Bloomberg Audio Studios, podcasts, radio news.
This is Bloomberg Business Week with Carol Messer and Tim Stenebek on Bloomberg Radio.
These are not upie. We've got to have some music.
Front, all right. We are talking new frontiers. We're talking actually we talk about it. I feel like every day anything and everything to do with artificial intelligence, specifically jen AI, LM's large language models, the components that play into it, the power that's going to be needed to power the data centers, and who is doing all of the build out. Having said that, we've been watching Dell Technology stock rallied in today's session, up about eight percent.
Some news.
Loop Capital, which has a buy on the stock, raised the price target from one twenty five to one five, boosted some investor confidence. Keep in mind, Dell does report earnings on May thirtieth, so that's tomorrow after the And then yesterday some news of Dell expanding its AI factory with Nvidia to include new server, edge workstation solutions and services advancements that speed AI adoption and innovation. There's a
lot going on. Michael Dell just about a week and a half ago talking about aipcs being pretty standard in twenty twenty five. So we have a great guest to talk about a lot of what is going on here. He participated in a panel with me put on by our Bloomberg Intelligence team looking at generative AI, specifically on the potential build out by companies of AI data capabilities on site or on premise with us as John Rose. He's global Chief Technology Officer at Dell Technology. He's joining
us from New Hampshire. John, So, nice to have you here. How are you.
I'm doing great? How are you all doing well?
Doing well? There is a lot coming at us right now. Talk to us about what you are seeing. We talked about at that Bloomberg Intelligence event about what companies were doing, what their demands would be, you know, to bring things on site on PREMI what's been some of the conversations that you are having with clients around this as of late.
Yeah, I mean, let me rewind a little bit. Last year, the first year of the AHI era. You know, I think most enterprises were trying to figure out what should they do, and largely most of the large enterprises did a lot of experimentation.
This year is different. This year, many of those.
Initial experiments which we're really trying to figure out where to apply this technology for the best return. You know, if you're an enterprise, you could apply aid anything, but if you apply it to your supply chain or your product development cycle or something that really moves the needle from an economic perspective, that'll have a bigger impact. So I think today most enterprises are triangulating on where to apply it, which then gets them to the discussion of
how to apply it. And that's where you know, the panel we add when we were chatting kind of went, which is this is not a workload that's you know, sitting on the side used by three people. This is the center of your enterprise. It's going to run non stop seven by twenty four. It's going to power your
sales processing tompower your customer satisfaction. And so choosing the right place to run it, whether that you know that gives you the best economic outcome, the best control, doesn't get you into compliance and regulatory challenges, is really the dialogue that's happening now. And so a lot of the work we're doing that you mentioned on the intro around with our ecosyst around Nvidia and Meta and everybody else, is how do we reduce the complexity to make that decision.
How do we make it easier for people to.
Get started, to not have to do everything, and to really get the platforms in place where they need them. And our opinion, one of the best places to do some of this stuff is clearly in their own owned infrastructure. One it tends to be a Capex model, and two it's in your control and it's much more predictable.
John.
For years we've heard about the cloud being what's nimble and the cloud being the place that we can do this stuff and secure, secure, do this stuff for less expensive. I'm wondering the shift on premise is not a shift. It's been around for but since before the cloud. But why are we seeing and top thing so much about the shift right now?
Yeah, this is.
Remember the cloud here was all about taking your existing workloads, your web servers, your email, your office productivity and maybe trying to figure out a different way to operate it. If by the cloud wasn't just public cloud, the cloud model pervade everything and so you know, the shift autonomous, automated elastic infrastructure happened with a set of applications that we understood well. The things we're building now look nothing
like that. A large scale generative AI system for an enterprise is arguably the most demanding and complex workload you will create in your lifetime, and so you know, we're you could argue that, you know, an OPX driven as a service model that gives you lots of agility but you pay by the drip is not a very good outcome. If the thing you're running runs seven x twenty four at enormous performance levels, you know you don't want that meter running. You want that meter to be predictable, and
so it's just a different class of workload. By the way, we're big proponents in multi cloud. A lot of the time, the best place to develop your AI system is in one of the cloud providers because they have a great tool chain. The best place to test it might be there, Maybe the best place to train your models might be there because you only need the infrastructure for a short
period of time. But the minute it becomes inference that you're putting it into production and it's using your data, which by the way, most of that is on prem even today, then it starts to become a very different discussion, which brings kind of this modern on prem architecture that we talk about with the AI factory into play as probably one of the more logical places to start.
