filing for an IPO with the SEC. We know

449
00:23:49,200 --> 00:23:51,359
Speaker 3: that this last valuation was about eleven billion dollars Z

450
00:23:51,440 --> 00:23:53,680
Speaker 3: We've got other key strategics and videos in their GP

451
00:23:53,760 --> 00:23:56,000
Speaker 3: Morgans in there as well, so all eyes on quantum

452
00:23:56,040 --> 00:23:58,160
Speaker 3: computing as we look towards twenty twenty six.

453
00:23:57,960 --> 00:23:58,680
Speaker 4: What are you looking at?

454
00:23:59,119 --> 00:24:01,360
Speaker 6: Yeah, we also have a big deal in private markets.

455
00:24:01,440 --> 00:24:04,600
Speaker 2: Robotic startup Skilled Ai has just closed a one point

456
00:24:04,640 --> 00:24:08,320
Speaker 2: four billion dollar funding around the valuation more than fourteen

457
00:24:08,440 --> 00:24:11,320
Speaker 2: billion dollars and that's for a company only founded in

458
00:24:11,359 --> 00:24:13,960
Speaker 2: twenty twenty three. Skilled is aiming to stand out in

459
00:24:14,000 --> 00:24:18,240
Speaker 2: what's a crowded field by building an AI powered robotic brain.

460
00:24:18,880 --> 00:24:20,960
Speaker 2: The goal is to develop a system that can adapt

461
00:24:21,040 --> 00:24:24,879
Speaker 2: across environments and form factors and crucially learn from its

462
00:24:24,960 --> 00:24:28,480
Speaker 2: own mistakes. Joining us to discuss a Skilled co founder

463
00:24:28,520 --> 00:24:31,119
Speaker 2: and CEO, Deepak Pathak, It's great to have you on

464
00:24:31,119 --> 00:24:35,080
Speaker 2: Bloomberg Tech. What's interesting. Let's start with the round. It's

465
00:24:35,160 --> 00:24:38,080
Speaker 2: not just the volume, it's not just the valuation. Some

466
00:24:38,200 --> 00:24:41,520
Speaker 2: of the strategics that are backing you in Nvidia, others

467
00:24:41,520 --> 00:24:45,320
Speaker 2: in consumer electronics, Samsung LG and in the software sector

468
00:24:45,359 --> 00:24:49,000
Speaker 2: as well. What are they recognizing is your core competence?

469
00:24:49,040 --> 00:24:50,479
Speaker 6: Do you think so?

470
00:24:51,800 --> 00:24:53,680
Speaker 14: If you're thinking about the robotics today is one of

471
00:24:53,720 --> 00:24:58,040
Speaker 14: the oldest areas in technology. In EI was big for robotics.

472
00:24:58,560 --> 00:25:01,200
Speaker 14: We have seen amazing they're more robotics for the last

473
00:25:01,440 --> 00:25:04,439
Speaker 14: seventy years, not four or five years, seventy years. But

474
00:25:04,480 --> 00:25:07,679
Speaker 14: we are robots around us. And this is what people

475
00:25:07,760 --> 00:25:10,440
Speaker 14: often get wrong. When you think of robots in your mind,

476
00:25:10,760 --> 00:25:13,560
Speaker 14: you imagine hardware like that comes to a point in

477
00:25:13,560 --> 00:25:15,960
Speaker 14: your mind first, because that's how Hollywood has primed us.

478
00:25:16,400 --> 00:25:19,800
Speaker 14: But the main thing in behind that not having robots

479
00:25:19,800 --> 00:25:22,360
Speaker 14: around us today is the brain is missing. And if

480
00:25:22,400 --> 00:25:25,520
Speaker 14: you have the brain for robots, you can really enable

481
00:25:25,560 --> 00:25:28,159
Speaker 14: them around in variety of areas. We active applications, and

482
00:25:28,240 --> 00:25:30,960
Speaker 14: that's what exactly SCARED is doing. We are building what

483
00:25:31,000 --> 00:25:35,080
Speaker 14: we call only bodied brain, any robot, any task one brain.

484
00:25:35,440 --> 00:25:36,840
Speaker 6: I'm interested in how you're building it.

485
00:25:36,880 --> 00:25:38,840
Speaker 2: The limitation that a lot of people are talking about

486
00:25:38,920 --> 00:25:42,199
Speaker 2: right now is real world data and training. There are

487
00:25:42,240 --> 00:25:47,879
Speaker 2: different approaches digital twins, trying to simulate physics in digital worlds.

488
00:25:47,920 --> 00:25:49,879
Speaker 2: This is very much what in video is focused on,

489
00:25:50,000 --> 00:25:53,879
Speaker 2: right is that the approach you've taken simulated or synthetic data.

490
00:25:53,960 --> 00:25:57,000
Speaker 14: So I think I will draw analogy to lllms like Jagpity,

491
00:25:57,080 --> 00:25:59,960
Speaker 14: like models like if you see open a head Jadgepitty,

492
00:26:00,240 --> 00:26:03,200
Speaker 14: And soon after many companies had their own LM models,

493
00:26:03,280 --> 00:26:05,359
Speaker 14: and the reason for that is data exist on the

494
00:26:05,440 --> 00:26:08,720
Speaker 14: Internet for language or vision, but robotics there is no

495
00:26:08,880 --> 00:26:12,320
Speaker 14: Internet of robotics today. So what we do is we

496
00:26:12,359 --> 00:26:16,560
Speaker 14: look for alternate sources. One such source is human videos.

