1
00:00:02,520 --> 00:00:06,880
Speaker 1: Bloomberg Audio Studios, podcasts, radio news.

2
00:00:07,200 --> 00:00:11,080
Speaker 2: This is Bloomberg Business Week with Carol Messer and Tim

3
00:00:11,119 --> 00:00:13,399
Speaker 2: Stenebek on Bloomberg Radio.

4
00:00:14,400 --> 00:00:19,480
Speaker 3: These are not upie. We've got to have some music.

5
00:00:20,040 --> 00:00:23,960
Speaker 4: Front, all right. We are talking new frontiers. We're talking

6
00:00:24,000 --> 00:00:25,759
Speaker 4: actually we talk about it. I feel like every day

7
00:00:26,079 --> 00:00:30,200
Speaker 4: anything and everything to do with artificial intelligence, specifically jen AI,

8
00:00:31,280 --> 00:00:34,760
Speaker 4: LM's large language models, the components that play into it,

9
00:00:34,800 --> 00:00:37,280
Speaker 4: the power that's going to be needed to power the

10
00:00:37,360 --> 00:00:40,919
Speaker 4: data centers, and who is doing all of the build out.

11
00:00:41,440 --> 00:00:44,800
Speaker 4: Having said that, we've been watching Dell Technology stock rallied

12
00:00:44,840 --> 00:00:46,519
Speaker 4: in today's session, up about eight percent.

13
00:00:47,200 --> 00:00:47,760
Speaker 1: Some news.

14
00:00:47,840 --> 00:00:50,240
Speaker 4: Loop Capital, which has a buy on the stock, raised

15
00:00:50,240 --> 00:00:52,560
Speaker 4: the price target from one twenty five to one five,

16
00:00:53,000 --> 00:00:56,280
Speaker 4: boosted some investor confidence. Keep in mind, Dell does report

17
00:00:56,320 --> 00:01:00,880
Speaker 4: earnings on May thirtieth, so that's tomorrow after the And

18
00:01:00,920 --> 00:01:04,080
Speaker 4: then yesterday some news of Dell expanding its AI factory

19
00:01:04,080 --> 00:01:07,480
Speaker 4: with Nvidia to include new server, edge workstation solutions and

20
00:01:07,800 --> 00:01:10,920
Speaker 4: services advancements that speed AI adoption and innovation. There's a

21
00:01:10,959 --> 00:01:13,560
Speaker 4: lot going on. Michael Dell just about a week and

22
00:01:13,600 --> 00:01:16,720
Speaker 4: a half ago talking about aipcs being pretty standard in

23
00:01:16,800 --> 00:01:19,880
Speaker 4: twenty twenty five. So we have a great guest to

24
00:01:19,920 --> 00:01:21,959
Speaker 4: talk about a lot of what is going on here.

25
00:01:22,400 --> 00:01:25,280
Speaker 4: He participated in a panel with me put on by

26
00:01:25,319 --> 00:01:29,080
Speaker 4: our Bloomberg Intelligence team looking at generative AI, specifically on

27
00:01:29,120 --> 00:01:33,520
Speaker 4: the potential build out by companies of AI data capabilities

28
00:01:33,520 --> 00:01:36,160
Speaker 4: on site or on premise with us as John Rose.

29
00:01:36,200 --> 00:01:39,280
Speaker 4: He's global Chief Technology Officer at Dell Technology. He's joining

30
00:01:39,319 --> 00:01:41,520
Speaker 4: us from New Hampshire. John, So, nice to have you here.

31
00:01:41,560 --> 00:01:42,039
Speaker 4: How are you.

32
00:01:43,120 --> 00:01:45,000
Speaker 3: I'm doing great? How are you all doing well?

33
00:01:45,040 --> 00:01:48,760
Speaker 4: Doing well? There is a lot coming at us right now.

34
00:01:49,240 --> 00:01:51,720
Speaker 4: Talk to us about what you are seeing. We talked

35
00:01:51,720 --> 00:01:56,200
Speaker 4: about at that Bloomberg Intelligence event about what companies were doing,

36
00:01:56,240 --> 00:01:58,520
Speaker 4: what their demands would be, you know, to bring things

37
00:01:58,640 --> 00:02:02,280
Speaker 4: on site on PREMI what's been some of the conversations

38
00:02:02,280 --> 00:02:04,480
Speaker 4: that you are having with clients around this as of late.

39
00:02:05,720 --> 00:02:08,840
Speaker 5: Yeah, I mean, let me rewind a little bit. Last year,

40
00:02:08,880 --> 00:02:11,960
Speaker 5: the first year of the AHI era. You know, I

41
00:02:11,960 --> 00:02:14,720
Speaker 5: think most enterprises were trying to figure out what should

42
00:02:14,760 --> 00:02:18,040
Speaker 5: they do, and largely most of the large enterprises did

43
00:02:18,040 --> 00:02:19,040
Speaker 5: a lot of experimentation.

44
00:02:19,120 --> 00:02:22,160
Speaker 3: This year is different. This year, many of those.

45
00:02:21,919 --> 00:02:25,080
Speaker 5: Initial experiments which we're really trying to figure out where

46
00:02:25,080 --> 00:02:27,960
Speaker 5: to apply this technology for the best return. You know,

47
00:02:28,120 --> 00:02:30,840
Speaker 5: if you're an enterprise, you could apply aid anything, but

48
00:02:30,919 --> 00:02:32,920
Speaker 5: if you apply it to your supply chain or your

49
00:02:32,919 --> 00:02:35,240
Speaker 5: product development cycle or something that really moves the needle

50
00:02:35,320 --> 00:02:38,000
Speaker 5: from an economic perspective, that'll have a bigger impact. So

51
00:02:38,040 --> 00:02:41,440
Speaker 5: I think today most enterprises are triangulating on where to

52
00:02:41,480 --> 00:02:43,920
Speaker 5: apply it, which then gets them to the discussion of

53
00:02:43,960 --> 00:02:46,960
Speaker 5: how to apply it. And that's where you know, the

54
00:02:47,000 --> 00:02:49,040
Speaker 5: panel we add when we were chatting kind of went,

55
00:02:49,200 --> 00:02:53,000
Speaker 5: which is this is not a workload that's you know,

56
00:02:53,080 --> 00:02:55,120
Speaker 5: sitting on the side used by three people. This is

57
00:02:55,160 --> 00:02:57,280
Speaker 5: the center of your enterprise. It's going to run non

58
00:02:57,320 --> 00:02:59,240
Speaker 5: stop seven by twenty four. It's going to power your

59
00:02:59,280 --> 00:03:03,640
Speaker 5: sales processing tompower your customer satisfaction. And so choosing the

60
00:03:03,720 --> 00:03:06,400
Speaker 5: right place to run it, whether that you know that

61
00:03:06,520 --> 00:03:09,000
Speaker 5: gives you the best economic outcome, the best control, doesn't

62
00:03:09,040 --> 00:03:11,800
Speaker 5: get you into compliance and regulatory challenges, is really the

63
00:03:11,880 --> 00:03:14,120
Speaker 5: dialogue that's happening now. And so a lot of the

64
00:03:14,120 --> 00:03:16,240
Speaker 5: work we're doing that you mentioned on the intro around

65
00:03:16,280 --> 00:03:19,360
Speaker 5: with our ecosyst around Nvidia and Meta and everybody else,

66
00:03:19,480 --> 00:03:22,920
Speaker 5: is how do we reduce the complexity to make that decision.