Well, what is exactly the concept of an AI factory.
Yeah, here's the important thing for people to realize. AI is a new workload and by the way, it actually needs a new class of infrastructure. The type of compute is not CPUs, it's GPUs. The type of data is not traditional databases, it's vectorized data that lives in large language models. The kind of tools you use are different, and so what we talked about at Dell Technologies World last year last week was was not a new set
of products exclusively. It was you really probably need to have a separate type of infrastructure where your AI than you do for the traditional things that you do, those workloads that went through the kind of cloud migration we talked about, and what the AI factory is is it's
an articulation of what that infrastructure looks like. That it is accelerated compute, that it's a different kind of data architecture, it's a better and different type of networking architecture, and that it probably lives in a different footprint because it itself has different requirements. Than your legacy applications and the other applications you run in either a public cloud or a private environment.
And so the II.
Factory is how do you create a methodology and organize all the technology to put that in play, whether it's at a rack level or even an entire data center, that builds you the optimal infrastructure to run these new workloads that you're going to need. So think of it as just a paradigm shift. We're going to have to build new infrastructure for this new workload. It might be redesigning or optimizing what we have, and it might be in fact being your data center build out.
John. It's really interesting and I think we might have talked about this at the Bloomberg Intelligence event, but about you know, technology companies, they compete, they work with each other, and it was an interesting at Dell Tech World the keynote stage to see Jensen Wog and videos CEO up there, and so there's really you know, you can see the partnership. You can see there's clearly a show support from in
video to you guys at Dell. What is the nature of this partnership, especially when they make their own AI server racks, which makes you competitors to some degree.
Yeah, Remember, Dell is a unique company. We are obviously very large, and you could argue we're the largest technology integrator in the world. Now it means something to different people, but basically, like I don't build my own CPUs or GPUs, but what I do is I organize that technology in the consumable units of it that my customers across the world can consume. Now, when you look at what in Video is doing, they clearly are the provider of some of the better GPUs in the world. There are other
choices that we also work with. They also have organized their stack in a way that makes it very easy to consume, and I think that did a great job there. And so they have an early lead and it's importantly that they're they're they're making it much more consumable and
they're keeping the innovation cycle up. However, their ability to engage with the large enterprise across the world, they don't have the We own much larger salesforce, We have a global services capability, we have the largest supply chain that's secure in the world and technology and so and by the way, you also don't just need the GPUs, you need the advanced storage services. You need to integrate it
with your existing infrastructure. You need to talk to your existing data, which, by the way, rides pretty much on Dell technology storage systems. And so the nature of the relationship is, look, hey, you're trying to build an aifactory. There are some leading edge parts that absolutely have to be produced and on way to. One way to actually make sure that they happen correctly is to integrate them
into a system, an early system like what Nvidia does. However, those parts are decomposable and then they can reassemble into other form factors that companies like Dell can take to a much more scalable market. The announcements around edge and other areas. We have the ability to reach more customers than anybody in the world. We need partners to help us build technology to bring to them.
Is it a deeper relationship? Just got about forty seconds and we'll come back and continue. But is your partnership with Nvidia a little bit different? Is it a deeper one because you do have partnerships with other server manufacturers as well in others? Is it a deeper relationship or how would you call it?
Yeah, we worked with you know We work with a lot of companies, and what I will tell you is the first one.
It was the deepest.
Last year we announced Project Helix, which was the first time anybody articulated putting all the parts together into something people could consume. So it has a significant first mover advantage. However, you know, like you said, we work with a lot of companies and we have a lot of partners.