497
00:26:17,000 --> 00:26:20,840
Speaker 14: We look at how humans operate in their scenarios in

498
00:26:20,880 --> 00:26:25,720
Speaker 14: your kchareen in while working in factories, and combine that

499
00:26:25,880 --> 00:26:28,520
Speaker 14: with training come simulation, which companies like Nvidia invest a

500
00:26:28,560 --> 00:26:29,960
Speaker 14: lot on, like omni.

501
00:26:29,760 --> 00:26:30,480
Speaker 4: Verse and all those.

502
00:26:30,720 --> 00:26:34,920
Speaker 14: So these two things watching people and then practicing in simulation.

503
00:26:35,160 --> 00:26:39,160
Speaker 14: That's how we combine and scale these models to large scale.

504
00:26:39,440 --> 00:26:42,200
Speaker 4: Your software can be put within the GPU.

505
00:26:42,840 --> 00:26:45,320
Speaker 3: And what's interesting is you're already going from what we

506
00:26:45,400 --> 00:26:48,680
Speaker 3: understand zero to tens of millions of dollars deepak in

507
00:26:48,760 --> 00:26:52,040
Speaker 3: revenue just a few months in twenty twenty five.

508
00:26:52,400 --> 00:26:55,720
Speaker 4: Where's that revenue coming from? Who is using your software already?

509
00:26:56,680 --> 00:26:59,720
Speaker 14: Yep, So this is the revenue has mostly been from

510
00:26:59,720 --> 00:27:05,479
Speaker 14: the price applications. And the idea is in like instead

511
00:27:05,480 --> 00:27:09,280
Speaker 14: of instead of going for our end goal is a

512
00:27:09,280 --> 00:27:14,320
Speaker 14: consumer homelike applications. But right now, since it's just a

513
00:27:14,359 --> 00:27:16,639
Speaker 14: beginning of the field, we are taking this model and

514
00:27:16,680 --> 00:27:18,919
Speaker 14: we are deploying it in variety of applications, like we

515
00:27:18,960 --> 00:27:24,400
Speaker 14: are deploying robots in sectors like point to point delivery applications,

516
00:27:25,040 --> 00:27:31,119
Speaker 14: security applications, data centers, and manufacturing and warehouse so the

517
00:27:31,160 --> 00:27:33,160
Speaker 14: revenue is coming from most of these applications.

518
00:27:33,480 --> 00:27:37,000
Speaker 3: You're born out of academia and you're still professor kind

519
00:27:37,040 --> 00:27:39,480
Speaker 3: of emlan. But what's interesting is you're not without some

520
00:27:39,480 --> 00:27:42,359
Speaker 3: competition out there. I think about physical intelligence for example,

521
00:27:42,480 --> 00:27:45,240
Speaker 3: you're all trying to intend to get robots to become

522
00:27:45,280 --> 00:27:48,720
Speaker 3: intelligent and situationally aware. Do you think you can take

523
00:27:48,760 --> 00:27:50,240
Speaker 3: on Are you worried about competition?

524
00:27:51,600 --> 00:27:54,800
Speaker 14: Well, competition is always good and turns out like the

525
00:27:54,880 --> 00:27:59,360
Speaker 14: AI community born from very small circles. So they're all

526
00:27:59,400 --> 00:28:04,520
Speaker 14: our friends and colleagues from earlier. But what's unique about

527
00:28:04,560 --> 00:28:07,879
Speaker 14: what we are building is this only bodied brain, like

528
00:28:08,320 --> 00:28:11,640
Speaker 14: any robot, any task one brain. Now, if you really

529
00:28:12,160 --> 00:28:16,000
Speaker 14: think deeply about this, it sounds absurd, like a humanoid

530
00:28:16,080 --> 00:28:18,919
Speaker 14: robot with a human like body, a dog robot and

531
00:28:18,960 --> 00:28:21,880
Speaker 14: they share the same brain. And that's what you see

532
00:28:22,119 --> 00:28:23,680
Speaker 14: in the reserves we release.

533
00:28:24,200 --> 00:28:25,760
Speaker 6: The difference in approach is fascinating.

534
00:28:25,800 --> 00:28:28,040
Speaker 2: It is a crowded failed you know, Burnt from one

535
00:28:28,160 --> 00:28:30,159
Speaker 2: X was on the show yesterday. They have a world

536
00:28:30,240 --> 00:28:35,560
Speaker 2: model that essentially allows NEO to carry out a task.

537
00:28:35,640 --> 00:28:39,360
Speaker 2: It's it's never seen before and that largely you know,

538
00:28:39,400 --> 00:28:41,640
Speaker 2: I'm not an engineer, right, but what you have in

539
00:28:41,640 --> 00:28:46,600
Speaker 2: common is VLN. Those are the inputs distinguish your approach

540
00:28:46,640 --> 00:28:48,520
Speaker 2: to one X's as an example.

541
00:28:48,640 --> 00:28:52,520
Speaker 14: So in our case, the robot learns by watching people.

542
00:28:52,640 --> 00:28:54,680
Speaker 14: Right now, imagine I have to pick up this cup

543
00:28:54,680 --> 00:28:57,520
Speaker 14: in front of mee do I have to exactly imagine?

544
00:28:57,520 --> 00:29:01,360
Speaker 14: So scary By the way, watching me you learned when

545
00:29:01,360 --> 00:29:03,320
Speaker 14: you were a kid. You learned by watching your periods.