67
00:03:22,960 --> 00:03:24,680
Speaker 3: How do we make it easier for people to.

68
00:03:24,720 --> 00:03:27,519
Speaker 5: Get started, to not have to do everything, and to

69
00:03:27,600 --> 00:03:29,840
Speaker 5: really get the platforms in place where they need them.

70
00:03:29,880 --> 00:03:32,000
Speaker 5: And our opinion, one of the best places to do

71
00:03:32,080 --> 00:03:35,200
Speaker 5: some of this stuff is clearly in their own owned infrastructure.

72
00:03:35,240 --> 00:03:37,760
Speaker 5: One it tends to be a Capex model, and two

73
00:03:37,960 --> 00:03:40,320
Speaker 5: it's in your control and it's much more predictable.

74
00:03:40,400 --> 00:03:40,640
Speaker 1: John.

75
00:03:40,680 --> 00:03:43,240
Speaker 2: For years we've heard about the cloud being what's nimble

76
00:03:43,400 --> 00:03:46,040
Speaker 2: and the cloud being the place that we can do

77
00:03:46,120 --> 00:03:50,800
Speaker 2: this stuff and secure, secure, do this stuff for less expensive.

78
00:03:52,240 --> 00:03:56,280
Speaker 2: I'm wondering the shift on premise is not a shift.

79
00:03:56,280 --> 00:03:58,760
Speaker 2: It's been around for but since before the cloud. But

80
00:03:58,760 --> 00:04:00,800
Speaker 2: why are we seeing and top thing so much about

81
00:04:00,800 --> 00:04:01,640
Speaker 2: the shift right now?

82
00:04:02,440 --> 00:04:03,080
Speaker 3: Yeah, this is.

83
00:04:03,000 --> 00:04:05,960
Speaker 5: Remember the cloud here was all about taking your existing workloads,

84
00:04:05,960 --> 00:04:09,440
Speaker 5: your web servers, your email, your office productivity and maybe

85
00:04:09,440 --> 00:04:10,920
Speaker 5: trying to figure out a different way to operate it.

86
00:04:10,920 --> 00:04:13,040
Speaker 5: If by the cloud wasn't just public cloud, the cloud

87
00:04:13,080 --> 00:04:16,680
Speaker 5: model pervade everything and so you know, the shift autonomous,

88
00:04:16,800 --> 00:04:20,479
Speaker 5: automated elastic infrastructure happened with a set of applications that

89
00:04:20,480 --> 00:04:23,200
Speaker 5: we understood well. The things we're building now look nothing

90
00:04:23,320 --> 00:04:26,080
Speaker 5: like that. A large scale generative AI system for an

91
00:04:26,160 --> 00:04:30,160
Speaker 5: enterprise is arguably the most demanding and complex workload you

92
00:04:30,200 --> 00:04:33,440
Speaker 5: will create in your lifetime, and so you know, we're

93
00:04:34,000 --> 00:04:36,920
Speaker 5: you could argue that, you know, an OPX driven as

94
00:04:36,960 --> 00:04:39,280
Speaker 5: a service model that gives you lots of agility but

95
00:04:39,400 --> 00:04:42,800
Speaker 5: you pay by the drip is not a very good outcome.

96
00:04:43,000 --> 00:04:45,640
Speaker 5: If the thing you're running runs seven x twenty four

97
00:04:45,680 --> 00:04:48,919
Speaker 5: at enormous performance levels, you know you don't want that

98
00:04:49,000 --> 00:04:51,839
Speaker 5: meter running. You want that meter to be predictable, and

99
00:04:51,920 --> 00:04:54,480
Speaker 5: so it's just a different class of workload. By the way,

100
00:04:54,520 --> 00:04:56,400
Speaker 5: we're big proponents in multi cloud. A lot of the time,

101
00:04:56,440 --> 00:04:59,440
Speaker 5: the best place to develop your AI system is in

102
00:04:59,440 --> 00:05:01,039
Speaker 5: one of the cloud providers because they have a great

103
00:05:01,080 --> 00:05:04,240
Speaker 5: tool chain. The best place to test it might be there,

104
00:05:04,240 --> 00:05:06,200
Speaker 5: Maybe the best place to train your models might be

105
00:05:06,200 --> 00:05:08,000
Speaker 5: there because you only need the infrastructure for a short

106
00:05:08,040 --> 00:05:10,440
Speaker 5: period of time. But the minute it becomes inference that

107
00:05:10,440 --> 00:05:12,960
Speaker 5: you're putting it into production and it's using your data,

108
00:05:13,000 --> 00:05:14,880
Speaker 5: which by the way, most of that is on prem

109
00:05:14,920 --> 00:05:18,320
Speaker 5: even today, then it starts to become a very different discussion,

110
00:05:18,320 --> 00:05:21,719
Speaker 5: which brings kind of this modern on prem architecture that

111
00:05:21,760 --> 00:05:24,520
Speaker 5: we talk about with the AI factory into play as

112
00:05:24,600 --> 00:05:26,720
Speaker 5: probably one of the more logical places to start.

113
00:05:26,800 --> 00:05:29,280
Speaker 4: Well, what is exactly the concept of an AI factory.

114
00:05:30,279 --> 00:05:34,200
Speaker 5: Yeah, here's the important thing for people to realize. AI

115
00:05:34,320 --> 00:05:36,520
Speaker 5: is a new workload and by the way, it actually

116
00:05:36,560 --> 00:05:39,640
Speaker 5: needs a new class of infrastructure. The type of compute

117
00:05:39,720 --> 00:05:43,400
Speaker 5: is not CPUs, it's GPUs. The type of data is

118
00:05:43,480 --> 00:05:46,920
Speaker 5: not traditional databases, it's vectorized data that lives in large

119
00:05:47,000 --> 00:05:49,920
Speaker 5: language models. The kind of tools you use are different,

120
00:05:49,920 --> 00:05:52,599
Speaker 5: and so what we talked about at Dell Technologies World

121
00:05:52,680 --> 00:05:55,160
Speaker 5: last year last week was was not a new set

122
00:05:55,200 --> 00:05:57,960
Speaker 5: of products exclusively. It was you really probably need to

123
00:05:58,000 --> 00:06:01,320
Speaker 5: have a separate type of infrastructure where your AI than

124
00:06:01,360 --> 00:06:03,560
Speaker 5: you do for the traditional things that you do, those

125
00:06:03,640 --> 00:06:06,000
Speaker 5: workloads that went through the kind of cloud migration we

126
00:06:06,000 --> 00:06:08,400
Speaker 5: talked about, and what the AI factory is is it's

127
00:06:08,400 --> 00:06:11,160
Speaker 5: an articulation of what that infrastructure looks like. That it

128
00:06:11,200 --> 00:06:14,359
Speaker 5: is accelerated compute, that it's a different kind of data architecture,

129
00:06:14,360 --> 00:06:17,640
Speaker 5: it's a better and different type of networking architecture, and

130
00:06:17,680 --> 00:06:21,120
Speaker 5: that it probably lives in a different footprint because it

131
00:06:21,120 --> 00:06:24,640
Speaker 5: itself has different requirements. Than your legacy applications and the

132
00:06:24,680 --> 00:06:27,000
Speaker 5: other applications you run in either a public cloud or

133
00:06:27,000 --> 00:06:27,760
Speaker 5: a private environment.