You want to continue with our guest, John Roses with US. Roses with us. He's a global chief technology officer at Dell Technology, still with us from New Hampshire. Hey, John, you know we were talking Tim and I in the break and just you know, curious about as people go right t him to build out their data centers, you do wonder how busy it kind of gets for Dell.
Yeah, I'm wondering how big John, the AI server opportunity is for Dell in twenty twenty five, and then how big it could be over the next few years. What are you folks talking about internally and externally?
Well, so I'm a CTO, so I'm not going to talk about.
Particular year, But are you really busy?
John?
I am extraordinarily busy, But let me paint a picture that's a little longer term, you know, and that is, look, we are, as I mentioned, we're in now year two of the AI cycle, the modern AI cycle, and year one was all about surprise, get organized. Year two is about the first kind of enterprise deployments and kind of the prototypes. What that means is that the enterprise buildout hasn't actually begun. And if we look at the size of the A market today and what's going on, it's
a pretty interesting market. What's happening is we're building the foundational technologies, we're training the large language models that this is a very robust ecosystem right now, we are developing the tool chains and as you can see from you know, just the state of the industry, it's a pretty exciting and a pretty significant shift that is in front of
the enterprise cycle. We're conservatively, you know, most most people over the long term, the AI cycle is about rebalancing a sizable portion of the work in the world into the machine layer, and so as that occurs, you know, it represents you know, sometimes we use the phrase we're in the training era. Now we're about to enter the infront zer for enterprise, which is AI gets put into production.
When it does, you know, you can.
Calculate imagine a world where you know a third of the work is now happening in a machine layer. It's being done by a machine an AI system. What does that look like? How big is that? Well, you can't calculate it accurately, but you know it's a gigantic number. And it's as big as the Internet build out, it's as big as the Industrial revolution. In many of the discussions that we have, the timing on it, it's an extended cycle. This'll be a twenty year cycle, but we're
about to enter that phase. And what it tells us is it's a significant amount of replumbing of enterprise. It's building out AI factories, it's rethinking your data strategy, it's rethinking your footprint in the multi cloud, and all of that tends to give it our breath and depth in our ecosystem, drag us into an awful lot of customer conversations in a pretty active world, even in advance of the significant buildouts occurring on the infront side.
So okay, so if we're only in year two when it comes to AI, where are corporates in their AI journey? How do they adopt their technology. We know Dell won a meaningful chunk of Tesla's AI roll out, for example, But where are corporates in their AI journey?
Yeah, the corporate As I mentioned before, last year, it was all about getting your feet on the ground, understanding the technology of earning what a large language model, earning what RAG was, and this year it's all about finding those first projects. The first projects are largely under development in most large enterprises, and some of them are emerging as chat thoughts and other services that we're starting to see. And there are definitely early examples of technology that have
been deployed as new offerings. But the full pivot where a company now declares that I am in the center of my business, you know, building my product, selling my product, servicing my product, engaging with my customers based on primary early in AI architecture, that is still work to be done.
And so so I think, you know, at.
This point, we're still right now in most large enterprises in the first proof of concepts, the first prototypes. But one of the things that's different about AI is it does not take three years to build one of those. You can go from idea based on the tool chains available to having something in production that you can start to really do, and we're doing that inside of Dell make your developers more productive in a matter of months.
And so the velocity of this cycle is something that we've never seen before, which means the gap between getting your feet on the ground and being in production and transforming your enterprise is not a ten year cycle.
Hey, John, I keep hearing you know you've said it, and I've had other guests, but when they talk about AI in inference, right, am I saying it correctly?
Yeah? Right?
How is that different from AI training? And is that kind of a new concept or is that just something more complicated when it comes to generative AI.
No, they're part of the same cycle that The idea behind AI is like you're trying to have a machine do some kind of cognitive work, answer a service, call, sell something, build a writ code. In order to do that, the first step is that you must have that machine have some access to the knowledge necessary to do that,
which is what training is about. The difference in large language models is that we've developed techniques that allow us to instead of trying to as human beings decide how to code, we've learned that if.