546
00:29:03,760 --> 00:29:07,200
Speaker 14: You learn watching others. Now, if you think about how

547
00:29:07,240 --> 00:29:09,640
Speaker 14: you how you operate. Every time you pick up a cup,

548
00:29:09,920 --> 00:29:12,080
Speaker 14: do you imagine in your head a picture of cup

549
00:29:12,120 --> 00:29:14,360
Speaker 14: and then putting your finger in the handle, And no,

550
00:29:14,480 --> 00:29:17,800
Speaker 14: it's all mushed in like somehow it's consciousness inside your

551
00:29:17,800 --> 00:29:20,480
Speaker 14: body and then you imagine in some abstract space and

552
00:29:20,680 --> 00:29:23,080
Speaker 14: then do it. And that is the main idea here.

553
00:29:23,360 --> 00:29:25,560
Speaker 14: That's how we do it. But just ted being said,

554
00:29:26,000 --> 00:29:28,520
Speaker 14: watching alone is not enough because if it was enough,

555
00:29:28,520 --> 00:29:31,320
Speaker 14: I could play like Federer, keep watching his games all day.

556
00:29:31,640 --> 00:29:35,080
Speaker 14: And that's where practice comes into play now. Practicing in

557
00:29:35,120 --> 00:29:38,440
Speaker 14: the real world in the making mistakes is expensive, and

558
00:29:38,480 --> 00:29:42,960
Speaker 14: that's where become mind simulation. So videos alone is not enough.

559
00:29:43,200 --> 00:29:45,719
Speaker 14: Same alone is not enough. But put them together and

560
00:29:45,840 --> 00:29:48,800
Speaker 14: boom you have a magic recipe to really figure out

561
00:29:48,840 --> 00:29:49,800
Speaker 14: how to scale robotics.

562
00:29:50,480 --> 00:29:53,440
Speaker 3: Magic recipe that values are forteen billion dollars as far

563
00:29:53,560 --> 00:29:57,240
Speaker 3: Deepak Pathact co found a CEO skilled AI. Thanks so

564
00:29:57,320 --> 00:30:00,520
Speaker 3: much for joining us today coming out more funn around news.

565
00:30:00,560 --> 00:30:03,280
Speaker 3: Excel partner Sarah Dlston joins us to talk about her

566
00:30:03,360 --> 00:30:07,200
Speaker 3: latest investment in AI security startup debt First.

567
00:30:07,560 --> 00:30:08,480
Speaker 4: There's somebody meg tech.

568
00:30:20,600 --> 00:30:23,800
Speaker 2: Security startup debt First, has closed a forty million dollars

569
00:30:23,920 --> 00:30:27,480
Speaker 2: Series A funding round. The company uses AI to identify

570
00:30:27,840 --> 00:30:29,560
Speaker 2: and harden software vulnerabilities.

571
00:30:29,560 --> 00:30:31,120
Speaker 6: The round led by Excel.

572
00:30:31,440 --> 00:30:35,280
Speaker 2: Joining us to discuss Excel partner Sarah Wilson Debt First.

573
00:30:35,760 --> 00:30:38,320
Speaker 2: This was a big theme actually at AWS reinvent the

574
00:30:38,400 --> 00:30:41,640
Speaker 2: idea that particularly sort of swarm based agents can go

575
00:30:41,680 --> 00:30:44,480
Speaker 2: on the offense, not just defense. So it's a sensible

576
00:30:44,520 --> 00:30:47,280
Speaker 2: place to start with. Why you made the investment in

577
00:30:47,320 --> 00:30:51,120
Speaker 2: general security intelligence is what can it do that existing

578
00:30:51,160 --> 00:30:55,400
Speaker 2: app SEC and other sort of static analysis tools just can't.

579
00:30:56,360 --> 00:30:59,240
Speaker 15: I think the main thing is performance and speed and

580
00:30:59,280 --> 00:31:01,880
Speaker 15: the ratio of signal to noise. So I think what

581
00:31:01,920 --> 00:31:04,440
Speaker 15: we have to realize in this moment is, as we're

582
00:31:04,480 --> 00:31:07,440
Speaker 15: all so excited about how AI is making it possible

583
00:31:07,440 --> 00:31:10,920
Speaker 15: to ship software so much more quickly, that same AI

584
00:31:10,960 --> 00:31:13,520
Speaker 15: tool is being used by the bad guys to ship

585
00:31:13,560 --> 00:31:17,320
Speaker 15: attacks more quickly. And so the historical ratio of signal

586
00:31:17,360 --> 00:31:20,160
Speaker 15: to noise just won't stand in the current era. And

587
00:31:20,240 --> 00:31:23,160
Speaker 15: so the security platform that Debt first has made is

588
00:31:23,320 --> 00:31:27,520
Speaker 15: AI native. And what they are doing is continuously agentically

589
00:31:27,560 --> 00:31:31,480
Speaker 15: scanning a company's code. They're looking at its architecture, they're

590
00:31:31,480 --> 00:31:36,200
Speaker 15: looking at its business logic, and they're finding verified vulnerabilities

591
00:31:36,440 --> 00:31:39,400
Speaker 15: eight times more than the previous best in class. And

592
00:31:39,440 --> 00:31:41,840
Speaker 15: I think the key there is verified vulnerability is they're

593
00:31:41,880 --> 00:31:44,520
Speaker 15: not raving their hands about things that aren't really a problem.

594
00:31:44,760 --> 00:31:47,479
Speaker 15: They're finding real problems, and then they're also serving up

595
00:31:47,480 --> 00:31:50,360
Speaker 15: a solution. You can press one button and say fix it,

596
00:31:51,160 --> 00:31:52,240
Speaker 15: and folks are doing that.

597
00:31:52,760 --> 00:31:56,400
Speaker 2: In mass I was looking at the round. It's a

598
00:31:56,440 --> 00:31:59,000
Speaker 2: sizable series A. You know, in the last couple of

599
00:31:59,120 --> 00:32:03,680
Speaker 2: years the market has been somewhat distorted depending on the

600
00:32:03,720 --> 00:32:05,480
Speaker 2: company's type.