134
00:06:27,839 --> 00:06:29,000
Speaker 3: And so the II.

135
00:06:28,800 --> 00:06:32,600
Speaker 5: Factory is how do you create a methodology and organize

136
00:06:32,640 --> 00:06:35,600
Speaker 5: all the technology to put that in play, whether it's

137
00:06:35,640 --> 00:06:38,040
Speaker 5: at a rack level or even an entire data center,

138
00:06:38,320 --> 00:06:41,040
Speaker 5: that builds you the optimal infrastructure to run these new

139
00:06:41,040 --> 00:06:43,120
Speaker 5: workloads that you're going to need. So think of it

140
00:06:43,160 --> 00:06:44,840
Speaker 5: as just a paradigm shift. We're going to have to

141
00:06:44,880 --> 00:06:47,800
Speaker 5: build new infrastructure for this new workload. It might be

142
00:06:47,839 --> 00:06:50,320
Speaker 5: redesigning or optimizing what we have, and it might be

143
00:06:50,360 --> 00:06:51,840
Speaker 5: in fact being your data center build out.

144
00:06:52,080 --> 00:06:54,040
Speaker 4: John. It's really interesting and I think we might have

145
00:06:54,080 --> 00:06:56,640
Speaker 4: talked about this at the Bloomberg Intelligence event, but about

146
00:06:56,640 --> 00:06:59,400
Speaker 4: you know, technology companies, they compete, they work with each other,

147
00:06:59,480 --> 00:07:02,880
Speaker 4: and it was an interesting at Dell Tech World the

148
00:07:02,960 --> 00:07:07,480
Speaker 4: keynote stage to see Jensen Wog and videos CEO up there,

149
00:07:08,400 --> 00:07:11,280
Speaker 4: and so there's really you know, you can see the partnership.

150
00:07:11,360 --> 00:07:13,920
Speaker 4: You can see there's clearly a show support from in

151
00:07:14,040 --> 00:07:16,800
Speaker 4: video to you guys at Dell. What is the nature

152
00:07:16,840 --> 00:07:19,560
Speaker 4: of this partnership, especially when they make their own AI

153
00:07:19,840 --> 00:07:23,840
Speaker 4: server racks, which makes you competitors to some degree.

154
00:07:24,920 --> 00:07:28,400
Speaker 5: Yeah, Remember, Dell is a unique company. We are obviously

155
00:07:28,560 --> 00:07:32,120
Speaker 5: very large, and you could argue we're the largest technology

156
00:07:32,120 --> 00:07:34,640
Speaker 5: integrator in the world. Now it means something to different people,

157
00:07:34,640 --> 00:07:37,440
Speaker 5: but basically, like I don't build my own CPUs or GPUs,

158
00:07:37,480 --> 00:07:40,280
Speaker 5: but what I do is I organize that technology in

159
00:07:40,320 --> 00:07:44,520
Speaker 5: the consumable units of it that my customers across the

160
00:07:44,560 --> 00:07:47,320
Speaker 5: world can consume. Now, when you look at what in

161
00:07:47,400 --> 00:07:50,120
Speaker 5: Video is doing, they clearly are the provider of some

162
00:07:50,160 --> 00:07:51,880
Speaker 5: of the better GPUs in the world. There are other

163
00:07:51,960 --> 00:07:54,920
Speaker 5: choices that we also work with. They also have organized

164
00:07:54,960 --> 00:07:56,880
Speaker 5: their stack in a way that makes it very easy

165
00:07:56,880 --> 00:07:58,640
Speaker 5: to consume, and I think that did a great job there.

166
00:07:58,680 --> 00:08:00,800
Speaker 5: And so they have an early lead and it's importantly

167
00:08:00,960 --> 00:08:04,120
Speaker 5: that they're they're they're making it much more consumable and

168
00:08:04,160 --> 00:08:08,160
Speaker 5: they're keeping the innovation cycle up. However, their ability to

169
00:08:08,200 --> 00:08:10,640
Speaker 5: engage with the large enterprise across the world, they don't

170
00:08:10,640 --> 00:08:12,880
Speaker 5: have the We own much larger salesforce, We have a

171
00:08:12,960 --> 00:08:16,040
Speaker 5: global services capability, we have the largest supply chain that's

172
00:08:16,040 --> 00:08:19,240
Speaker 5: secure in the world and technology and so and by

173
00:08:19,280 --> 00:08:22,040
Speaker 5: the way, you also don't just need the GPUs, you

174
00:08:22,080 --> 00:08:24,200
Speaker 5: need the advanced storage services. You need to integrate it

175
00:08:24,200 --> 00:08:26,000
Speaker 5: with your existing infrastructure. You need to talk to your

176
00:08:26,040 --> 00:08:28,000
Speaker 5: existing data, which, by the way, rides pretty much on

177
00:08:28,080 --> 00:08:31,840
Speaker 5: Dell technology storage systems. And so the nature of the

178
00:08:31,840 --> 00:08:34,640
Speaker 5: relationship is, look, hey, you're trying to build an aifactory.

179
00:08:34,640 --> 00:08:37,120
Speaker 5: There are some leading edge parts that absolutely have to

180
00:08:37,160 --> 00:08:39,320
Speaker 5: be produced and on way to. One way to actually

181
00:08:39,320 --> 00:08:41,600
Speaker 5: make sure that they happen correctly is to integrate them

182
00:08:41,640 --> 00:08:45,000
Speaker 5: into a system, an early system like what Nvidia does. However,

183
00:08:45,080 --> 00:08:48,000
Speaker 5: those parts are decomposable and then they can reassemble into

184
00:08:48,120 --> 00:08:51,040
Speaker 5: other form factors that companies like Dell can take to

185
00:08:51,160 --> 00:08:54,240
Speaker 5: a much more scalable market. The announcements around edge and

186
00:08:54,280 --> 00:08:57,280
Speaker 5: other areas. We have the ability to reach more customers

187
00:08:57,320 --> 00:08:59,480
Speaker 5: than anybody in the world. We need partners to help

188
00:08:59,559 --> 00:09:01,760
Speaker 5: us build technology to bring to them.

189
00:09:01,880 --> 00:09:04,199
Speaker 4: Is it a deeper relationship? Just got about forty seconds

190
00:09:04,240 --> 00:09:07,160
Speaker 4: and we'll come back and continue. But is your partnership

191
00:09:07,200 --> 00:09:09,000
Speaker 4: with Nvidia a little bit different? Is it a deeper

192
00:09:09,040 --> 00:09:12,239
Speaker 4: one because you do have partnerships with other server manufacturers

193
00:09:12,240 --> 00:09:14,640
Speaker 4: as well in others? Is it a deeper relationship or

194
00:09:14,640 --> 00:09:15,400
Speaker 4: how would you call it?