601
00:32:05,920 --> 00:32:08,320
Speaker 6: But what you just said eight times invest in class.

602
00:32:08,360 --> 00:32:10,880
Speaker 2: How critical is a data point like that to you

603
00:32:10,960 --> 00:32:13,040
Speaker 2: when you decide do I want to be in this

604
00:32:13,160 --> 00:32:14,760
Speaker 2: round a tool or not evidence?

605
00:32:15,280 --> 00:32:18,400
Speaker 15: It's really critical, and I think, particularly in the case

606
00:32:18,400 --> 00:32:22,000
Speaker 15: of security, there's just a societal imperative that we get

607
00:32:22,040 --> 00:32:24,600
Speaker 15: this right. And I think Excel We've been investing for

608
00:32:24,600 --> 00:32:28,320
Speaker 15: forty three years. Throughout that time, security has been a

609
00:32:28,440 --> 00:32:32,720
Speaker 15: key category of investment for US folks like CrowdStrike or netscope,

610
00:32:32,760 --> 00:32:36,400
Speaker 15: that IPO just last year, and what we've seen is

611
00:32:36,440 --> 00:32:41,400
Speaker 15: in these major innovation transitions, these major you know, cycles

612
00:32:41,720 --> 00:32:45,880
Speaker 15: of transition, the security apparatus has to be replaced.

613
00:32:45,960 --> 00:32:47,600
Speaker 4: It has to evolve to that new era.

614
00:32:48,040 --> 00:32:50,760
Speaker 15: And that's why we're so excited about companies like Depth first.

615
00:32:50,960 --> 00:32:54,160
Speaker 15: But you're exactly right, you know, it has to work.

616
00:32:54,920 --> 00:32:56,960
Speaker 15: It's imperative that it works, and so those stats are

617
00:32:57,000 --> 00:32:59,560
Speaker 15: really important. They're important to the CISOs, they're important to

618
00:32:59,560 --> 00:33:02,560
Speaker 15: the fortune fifty that are evaluating these companies. And so

619
00:33:03,040 --> 00:33:06,000
Speaker 15: if we're gonna, you know, put ourselves in our funds

620
00:33:06,040 --> 00:33:09,560
Speaker 15: behind a company, it's imperative that we believe that the

621
00:33:09,600 --> 00:33:12,600
Speaker 15: thing works and not just not just today, but that

622
00:33:12,680 --> 00:33:16,120
Speaker 15: you have the team that can drive over this next

623
00:33:16,200 --> 00:33:19,560
Speaker 15: era of transition, that you have the talent that can

624
00:33:19,600 --> 00:33:21,880
Speaker 15: sort of weather this moment.

625
00:33:22,520 --> 00:33:27,560
Speaker 3: The talent is Hanley Soltofta, and you're seeing Palo Alto Networks.

626
00:33:27,760 --> 00:33:30,479
Speaker 3: Even so it's now he's in big acquisitions in the space,

627
00:33:31,040 --> 00:33:33,360
Speaker 3: trying to get ahead of the cyber risks around AI.

628
00:33:33,920 --> 00:33:36,920
Speaker 4: Do you think that this is a long term bet?

629
00:33:36,920 --> 00:33:39,200
Speaker 4: Will this company go public or do you think it

630
00:33:39,240 --> 00:33:40,400
Speaker 4: will be an M and A story.

631
00:33:41,960 --> 00:33:44,440
Speaker 15: You know, I think this is absolutely a long term bet.

632
00:33:44,520 --> 00:33:48,200
Speaker 15: That is always our you know, our focus when we

633
00:33:48,400 --> 00:33:51,520
Speaker 15: do our investing, and I think that this is a

634
00:33:51,560 --> 00:33:54,480
Speaker 15: team that's really uniquely qualified to go that distance. So

635
00:33:55,120 --> 00:33:58,520
Speaker 15: this is technical leaders from companies like Data Bricks, from

636
00:33:58,600 --> 00:34:02,719
Speaker 15: Google DeepMind, from fair These are folks who have built

637
00:34:03,160 --> 00:34:08,480
Speaker 15: you know, large durable companies and confronted these security problems.

638
00:34:08,000 --> 00:34:09,400
Speaker 4: At an incredible scale.

639
00:34:09,800 --> 00:34:13,919
Speaker 15: But mirroring that cyber experience with the AI experience, very

640
00:34:13,920 --> 00:34:18,240
Speaker 15: few people have had sort of the experience of building.

641
00:34:17,920 --> 00:34:19,759
Speaker 4: At a place like Google.

642
00:34:19,480 --> 00:34:23,040
Speaker 15: Deep Mind and can now apply that to this context.

643
00:34:23,040 --> 00:34:26,960
Speaker 15: So it's a really interdisciplinary team, a highly technical team

644
00:34:27,520 --> 00:34:31,200
Speaker 15: that we believe is very unique in their ability to

645
00:34:31,200 --> 00:34:32,120
Speaker 15: tackle this problem.

646
00:34:32,360 --> 00:34:35,239
Speaker 3: Sarah, you a head of strategy Partnerships at fair So

647
00:34:36,400 --> 00:34:38,600
Speaker 3: when you're thinking about who are the talent you want

648
00:34:38,640 --> 00:34:40,680
Speaker 3: to go for, is it your own network you need

649
00:34:40,760 --> 00:34:44,000
Speaker 3: upon Is it that you're hearing these companies being born

650
00:34:44,080 --> 00:34:46,440
Speaker 3: out of these big deep minds even before they're a

651
00:34:46,440 --> 00:34:50,040
Speaker 3: really created. How you tapping into a really early stage

652
00:34:50,040 --> 00:34:51,520
Speaker 3: business that needs financing?