195
00:09:15,480 --> 00:09:17,599
Speaker 5: Yeah, we worked with you know We work with a

196
00:09:17,679 --> 00:09:19,240
Speaker 5: lot of companies, and what I will tell you is

197
00:09:19,520 --> 00:09:20,560
Speaker 5: the first one.

198
00:09:20,600 --> 00:09:21,520
Speaker 3: It was the deepest.

199
00:09:21,679 --> 00:09:24,280
Speaker 5: Last year we announced Project Helix, which was the first

200
00:09:24,280 --> 00:09:27,280
Speaker 5: time anybody articulated putting all the parts together into something

201
00:09:27,280 --> 00:09:31,400
Speaker 5: people could consume. So it has a significant first mover advantage. However,

202
00:09:31,559 --> 00:09:32,880
Speaker 5: you know, like you said, we work with a lot

203
00:09:32,960 --> 00:09:34,440
Speaker 5: of companies and we have a lot of partners.

204
00:09:34,559 --> 00:09:36,839
Speaker 4: You want to continue with our guest, John Roses with US.

205
00:09:37,080 --> 00:09:39,559
Speaker 4: Roses with us. He's a global chief technology officer at

206
00:09:39,559 --> 00:09:42,920
Speaker 4: Dell Technology, still with us from New Hampshire. Hey, John,

207
00:09:42,920 --> 00:09:44,960
Speaker 4: you know we were talking Tim and I in the

208
00:09:44,960 --> 00:09:48,400
Speaker 4: break and just you know, curious about as people go

209
00:09:48,559 --> 00:09:50,880
Speaker 4: right t him to build out their data centers, you

210
00:09:50,920 --> 00:09:53,320
Speaker 4: do wonder how busy it kind of gets for Dell.

211
00:09:53,640 --> 00:09:57,280
Speaker 2: Yeah, I'm wondering how big John, the AI server opportunity

212
00:09:57,640 --> 00:09:59,880
Speaker 2: is for Dell in twenty twenty five, and then how

213
00:10:00,040 --> 00:10:02,160
Speaker 2: big it could be over the next few years. What

214
00:10:02,200 --> 00:10:04,599
Speaker 2: are you folks talking about internally and externally?

215
00:10:05,480 --> 00:10:07,760
Speaker 3: Well, so I'm a CTO, so I'm not going to

216
00:10:07,840 --> 00:10:08,480
Speaker 3: talk about.

217
00:10:08,480 --> 00:10:10,400
Speaker 4: Particular year, But are you really busy?

218
00:10:10,520 --> 00:10:10,800
Speaker 2: John?

219
00:10:11,840 --> 00:10:14,800
Speaker 5: I am extraordinarily busy, But let me paint a picture

220
00:10:14,840 --> 00:10:17,320
Speaker 5: that's a little longer term, you know, and that is, look,

221
00:10:17,320 --> 00:10:20,040
Speaker 5: we are, as I mentioned, we're in now year two

222
00:10:20,120 --> 00:10:22,840
Speaker 5: of the AI cycle, the modern AI cycle, and year

223
00:10:22,880 --> 00:10:26,120
Speaker 5: one was all about surprise, get organized. Year two is

224
00:10:26,160 --> 00:10:29,920
Speaker 5: about the first kind of enterprise deployments and kind of

225
00:10:29,960 --> 00:10:33,400
Speaker 5: the prototypes. What that means is that the enterprise buildout

226
00:10:33,440 --> 00:10:36,640
Speaker 5: hasn't actually begun. And if we look at the size

227
00:10:36,640 --> 00:10:38,679
Speaker 5: of the A market today and what's going on, it's

228
00:10:38,679 --> 00:10:41,200
Speaker 5: a pretty interesting market. What's happening is we're building the

229
00:10:41,200 --> 00:10:44,480
Speaker 5: foundational technologies, we're training the large language models that this

230
00:10:44,520 --> 00:10:47,800
Speaker 5: is a very robust ecosystem right now, we are developing

231
00:10:47,800 --> 00:10:50,400
Speaker 5: the tool chains and as you can see from you know,

232
00:10:50,440 --> 00:10:52,480
Speaker 5: just the state of the industry, it's a pretty exciting

233
00:10:52,520 --> 00:10:56,080
Speaker 5: and a pretty significant shift that is in front of

234
00:10:56,480 --> 00:10:59,840
Speaker 5: the enterprise cycle. We're conservatively, you know, most most people

235
00:11:00,240 --> 00:11:03,839
Speaker 5: over the long term, the AI cycle is about rebalancing

236
00:11:04,080 --> 00:11:06,520
Speaker 5: a sizable portion of the work in the world into

237
00:11:06,520 --> 00:11:09,400
Speaker 5: the machine layer, and so as that occurs, you know,

238
00:11:09,720 --> 00:11:11,800
Speaker 5: it represents you know, sometimes we use the phrase we're

239
00:11:11,800 --> 00:11:13,680
Speaker 5: in the training era. Now we're about to enter the

240
00:11:13,720 --> 00:11:16,640
Speaker 5: infront zer for enterprise, which is AI gets put into production.

241
00:11:16,760 --> 00:11:18,840
Speaker 3: When it does, you know, you can.

242
00:11:18,720 --> 00:11:21,480
Speaker 5: Calculate imagine a world where you know a third of

243
00:11:21,480 --> 00:11:24,480
Speaker 5: the work is now happening in a machine layer. It's

244
00:11:24,480 --> 00:11:26,920
Speaker 5: being done by a machine an AI system. What does

245
00:11:26,960 --> 00:11:29,720
Speaker 5: that look like? How big is that? Well, you can't

246
00:11:29,720 --> 00:11:32,319
Speaker 5: calculate it accurately, but you know it's a gigantic number.

247
00:11:32,360 --> 00:11:34,680
Speaker 5: And it's as big as the Internet build out, it's

248
00:11:34,720 --> 00:11:37,640
Speaker 5: as big as the Industrial revolution. In many of the

249
00:11:37,679 --> 00:11:40,360
Speaker 5: discussions that we have, the timing on it, it's an

250
00:11:40,360 --> 00:11:43,160
Speaker 5: extended cycle. This'll be a twenty year cycle, but we're

251
00:11:43,160 --> 00:11:45,960
Speaker 5: about to enter that phase. And what it tells us

252
00:11:46,040 --> 00:11:48,839
Speaker 5: is it's a significant amount of replumbing of enterprise. It's

253
00:11:48,840 --> 00:11:52,160
Speaker 5: building out AI factories, it's rethinking your data strategy, it's

254
00:11:52,200 --> 00:11:55,080
Speaker 5: rethinking your footprint in the multi cloud, and all of

255
00:11:55,080 --> 00:11:57,360
Speaker 5: that tends to give it our breath and depth in

256
00:11:57,400 --> 00:12:00,600
Speaker 5: our ecosystem, drag us into an awful lot of customer

257
00:12:00,800 --> 00:12:04,120
Speaker 5: conversations in a pretty active world, even in advance of

258
00:12:04,160 --> 00:12:06,760
Speaker 5: the significant buildouts occurring on the infront side.