653
00:34:53,040 --> 00:34:55,280
Speaker 4: Yeah, I think it's all of the above.

654
00:34:55,600 --> 00:34:57,560
Speaker 15: And then you know, we hope to spend a lot

655
00:34:57,600 --> 00:34:59,919
Speaker 15: of time with folks during the early days when they're

656
00:35:00,000 --> 00:35:03,520
Speaker 15: really you know, considering starting a company or the early

657
00:35:03,560 --> 00:35:06,200
Speaker 15: definition of the company. You know, our goal is to

658
00:35:06,280 --> 00:35:11,400
Speaker 15: be along the journey for decades, right the generational companies

659
00:35:11,440 --> 00:35:14,280
Speaker 15: that are going to solve these security problems will last

660
00:35:14,320 --> 00:35:16,359
Speaker 15: that long, and so we do we try and find

661
00:35:16,360 --> 00:35:18,240
Speaker 15: them early. And it can be through a lot of things,

662
00:35:18,280 --> 00:35:20,680
Speaker 15: but certainly one of the greatest signals is when we

663
00:35:20,719 --> 00:35:24,640
Speaker 15: see incredibly talented people that we know to be excellent

664
00:35:25,120 --> 00:35:27,080
Speaker 15: aggregating around an opportunity.

665
00:35:27,640 --> 00:35:29,600
Speaker 4: And I think what really emerged.

666
00:35:29,200 --> 00:35:33,200
Speaker 15: About this team is that they're incredibly motivated by this mission.

667
00:35:33,440 --> 00:35:36,920
Speaker 15: And if you look around all of our critical systems

668
00:35:36,920 --> 00:35:40,520
Speaker 15: in society are powered by software. Software that people can

669
00:35:40,520 --> 00:35:45,120
Speaker 15: write increasingly quickly with AI, and the other dimension that's

670
00:35:45,200 --> 00:35:49,000
Speaker 15: changing is increasingly we're giving this software more autonomy, we're

671
00:35:49,040 --> 00:35:52,040
Speaker 15: giving it more access, we're giving it the ability to reason,

672
00:35:52,440 --> 00:35:57,200
Speaker 15: and that's really exciting, but it's also something that would

673
00:35:57,239 --> 00:35:59,719
Speaker 15: be concerning if those things aren't secure. We just watch

674
00:35:59,800 --> 00:36:02,759
Speaker 15: your segment talking about the robots, right. We want to

675
00:36:02,800 --> 00:36:06,160
Speaker 15: make sure if we're getting these devices access to our homes,

676
00:36:06,200 --> 00:36:09,120
Speaker 15: to our businesses, that they're secure. And I think that

677
00:36:09,120 --> 00:36:11,520
Speaker 15: that is a is a belief that's shared by the

678
00:36:11,520 --> 00:36:15,040
Speaker 15: Depth First team that they this is sort of an

679
00:36:15,080 --> 00:36:17,440
Speaker 15: imperative for society that we get this right, that we

680
00:36:17,600 --> 00:36:21,000
Speaker 15: ensure that all of this new AI generated code that's

681
00:36:21,040 --> 00:36:23,720
Speaker 15: being shipped at increasing rates is all secure.

682
00:36:23,880 --> 00:36:26,280
Speaker 3: What's being shipped by a load of your portfolio companies Cursor,

683
00:36:26,320 --> 00:36:29,080
Speaker 3: Lovable and Thrope among some of the bets over Accel.

684
00:36:29,120 --> 00:36:32,520
Speaker 3: We really appreciate your time, Accel partner Sarah Itttleston. Now

685
00:36:32,600 --> 00:36:36,160
Speaker 3: coming up Spotify, f co, CEOs my Face are host

686
00:36:36,200 --> 00:36:39,040
Speaker 3: of challenges more the Audio Giants plans next.

687
00:36:38,880 --> 00:36:39,480
Speaker 4: That's a blue bag.

688
00:36:39,560 --> 00:36:51,840
Speaker 2: Tech Salesforce is looking to bolster AI capabilities for Slack users.

689
00:36:51,840 --> 00:36:54,640
Speaker 2: The company is releasing an upgraded slack bot that uses

690
00:36:54,680 --> 00:36:58,160
Speaker 2: anthropics claud model, joining us to discuss Rob Seaman's Slack

691
00:36:58,200 --> 00:37:01,480
Speaker 2: interim CEO and Chief Products Office. You know, I have

692
00:37:01,640 --> 00:37:05,040
Speaker 2: not yet got access to this, this updated slack bot,

693
00:37:05,280 --> 00:37:08,560
Speaker 2: but as a Slack user, very interested to see where

694
00:37:08,600 --> 00:37:09,080
Speaker 2: this goes.

695
00:37:09,480 --> 00:37:09,719
Speaker 6: Rob.

696
00:37:09,960 --> 00:37:13,919
Speaker 2: The general idea, right was that Mark Bennioff, the parent

697
00:37:14,000 --> 00:37:17,360
Speaker 2: company CEO, was using chat GPT and you guys are like,

698
00:37:17,440 --> 00:37:19,320
Speaker 2: we want to have an internal tool.

699
00:37:19,560 --> 00:37:22,560
Speaker 6: So has it worked? What is he using it for?

700
00:37:22,880 --> 00:37:26,000
Speaker 6: What are you using it for? Absolutely? So one.