259
00:12:06,920 --> 00:12:09,360
Speaker 2: So okay, so if we're only in year two when

260
00:12:09,400 --> 00:12:14,880
Speaker 2: it comes to AI, where are corporates in their AI journey?

261
00:12:14,880 --> 00:12:17,080
Speaker 2: How do they adopt their technology. We know Dell won

262
00:12:17,160 --> 00:12:19,679
Speaker 2: a meaningful chunk of Tesla's AI roll out, for example,

263
00:12:19,720 --> 00:12:21,960
Speaker 2: But where are corporates in their AI journey?

264
00:12:22,960 --> 00:12:25,199
Speaker 5: Yeah, the corporate As I mentioned before, last year, it

265
00:12:25,280 --> 00:12:27,280
Speaker 5: was all about getting your feet on the ground, understanding

266
00:12:27,280 --> 00:12:29,600
Speaker 5: the technology of earning what a large language model, earning

267
00:12:29,640 --> 00:12:32,720
Speaker 5: what RAG was, and this year it's all about finding

268
00:12:32,760 --> 00:12:36,120
Speaker 5: those first projects. The first projects are largely under development

269
00:12:36,160 --> 00:12:38,880
Speaker 5: in most large enterprises, and some of them are emerging

270
00:12:38,920 --> 00:12:41,960
Speaker 5: as chat thoughts and other services that we're starting to see.

271
00:12:42,559 --> 00:12:45,360
Speaker 5: And there are definitely early examples of technology that have

272
00:12:45,400 --> 00:12:49,400
Speaker 5: been deployed as new offerings. But the full pivot where

273
00:12:49,400 --> 00:12:52,640
Speaker 5: a company now declares that I am in the center

274
00:12:52,679 --> 00:12:55,640
Speaker 5: of my business, you know, building my product, selling my product,

275
00:12:55,800 --> 00:13:00,000
Speaker 5: servicing my product, engaging with my customers based on primary

276
00:13:00,080 --> 00:13:03,040
Speaker 5: early in AI architecture, that is still work to be done.

277
00:13:03,400 --> 00:13:04,800
Speaker 3: And so so I think, you know, at.

278
00:13:04,720 --> 00:13:07,880
Speaker 5: This point, we're still right now in most large enterprises

279
00:13:07,920 --> 00:13:11,480
Speaker 5: in the first proof of concepts, the first prototypes. But

280
00:13:11,559 --> 00:13:13,240
Speaker 5: one of the things that's different about AI is it

281
00:13:13,240 --> 00:13:15,000
Speaker 5: does not take three years to build one of those.

282
00:13:15,360 --> 00:13:17,640
Speaker 5: You can go from idea based on the tool chains

283
00:13:17,640 --> 00:13:20,960
Speaker 5: available to having something in production that you can start

284
00:13:21,000 --> 00:13:22,760
Speaker 5: to really do, and we're doing that inside of Dell

285
00:13:22,880 --> 00:13:25,480
Speaker 5: make your developers more productive in a matter of months.

286
00:13:25,559 --> 00:13:27,880
Speaker 5: And so the velocity of this cycle is something that

287
00:13:27,880 --> 00:13:31,520
Speaker 5: we've never seen before, which means the gap between getting

288
00:13:31,520 --> 00:13:33,640
Speaker 5: your feet on the ground and being in production and

289
00:13:33,640 --> 00:13:35,920
Speaker 5: transforming your enterprise is not a ten year cycle.

290
00:13:36,000 --> 00:13:38,120
Speaker 4: Hey, John, I keep hearing you know you've said it,

291
00:13:38,200 --> 00:13:40,280
Speaker 4: and I've had other guests, but when they talk about

292
00:13:40,280 --> 00:13:43,480
Speaker 4: AI in inference, right, am I saying it correctly?

293
00:13:43,920 --> 00:13:44,839
Speaker 1: Yeah? Right?

294
00:13:45,400 --> 00:13:48,160
Speaker 4: How is that different from AI training? And is that

295
00:13:48,320 --> 00:13:50,680
Speaker 4: kind of a new concept or is that just something

296
00:13:50,720 --> 00:13:53,760
Speaker 4: more complicated when it comes to generative AI.

297
00:13:54,160 --> 00:13:56,640
Speaker 5: No, they're part of the same cycle that The idea

298
00:13:56,720 --> 00:13:59,680
Speaker 5: behind AI is like you're trying to have a machine

299
00:14:00,480 --> 00:14:04,160
Speaker 5: do some kind of cognitive work, answer a service, call,

300
00:14:04,320 --> 00:14:07,480
Speaker 5: sell something, build a writ code. In order to do that,

301
00:14:07,559 --> 00:14:10,360
Speaker 5: the first step is that you must have that machine

302
00:14:10,520 --> 00:14:13,960
Speaker 5: have some access to the knowledge necessary to do that,

303
00:14:14,000 --> 00:14:16,680
Speaker 5: which is what training is about. The difference in large

304
00:14:16,720 --> 00:14:19,920
Speaker 5: language models is that we've developed techniques that allow us

305
00:14:19,960 --> 00:14:22,880
Speaker 5: to instead of trying to as human beings decide how

306
00:14:22,880 --> 00:14:24,760
Speaker 5: to code, we've learned that if.

307
00:14:24,640 --> 00:14:27,320
Speaker 3: You just expose these new.

308
00:14:27,200 --> 00:14:31,000
Speaker 5: Techniques, these new technologies, large language models, to a gigantic

309
00:14:31,120 --> 00:14:34,840
Speaker 5: set of coding, just examples of coding, they will classify them,

310
00:14:34,920 --> 00:14:37,760
Speaker 5: organize them, and create a neural network. And interestingly enough,

311
00:14:37,800 --> 00:14:41,320
Speaker 5: they will then be able to replicate that intelligence, that behavior.

312
00:14:42,520 --> 00:14:45,320
Speaker 5: And so the training phase is about taking gobs of data.

313
00:14:45,400 --> 00:14:49,600
Speaker 5: In the current phase, it's the entire Internet and run

314
00:14:49,640 --> 00:14:53,360
Speaker 5: it into these models that create systems that can understand

315
00:14:53,520 --> 00:14:56,560
Speaker 5: or communicate human language, that can code. And all that

316
00:14:56,640 --> 00:15:00,520
Speaker 5: is is them deriving from a gigantic data set the

317
00:15:00,600 --> 00:15:03,120
Speaker 5: knowledge that's contained within it to create a set of skills.

318
00:15:03,280 --> 00:15:06,760
Speaker 5: That's training. Inference is totally different. Inference is once you

319
00:15:06,840 --> 00:15:08,680
Speaker 5: have that model, now you want to do.

320
00:15:08,680 --> 00:15:09,320
Speaker 3: Something with it.