701
00:37:26,000 --> 00:37:28,279
Speaker 16: I can't wait for you to try it. So we're

702
00:37:28,360 --> 00:37:30,799
Speaker 16: using it internally. Marcus certainly using it. I'm using it.

703
00:37:30,840 --> 00:37:32,960
Speaker 16: I'll give you a few examples of how I use it.

704
00:37:33,640 --> 00:37:36,600
Speaker 16: I used it recently. I just said, hey, go find

705
00:37:36,680 --> 00:37:39,319
Speaker 16: the most recent all hands deck that we're working on.

706
00:37:39,320 --> 00:37:40,960
Speaker 16: One of the things that we do is actually introduce

707
00:37:41,040 --> 00:37:43,520
Speaker 16: every single new hire every month, and I said, tell

708
00:37:43,520 --> 00:37:46,440
Speaker 16: me how to pronounce the names of all the new hires,

709
00:37:46,480 --> 00:37:49,360
Speaker 16: and that goes off and finds the presentation without me

710
00:37:49,440 --> 00:37:53,279
Speaker 16: saying reads the presentation figures out where the new hires are,

711
00:37:53,360 --> 00:37:56,000
Speaker 16: loops through all of them, and figures out their highest

712
00:37:56,000 --> 00:37:58,239
Speaker 16: probability nationality and tells me how to pronounce them. I

713
00:37:58,239 --> 00:37:59,600
Speaker 16: don't have to go bug people to figure that out.

714
00:37:59,640 --> 00:38:00,879
Speaker 16: It's awesome and it saves me time.

715
00:38:01,800 --> 00:38:02,760
Speaker 6: In the first instance.

716
00:38:02,800 --> 00:38:05,239
Speaker 2: It is underpinned by Anthropics Claude, Right, and you're going

717
00:38:05,280 --> 00:38:07,319
Speaker 2: to tell me, well, this is interim and we're going

718
00:38:07,360 --> 00:38:09,360
Speaker 2: to look at all kinds of different model options.

719
00:38:09,360 --> 00:38:10,320
Speaker 6: We're model agnostic.

720
00:38:10,680 --> 00:38:13,480
Speaker 2: But it is an interesting and the latest example of

721
00:38:13,520 --> 00:38:17,160
Speaker 2: how Claude helps people in the enterprise get to market quick.

722
00:38:17,400 --> 00:38:19,360
Speaker 6: You know what was your experience of that, rob.

723
00:38:20,160 --> 00:38:22,080
Speaker 16: The biggest thing for us was being able to take

724
00:38:22,120 --> 00:38:24,000
Speaker 16: an m pick it up and set it down in

725
00:38:24,040 --> 00:38:27,319
Speaker 16: our infrastructure in a way that our customer's data never

726
00:38:27,440 --> 00:38:30,080
Speaker 16: left our security boundary and wasn't used for the training

727
00:38:30,080 --> 00:38:32,279
Speaker 16: of the model. And that's what we got with Anthropic

728
00:38:32,400 --> 00:38:34,759
Speaker 16: and Claude. And on top of that, we've done a

729
00:38:34,760 --> 00:38:38,640
Speaker 16: bunch of sophisticated prompt and context engineering that makes slack

730
00:38:38,680 --> 00:38:42,560
Speaker 16: bot deeply personal and a deeply personal agent to every

731
00:38:42,600 --> 00:38:43,759
Speaker 16: single person that uses it.

732
00:38:44,360 --> 00:38:45,960
Speaker 4: Well, but how are you thinking about security?

733
00:38:47,480 --> 00:38:49,560
Speaker 16: Oh, I mean, it's a paramount for us. It's been

734
00:38:49,560 --> 00:38:52,239
Speaker 16: something that's been the very very top of our priorities

735
00:38:52,520 --> 00:38:55,880
Speaker 16: since the foundation of slack and so one slack bot

736
00:38:55,960 --> 00:38:58,800
Speaker 16: is a personal agent. It needs to be deeply trusted

737
00:38:58,880 --> 00:39:01,400
Speaker 16: by the user, but also trusted by the enterprise. And

738
00:39:01,440 --> 00:39:05,759
Speaker 16: so what we do is the user, the slack bot

739
00:39:05,840 --> 00:39:07,839
Speaker 16: is acting on behalf of the user, so it can

740
00:39:07,880 --> 00:39:10,840
Speaker 16: access the information that a user can their direct messages,

741
00:39:11,120 --> 00:39:14,240
Speaker 16: their private channels, and the public channels within a particular

742
00:39:14,280 --> 00:39:15,000
Speaker 16: slack instance.

743
00:39:15,080 --> 00:39:15,560
Speaker 6: And as I.

744
00:39:15,560 --> 00:39:19,200
Speaker 16: Mentioned before, from a company perspective, you can rest assured

745
00:39:19,239 --> 00:39:21,600
Speaker 16: that your data is not being used to train the model,

746
00:39:21,600 --> 00:39:24,040
Speaker 16: nor is it leaving the security boundary. So it's all

747
00:39:24,080 --> 00:39:27,360
Speaker 16: adhering to your existing security policies and slack this.

748
00:39:27,239 --> 00:39:29,840
Speaker 4: Is all about return on AI investment.

749
00:39:30,040 --> 00:39:32,600
Speaker 3: It's all about making it clear to me, to you,

750
00:39:32,760 --> 00:39:35,520
Speaker 3: to the users that the stuff saves you time makes

751
00:39:35,680 --> 00:39:36,240
Speaker 3: more productive.

752
00:39:36,400 --> 00:39:38,040
Speaker 4: How are you measuring that role.

753
00:39:39,360 --> 00:39:42,520
Speaker 16: We don't measure individual productivity, but we do measure the

754
00:39:42,640 --> 00:39:45,040
Speaker 16: usage of slack bots. So we measure the penetration, so

755
00:39:45,080 --> 00:39:47,440
Speaker 16: the number of users that actually use slack bot over

756
00:39:47,480 --> 00:39:49,319
Speaker 16: the number of users that have access to it, but

757
00:39:49,360 --> 00:39:52,400
Speaker 16: at week over week retention and then message intensity, and

758
00:39:52,400 --> 00:39:55,360
Speaker 16: that allows us to approximate the value that's being created,

759
00:39:55,600 --> 00:39:58,160
Speaker 16: and the week over week retention force slackbot is higher

760
00:39:58,200 --> 00:40:00,920
Speaker 16: than any other feature we've ever had, and slack and

761
00:40:00,960 --> 00:40:04,040
Speaker 16: the message intensity keeps growing, and to us, that's a

762
00:40:04,120 --> 00:40:07,400
Speaker 16: quantitative signal of qualitative value that's being derived from our

763
00:40:07,520 --> 00:40:10,680
Speaker 16: users for time savings, but also we're seeing through our

764
00:40:10,719 --> 00:40:14,240
Speaker 16: qualitative surveys people use this for brainstorming and creative pursuits

765
00:40:14,239 --> 00:40:15,440
Speaker 16: that we didn't even expect.

766
00:40:16,480 --> 00:40:19,640
Speaker 3: Rob Seman, Slack intro CEO, and great speaking with you

767
00:40:20,120 --> 00:40:20,919
Speaker 3: about slack bot.

768
00:40:21,040 --> 00:40:21,960
Speaker 4: We'll see how it catches on.

769
00:40:22,040 --> 00:40:24,399
Speaker 3: Meanwhile, let's take a look at shares of Spotify right

770
00:40:24,440 --> 00:40:27,759
Speaker 3: now and indeed over the start of the year well

771
00:40:27,800 --> 00:40:30,560
Speaker 3: off by nine percent. That's, of course, when the new

772
00:40:30,640 --> 00:40:35,640
Speaker 3: co CEOs have joined the company. Gusta Sworderstrom Alex Norstrom

773
00:40:35,880 --> 00:40:39,560
Speaker 3: took over as from the Spotify founder Daniel Eck. Now

774
00:40:39,560 --> 00:40:41,399
Speaker 3: they've both been out the business for a long time,

775
00:40:41,440 --> 00:40:42,840
Speaker 3: but Eck himself.

776
00:40:42,520 --> 00:40:45,120
Speaker 4: Grew the platform from money suck to.

777
00:40:45,040 --> 00:40:47,280
Speaker 3: One of the most powerful businesses in the music industry,

778
00:40:47,280 --> 00:40:50,120
Speaker 3: and the new CEOs have to steer that powerful ship

779
00:40:50,160 --> 00:40:54,880
Speaker 3: around some pretty tough headwinds AI angry creators, user fatigue

780
00:40:55,080 --> 00:40:57,479
Speaker 3: to turn around that share drop and most actually Commen

781
00:40:57,520 --> 00:40:59,760
Speaker 3: has been writing about all of this by the Audio

782
00:40:59,800 --> 00:41:01,160
Speaker 3: GI's leadership transition.

783
00:41:01,200 --> 00:41:05,000
Speaker 4: It's in BusinessWeek and there are a lot of headwinds.

784
00:41:05,000 --> 00:41:06,560
Speaker 4: What is front and center for them?

785
00:41:06,600 --> 00:41:11,000
Speaker 3: How are they're taking on the algos and the AI story.

786
00:41:10,880 --> 00:41:13,640
Speaker 17: They're coming up with a lot of different products that

787
00:41:13,680 --> 00:41:16,480
Speaker 17: they're hoping will kind of buy them some time and

788
00:41:16,520 --> 00:41:18,799
Speaker 17: also maybe address some of what users want to do. So,

789
00:41:18,840 --> 00:41:23,319
Speaker 17: for example, they're integrating Spotify into jack shipt. They've made

790
00:41:23,360 --> 00:41:25,400
Speaker 17: a partnership with Meta on its glasses.

791
00:41:25,840 --> 00:41:27,000
Speaker 4: They are letting you.

792
00:41:27,400 --> 00:41:30,600
Speaker 17: Enter prompts to get a playlist in return, using even

793
00:41:30,680 --> 00:41:34,400
Speaker 17: like online contexts so that you can get the top

794
00:41:34,760 --> 00:41:37,480
Speaker 17: rated albums of the year from Pitchfork or something like that.

795
00:41:37,640 --> 00:41:39,840
Speaker 17: So they really want to give people both this personalization,

796
00:41:40,040 --> 00:41:42,919
Speaker 17: more options showing up where people are. Podcasts are coming

797
00:41:42,960 --> 00:41:46,160
Speaker 17: to Netflix. They're really trying to meet people where they

798
00:41:46,200 --> 00:41:51,120
Speaker 17: are rather than forcing them to just come to Spotify.

799
00:41:51,160 --> 00:41:56,040
Speaker 2: YouTube not a new issue, but the biggest in front

800
00:41:56,080 --> 00:41:58,160
Speaker 2: of them. It seems like, you know, it's very detailed

801
00:41:58,160 --> 00:42:01,720
Speaker 2: in the BusinessWeek story. This competition for eyeballs and YouTube

802
00:42:01,800 --> 00:42:04,399
Speaker 2: is the thing on both of those CEO's deaths right

803
00:42:04,400 --> 00:42:05,279
Speaker 2: now Acshually.