321
00:15:09,440 --> 00:15:12,400
Speaker 5: You have a thing that can code great well inferences

322
00:15:12,400 --> 00:15:14,960
Speaker 5: when you tell it to code a program to actually

323
00:15:15,000 --> 00:15:17,560
Speaker 5: produce source code that does something. So they're just two

324
00:15:17,560 --> 00:15:19,680
Speaker 5: habs at the same coin. One is the learning to

325
00:15:19,680 --> 00:15:21,720
Speaker 5: create the capability. The other is the act of using

326
00:15:21,760 --> 00:15:22,920
Speaker 5: the capability and production.

327
00:15:23,520 --> 00:15:25,160
Speaker 4: Hey listen, something I got to ask you. It's a

328
00:15:25,200 --> 00:15:27,600
Speaker 4: story that's on the Bloomberg and this has to do

329
00:15:27,640 --> 00:15:30,040
Speaker 4: with Nobel Laureate he's also an economics professor. We're talking

330
00:15:30,040 --> 00:15:33,560
Speaker 4: about Paul Romer, and he was talking with our team

331
00:15:33,640 --> 00:15:37,640
Speaker 4: here and he said, runaway confidence and artificial intelligence risks

332
00:15:37,640 --> 00:15:39,800
Speaker 4: repeating the mistakes of the crypto hype bubble of only

333
00:15:39,840 --> 00:15:42,440
Speaker 4: two years ago. He says, right now, there's way too

334
00:15:42,520 --> 00:15:45,920
Speaker 4: much confidence about the future trajectory of AI. When people

335
00:15:45,960 --> 00:15:47,920
Speaker 4: project this phot I think they're at risk of making

336
00:15:47,960 --> 00:15:51,200
Speaker 4: a very serious mistake. Do you think that there's too

337
00:15:51,320 --> 00:15:55,240
Speaker 4: much confidence about AI? Do you think there's too much euphoria?

338
00:15:55,480 --> 00:15:58,240
Speaker 5: Two answers to that in the general population. I think

339
00:15:58,280 --> 00:16:01,520
Speaker 5: it's a very confusing space because we have basically the

340
00:16:01,600 --> 00:16:06,760
Speaker 5: human race is maybe susceptible to science fiction, and we

341
00:16:06,840 --> 00:16:09,440
Speaker 5: think that what we're producing is artificial general intelligence, and

342
00:16:09,480 --> 00:16:11,800
Speaker 5: we're producing these these sentient beings.

343
00:16:11,840 --> 00:16:14,800
Speaker 3: These are not sentient beings. These are technologies that yah.

344
00:16:15,720 --> 00:16:18,400
Speaker 5: Yeah, maybe sometimes we all believe there's a path to AGI.

345
00:16:18,480 --> 00:16:21,720
Speaker 5: It's just not in the near term. That whole dialogue,

346
00:16:21,720 --> 00:16:24,840
Speaker 5: and let's call it the consumer general market, which is

347
00:16:24,880 --> 00:16:28,000
Speaker 5: primarily where most of the big public AI players play,

348
00:16:28,360 --> 00:16:31,080
Speaker 5: is a risk because people think these have personalities. They

349
00:16:31,080 --> 00:16:33,840
Speaker 5: don't understand the technology. When you go to the enterprise world,

350
00:16:33,960 --> 00:16:37,080
Speaker 5: it is far more conservative. There is no enterprise in

351
00:16:37,120 --> 00:16:39,360
Speaker 5: the world that's trying to build the terminator or a

352
00:16:39,440 --> 00:16:41,760
Speaker 5: sentient being. What we are doing is we're applying the

353
00:16:41,800 --> 00:16:44,920
Speaker 5: techniques that we're pioneered in the public AI world in

354
00:16:45,000 --> 00:16:49,000
Speaker 5: large language model space. Two very specific problems within a

355
00:16:49,040 --> 00:16:53,480
Speaker 5: corporation that could benefit from it, writing code, finding your customers,

356
00:16:53,880 --> 00:16:55,640
Speaker 5: engaging to solve problems.

357
00:16:55,960 --> 00:16:57,400
Speaker 3: Those are the things that are mattering.

358
00:16:57,400 --> 00:17:00,080
Speaker 5: So enterprise is kind of boring to be perfectly on

359
00:17:00,320 --> 00:17:02,400
Speaker 5: versus what is in the part of the possible in

360
00:17:02,440 --> 00:17:05,960
Speaker 5: the public world. However, as you know, our industrial complex

361
00:17:06,000 --> 00:17:08,040
Speaker 5: is quite large and the impact is much bigger.

362
00:17:09,080 --> 00:17:11,840
Speaker 4: Great stuff already looking forward the next time we get

363
00:17:11,840 --> 00:17:13,720
Speaker 4: to catch up John, Thank you so much. B. While

364
00:17:13,760 --> 00:17:17,159
Speaker 4: John Rose, he's global chief technology officer Dell Technology, is

365
00:17:17,240 --> 00:17:20,280
Speaker 4: joining us from New Hampshire. You are listening and watching

366
00:17:20,280 --> 00:17:23,679
Speaker 4: Bloomberg Business Week Carol Masser along with Tim Stanovic, and

367
00:17:23,720 --> 00:17:25,320
Speaker 4: this is Bloomberg.

368
00:17:29,720 --> 00:17:33,280
Speaker 2: Some disturbing information that we found our disturbing This next

369
00:17:33,280 --> 00:17:36,000
Speaker 2: interview report by the global media company with a health

370
00:17:36,000 --> 00:17:39,439
Speaker 2: and biotech focus. Marianne Liebert found that quote in nearly

371
00:17:39,600 --> 00:17:42,879
Speaker 2: three quarters of the cases where a disease afflicts primarily

372
00:17:42,960 --> 00:17:47,320
Speaker 2: one gender. The funding pattern favors males in that either

373
00:17:47,359 --> 00:17:50,000
Speaker 2: the disease affects more women and is underfunded with respect

374
00:17:50,040 --> 00:17:53,880
Speaker 2: to burden, or the disease affects more men and it's overfunded.

375
00:17:53,960 --> 00:17:56,800
Speaker 2: So double whammy when it comes to the gender health gap.

376
00:17:57,000 --> 00:17:59,200
Speaker 4: Yeah. Meantime, in a recent report for McKinsey, it found

377
00:17:59,200 --> 00:18:02,400
Speaker 4: that reducing the time women spend in poor health by

378
00:18:02,440 --> 00:18:05,240
Speaker 4: twenty five percent could be worth one trillion dollars, in

379
00:18:05,280 --> 00:18:08,920
Speaker 4: large part because health disparities disproportionately hit women during their

380
00:18:08,960 --> 00:18:12,320
Speaker 4: working years. So let's get to it. There are huge

381
00:18:12,359 --> 00:18:15,640
Speaker 4: disparities between men and women's health that definitely tim need

382
00:18:15,680 --> 00:18:16,360
Speaker 4: to be addressed.