804
00:42:05,680 --> 00:42:09,600
Speaker 17: Yes, YouTube, Spotify has moved into video. One of the

805
00:42:09,760 --> 00:42:12,920
Speaker 17: scoops from our story is that they're now actually focusing

806
00:42:12,960 --> 00:42:16,880
Speaker 17: on fitness content. So as a result of them pushing

807
00:42:16,920 --> 00:42:20,080
Speaker 17: into video podcasts, creators who we see on YouTube that

808
00:42:20,440 --> 00:42:23,680
Speaker 17: actually put out guided workouts have started uploading their content

809
00:42:23,719 --> 00:42:25,960
Speaker 17: to Spotify, and Spotify is seeing that.

810
00:42:25,920 --> 00:42:27,400
Speaker 4: People really enjoy that content.

811
00:42:27,480 --> 00:42:30,560
Speaker 17: So instead of just using playlists to soundtrack their runs,

812
00:42:30,719 --> 00:42:33,719
Speaker 17: they're actually watching yoga on Spotify, for example. So this

813
00:42:33,800 --> 00:42:35,920
Speaker 17: is an area they're pretty excited about and seemed to

814
00:42:35,920 --> 00:42:36,600
Speaker 17: be investing in.

815
00:42:37,520 --> 00:42:39,640
Speaker 4: We know a lot about Daniellek.

816
00:42:39,680 --> 00:42:42,120
Speaker 3: We've sort of grown up alongside him as he's taken

817
00:42:42,160 --> 00:42:44,960
Speaker 3: this European success story global.

818
00:42:45,520 --> 00:42:46,840
Speaker 4: What do we know about the co CEOs?

819
00:42:46,880 --> 00:42:49,720
Speaker 17: Who are they? Well, their descents, what are they about?

820
00:42:50,320 --> 00:42:53,839
Speaker 17: So both of them are Swedish. Alex is of half

821
00:42:53,920 --> 00:42:56,279
Speaker 17: Chinese half Swedish descent. His mother was a single mother,

822
00:42:56,400 --> 00:42:58,719
Speaker 17: Sharan a Chinese restaurant. Growing up, he's really kind of

823
00:42:58,719 --> 00:43:02,399
Speaker 17: the culture guy. He gets a playlist regularly that keeps

824
00:43:02,480 --> 00:43:04,520
Speaker 17: him on top of trends. It seems like he's really

825
00:43:04,520 --> 00:43:06,759
Speaker 17: sort of a social person who's doing all these partnership

826
00:43:06,760 --> 00:43:10,440
Speaker 17: deals with the labels and audiobooks and such. Gustave formerly

827
00:43:10,480 --> 00:43:13,400
Speaker 17: really focused on product now co CEO also Swedish. He

828
00:43:13,480 --> 00:43:15,799
Speaker 17: sold a startup to Yahoo back in two thousand and six.

829
00:43:16,160 --> 00:43:18,319
Speaker 17: He did a stint in the Swedish military, and he's

830
00:43:18,400 --> 00:43:21,280
Speaker 17: really tech philosophy focused.

831
00:43:21,520 --> 00:43:23,280
Speaker 4: He was really excited about a book.

832
00:43:23,040 --> 00:43:25,960
Speaker 17: He read from the inventor of quantum computing, so he

833
00:43:26,000 --> 00:43:26,640
Speaker 17: really gets.

834
00:43:26,480 --> 00:43:27,040
Speaker 4: Deep on that.

835
00:43:27,080 --> 00:43:29,360
Speaker 17: And supposedly the two of them together are really like

836
00:43:29,400 --> 00:43:30,080
Speaker 17: a yin and yang.

837
00:43:30,120 --> 00:43:30,920
Speaker 4: They go well together.

838
00:43:31,440 --> 00:43:35,680
Speaker 17: A fun tidbit, Nordstrom means north stream, Soderstrom means south stream,

839
00:43:36,600 --> 00:43:38,880
Speaker 17: so it's written at the universe.

840
00:43:38,920 --> 00:43:40,000
Speaker 4: I guess.

841
00:43:41,120 --> 00:43:44,120
Speaker 2: This really is a must read on Bloomberg online or

842
00:43:44,120 --> 00:43:46,719
Speaker 2: in a magazine coming to you very soon in BusinessWeek.

843
00:43:46,719 --> 00:43:49,160
Speaker 2: Bloomberg's actually come and with the deep dive in Spotify.

844
00:43:49,200 --> 00:43:51,560
Speaker 2: Thank you very much, Caro. That does it for another

845
00:43:51,719 --> 00:43:55,640
Speaker 2: edition of Bloomberg. Tech markets still continuing to start the

846
00:43:55,719 --> 00:43:56,640
Speaker 2: year pretty.

847
00:43:56,320 --> 00:43:58,000
Speaker 4: Hot, pretty hot.

848
00:43:58,040 --> 00:43:59,799
Speaker 3: Well, we're under pressure if you're not a big tech

849
00:44:00,000 --> 00:44:02,040
Speaker 3: they're hot if you're looking at commodities. And that too

850
00:44:02,200 --> 00:44:04,040
Speaker 3: is an AI story. I mean, the joke of twenty

851
00:44:04,080 --> 00:44:06,879
Speaker 3: twenty two and onwards. Everything's an AI story, don't forget

852
00:44:06,880 --> 00:44:07,879
Speaker 3: to check out or all of them.

853
00:44:07,760 --> 00:44:08,480
Speaker 4: On our podcast.

854
00:44:08,760 --> 00:44:10,560
Speaker 3: Find on The Terminal, as well as online on Apple,

855
00:44:10,920 --> 00:44:13,480
Speaker 3: on Spotify, and on iHeart as a Bloomberg