383
00:18:16,520 --> 00:18:18,560
Speaker 2: Back with us to talk about that gender health gap

384
00:18:18,640 --> 00:18:21,840
Speaker 2: is Elizabeth Staddinger, managing board member at the sixty five

385
00:18:21,880 --> 00:18:25,320
Speaker 2: billion dollar medical tech company Semen's Health and Ears, publicly

386
00:18:25,320 --> 00:18:27,960
Speaker 2: traded one. I should note she joins us from Germany. Elizabeth,

387
00:18:27,960 --> 00:18:29,800
Speaker 2: good to have you back with us. Thanks for staying

388
00:18:29,840 --> 00:18:32,440
Speaker 2: up a little bit later once again to join us

389
00:18:32,560 --> 00:18:35,040
Speaker 2: from Germany. Talk a little bit about what you're doing

390
00:18:35,040 --> 00:18:37,320
Speaker 2: over at Semens Health and Ears to address this gender

391
00:18:37,359 --> 00:18:37,879
Speaker 2: health gap.

392
00:18:40,320 --> 00:18:44,280
Speaker 1: As you already mentioned, in the opening, there is significant disparities.

393
00:18:44,359 --> 00:18:47,119
Speaker 1: And if you just to make this tangible, if you

394
00:18:47,280 --> 00:18:51,680
Speaker 1: think of the typical Hollywood heart attack where somebody kind

395
00:18:51,720 --> 00:18:55,119
Speaker 1: of reaches to his breast, has this burning sensation in

396
00:18:55,160 --> 00:18:57,720
Speaker 1: the arm, and everybody will start rushing and saying, oh

397
00:18:57,720 --> 00:19:00,639
Speaker 1: my god, something really serious is going on. Let's rush

398
00:19:00,720 --> 00:19:05,879
Speaker 1: that patient to the ED and take care of him. Unfortunately,

399
00:19:06,080 --> 00:19:08,960
Speaker 1: when you're a woman, most likely the symptoms you will

400
00:19:09,000 --> 00:19:13,440
Speaker 1: be feeling are different. And that matters because the likelihood

401
00:19:13,480 --> 00:19:16,560
Speaker 1: as a woman to be misdiagnosed when you're having a

402
00:19:16,560 --> 00:19:19,560
Speaker 1: heart attack is fifty percent higher than it is for men,

403
00:19:20,320 --> 00:19:22,800
Speaker 1: and the likelihood that you will actually not make it

404
00:19:22,840 --> 00:19:25,160
Speaker 1: when you admit it to the hospital is actually twice

405
00:19:25,200 --> 00:19:27,320
Speaker 1: as high as it is for men. And it gives

406
00:19:27,359 --> 00:19:30,000
Speaker 1: you a sense of the huge gap that we have

407
00:19:30,080 --> 00:19:32,600
Speaker 1: and the huge disparities we have when it comes to

408
00:19:33,000 --> 00:19:35,640
Speaker 1: looking at women's health and men's health, and the big

409
00:19:35,680 --> 00:19:38,880
Speaker 1: biases which are deeply ingrained into how we do things.

410
00:19:39,160 --> 00:19:43,800
Speaker 4: Do men and women medical professionals are they biased equally?

411
00:19:44,119 --> 00:19:48,560
Speaker 4: In other words, do women doctors, women nurses also misdiagnose

412
00:19:49,080 --> 00:19:50,439
Speaker 4: or underdiagnosed women.

413
00:19:51,359 --> 00:19:54,840
Speaker 1: That's actually a very interesting question. And if you think

414
00:19:54,880 --> 00:19:58,200
Speaker 1: of what people learn when they get their training bit

415
00:19:58,280 --> 00:20:02,560
Speaker 1: in nursing or in medical school, the typical default everything

416
00:20:02,600 --> 00:20:06,080
Speaker 1: starts from in research, but also in the training in

417
00:20:06,080 --> 00:20:09,199
Speaker 1: mid school is the man and the women. Sometimes is

418
00:20:09,240 --> 00:20:13,919
Speaker 1: considered that's the abnormal version of the human being, and

419
00:20:13,960 --> 00:20:17,080
Speaker 1: that leads to a situation that both men and women

420
00:20:18,160 --> 00:20:21,639
Speaker 1: may have that bias. At the same time. There is

421
00:20:21,680 --> 00:20:24,880
Speaker 1: some data which shows that on average, women are more

422
00:20:24,960 --> 00:20:28,359
Speaker 1: sensitive to some of these differences than men are, just

423
00:20:28,400 --> 00:20:30,600
Speaker 1: maybe because they've also experienced it themselves.

424
00:20:30,600 --> 00:20:32,760
Speaker 4: It's funny even kind of preparing for this segment working

425
00:20:32,760 --> 00:20:35,760
Speaker 4: with our producer Elizabeth Cedric, Like you know, even research

426
00:20:36,000 --> 00:20:39,400
Speaker 4: that's been done is the baseline is off of what's

427
00:20:39,440 --> 00:20:42,600
Speaker 4: going on with men, and it's just interesting, like trying

428
00:20:42,600 --> 00:20:46,200
Speaker 4: to find data that everything is kind of off of

429
00:20:46,240 --> 00:20:48,879
Speaker 4: that base, if you will. It was a little discouraging,

430
00:20:48,920 --> 00:20:49,720
Speaker 4: to say the least.

431
00:20:50,280 --> 00:20:53,159
Speaker 1: Right, there is a lot of bias which is somehow

432
00:20:53,240 --> 00:20:56,119
Speaker 1: built into the system in R and D funding. You

433
00:20:56,200 --> 00:20:59,280
Speaker 1: mentioned that in your opening statement, but also in clinical studies.

434
00:20:59,320 --> 00:21:04,479
Speaker 1: For many, many many years, women were completely underrepresented in

435
00:21:04,520 --> 00:21:08,360
Speaker 1: the clinical studies, which then leads to medication being approved

436
00:21:08,359 --> 00:21:10,000
Speaker 1: based on male samples only.

437
00:21:11,840 --> 00:21:15,080
Speaker 2: I'm wondering, Elizabeth, about the role of technology here, and

438
00:21:15,119 --> 00:21:17,080
Speaker 2: we've had you on to talk about this in the past,

439
00:21:18,080 --> 00:21:21,359
Speaker 2: but I'm wondering about bias when it comes to AI

440
00:21:21,680 --> 00:21:26,120
Speaker 2: or when it comes to diagnoses, because if we talk

441
00:21:26,160 --> 00:21:31,320
Speaker 2: about bias, it's something that's within every person and it's

442
00:21:31,320 --> 00:21:34,280
Speaker 2: within doctors as well. Can we remove some of that

443
00:21:34,400 --> 00:21:38,399
Speaker 2: bias if we use technology instead to make these medical decisions?

444
00:21:39,480 --> 00:21:42,080
Speaker 1: And that's actually a very interesting topic because if you

445
00:21:42,119 --> 00:21:46,680
Speaker 1: think of bias in people and bias in physicians, there

446
00:21:46,760 --> 00:21:48,879
Speaker 1: is a lot of bias in the data which is

447
00:21:49,040 --> 00:21:50,520
Speaker 1: used to train AIDS.

448
00:21:50,600 --> 00:21:53,000
Speaker 2: I was afraid you were going to say this, and

449
00:21:53,040 --> 00:21:53,399
Speaker 2: this is.

450
00:21:53,440 --> 00:21:55,280
Speaker 1: But the good news is there is a lot of

451
00:21:55,320 --> 00:21:58,240
Speaker 1: ways of dealing with that. I mean, we enceemental healthy mes.

452
00:21:58,240 --> 00:22:00,800
Speaker 1: We have more than twenty years of working on deep

453
00:22:00,880 --> 00:22:05,440
Speaker 1: learning AI based algorithms and dealing with patient data. We've

454
00:22:05,480 --> 00:22:08,880
Speaker 1: built a huge database of more than two billion data points,

455
00:22:09,680 --> 00:22:12,800
Speaker 1: and we can make sure as we're training our algorithms

456
00:22:12,840 --> 00:22:16,239
Speaker 1: that we have a well balanced and diverse sample that

457
00:22:16,359 --> 00:22:21,320
Speaker 1: goes into creating the algorithms and the prediction of the

458
00:22:21,359 --> 00:22:24,399
Speaker 1: results that are based on the software what faith do

459
00:22:24,440 --> 00:22:27,600
Speaker 1: There even ways of kind of adjusting for that, but

460
00:22:27,640 --> 00:22:29,359
Speaker 1: you have to make a very conscious effort.

461
00:22:29,520 --> 00:22:30,960
Speaker 4: That's what I was going to say, what faith do

462
00:22:31,000 --> 00:22:33,200
Speaker 4: you have in that happening? Because I feel like, you know,

463
00:22:34,040 --> 00:22:35,600
Speaker 4: you know this, you know where I'm going to go.

464
00:22:36,040 --> 00:22:40,040
Speaker 4: These are not new issues, concerns, problems in terms of

465
00:22:40,040 --> 00:22:41,639
Speaker 4: certainly when it comes I feel like with women and

466
00:22:41,720 --> 00:22:45,119
Speaker 4: men and health and data. So what hopes do you

467
00:22:45,160 --> 00:22:47,800
Speaker 4: have that that you know that that effort will be

468
00:22:47,920 --> 00:22:51,600
Speaker 4: made to make sure that women are incorporated into the

469
00:22:51,640 --> 00:22:52,280
Speaker 4: data points.

470
00:22:53,960 --> 00:22:57,080
Speaker 1: I think there is a very strong link to how

471
00:22:57,119 --> 00:23:00,919
Speaker 1: do we kind of ensure and also demand that the

472
00:23:01,000 --> 00:23:05,000
Speaker 1: data that goes into the training of the AI does

473
00:23:05,080 --> 00:23:08,160
Speaker 1: have the right balance and the right mix. So regulatory

474
00:23:08,280 --> 00:23:11,920
Speaker 1: authorities can help here, but also companies of course can

475
00:23:12,000 --> 00:23:15,040
Speaker 1: make sure and we, for instance, we put a lot

476
00:23:15,040 --> 00:23:17,520
Speaker 1: of effort and energy into making sure that we use

477
00:23:17,640 --> 00:23:21,800
Speaker 1: high quality data which is vetted where we are confident

478
00:23:21,920 --> 00:23:25,960
Speaker 1: and certain that the input we provide to the software

479
00:23:25,960 --> 00:23:30,239
Speaker 1: and to the algorithms really reflects both the disease we

480
00:23:30,280 --> 00:23:32,639
Speaker 1: want to see as well as reflects a good sample

481
00:23:32,720 --> 00:23:35,920
Speaker 1: across not only JENDA but also other dimensions which matter.

482
00:23:36,640 --> 00:23:38,960
Speaker 2: So where do you go for the data to find

483
00:23:39,040 --> 00:23:43,360
Speaker 2: unbiased data well, ultimately or do you just correct for it?

484
00:23:44,600 --> 00:23:47,119
Speaker 1: You essentially correct for it. So when you pull together

485
00:23:47,200 --> 00:23:49,080
Speaker 1: the sample you want to use, you make sure, Okay,

486
00:23:49,080 --> 00:23:51,040
Speaker 1: how many women do I have, how many men? What

487
00:23:51,200 --> 00:23:54,520
Speaker 1: kind of Also do I have enough representation of Asians

488
00:23:54,560 --> 00:24:00,280
Speaker 1: of India and of African of Caucasian patients in order

489
00:24:00,359 --> 00:24:03,480
Speaker 1: to make sure that we do not kind of over

490
00:24:03,600 --> 00:24:05,959
Speaker 1: index on any of those dimensions.

491
00:24:07,119 --> 00:24:09,720
Speaker 4: Yeah, fascinating if we don't get this right, I mean,

492
00:24:10,880 --> 00:24:13,919
Speaker 4: what's at stake. It sounds like a lot in terms

493
00:24:13,920 --> 00:24:17,760
Speaker 4: of health of women overall. And I wonder how much

494
00:24:17,880 --> 00:24:22,560
Speaker 4: in terms of gaps between developed world and developing world,

495
00:24:22,840 --> 00:24:25,359
Speaker 4: you know, where that really maybe has an impact or

496
00:24:25,359 --> 00:24:26,280
Speaker 4: maybe it's everywhere.

497
00:24:27,680 --> 00:24:32,439
Speaker 1: I would say it is really everywhere. You know, women

498
00:24:32,520 --> 00:24:35,199
Speaker 1: are not simply small men, regardless of where you go

499
00:24:35,280 --> 00:24:38,080
Speaker 1: on this planet. And I think it's also reality that

500
00:24:38,160 --> 00:24:41,639
Speaker 1: we cannot improve global health if we ignore half the

501
00:24:41,680 --> 00:24:45,840
Speaker 1: world's population. And there is not only a risk speaking

502
00:24:45,880 --> 00:24:47,480
Speaker 1: now a lot about the risk, but there's also a

503
00:24:47,560 --> 00:24:52,080
Speaker 1: real opportunity because if we manage to remove some of

504
00:24:52,119 --> 00:24:57,399
Speaker 1: these barriers and can improve the health of women in

505
00:24:57,440 --> 00:25:01,320
Speaker 1: the world globally. It will help everyone rise. You mentioned

506
00:25:01,359 --> 00:25:04,800
Speaker 1: the one trillion of GDP by twenty forty that we

507
00:25:04,880 --> 00:25:08,359
Speaker 1: could be tapping into if we did a better job

508
00:25:08,480 --> 00:25:13,000
Speaker 1: at removing the barriers and the disparities in health.

509
00:25:14,119 --> 00:25:16,960
Speaker 2: Elizabeth, thank you for joining us an important conversation. We

510
00:25:17,000 --> 00:25:18,639
Speaker 2: always love it when you take the time to join us,

511
00:25:18,680 --> 00:25:20,800
Speaker 2: and again thanks for stand up late over there in

512
00:25:20,840 --> 00:25:24,600
Speaker 2: a Germany. Elizabeth Stattinger is managing board member over at

513
00:25:24,720 --> 00:25:28,159
Speaker 2: Siemens Health and Ears. It's a publicly traded medtech company

514
00:25:28,240 --> 00:25:30,119
Speaker 2: sixty five billion dollar market cap