You just expose these new.
Techniques, these new technologies, large language models, to a gigantic set of coding, just examples of coding, they will classify them, organize them, and create a neural network. And interestingly enough, they will then be able to replicate that intelligence, that behavior. And so the training phase is about taking gobs of data. In the current phase, it's the entire Internet and run it into these models that create systems that can understand
or communicate human language, that can code. And all that is is them deriving from a gigantic data set the knowledge that's contained within it to create a set of skills. That's training. Inference is totally different. Inference is once you have that model, now you want to do.
Something with it.
You have a thing that can code great well inferences when you tell it to code a program to actually produce source code that does something. So they're just two habs at the same coin. One is the learning to create the capability. The other is the act of using the capability and production.
Hey listen, something I got to ask you. It's a story that's on the Bloomberg and this has to do with Nobel Laureate he's also an economics professor. We're talking about Paul Romer, and he was talking with our team here and he said, runaway confidence and artificial intelligence risks repeating the mistakes of the crypto hype bubble of only two years ago. He says, right now, there's way too
much confidence about the future trajectory of AI. When people project this phot I think they're at risk of making a very serious mistake. Do you think that there's too much confidence about AI? Do you think there's too much euphoria?
Two answers to that in the general population. I think it's a very confusing space because we have basically the human race is maybe susceptible to science fiction, and we think that what we're producing is artificial general intelligence, and we're producing these these sentient beings.
These are not sentient beings. These are technologies that yah.
Yeah, maybe sometimes we all believe there's a path to AGI. It's just not in the near term. That whole dialogue, and let's call it the consumer general market, which is primarily where most of the big public AI players play, is a risk because people think these have personalities. They don't understand the technology. When you go to the enterprise world, it is far more conservative. There is no enterprise in the world that's trying to build the terminator or a
sentient being. What we are doing is we're applying the techniques that we're pioneered in the public AI world in large language model space. Two very specific problems within a corporation that could benefit from it, writing code, finding your customers, engaging to solve problems.
Those are the things that are mattering.
So enterprise is kind of boring to be perfectly on versus what is in the part of the possible in the public world. However, as you know, our industrial complex is quite large and the impact is much bigger.
Great stuff already looking forward the next time we get to catch up John, Thank you so much. B. While John Rose, he's global chief technology officer Dell Technology, is joining us from New Hampshire. You are listening and watching Bloomberg Business Week Carol Masser along with Tim Stanovic, and this is Bloomberg.
Some disturbing information that we found our disturbing This next interview report by the global media company with a health and biotech focus. Marianne Liebert found that quote in nearly three quarters of the cases where a disease afflicts primarily one gender. The funding pattern favors males in that either the disease affects more women and is underfunded with respect to burden, or the disease affects more men and it's overfunded. So double whammy when it comes to the gender health gap.
Yeah. Meantime, in a recent report for McKinsey, it found that reducing the time women spend in poor health by twenty five percent could be worth one trillion dollars, in large part because health disparities disproportionately hit women during their working years. So let's get to it. There are huge disparities between men and women's health that definitely tim need to be addressed.
Back with us to talk about that gender health gap is Elizabeth Staddinger, managing board member at the sixty five billion dollar medical tech company Semen's Health and Ears, publicly traded one. I should note she joins us from Germany. Elizabeth, good to have you back with us. Thanks for staying up a little bit later once again to join us from Germany. Talk a little bit about what you're doing over at Semens Health and Ears to address this gender health gap.
As you already mentioned, in the opening, there is significant disparities. And if you just to make this tangible, if you think of the typical Hollywood heart attack where somebody kind of reaches to his breast, has this burning sensation in the arm, and everybody will start rushing and saying, oh my god, something really serious is going on. Let's rush that patient to the ED and take care of him. Unfortunately, when you're a woman, most likely the symptoms you will
be feeling are different. And that matters because the likelihood as a woman to be misdiagnosed when you're having a heart attack is fifty percent higher than it is for men, and the likelihood that you will actually not make it when you admit it to the hospital is actually twice
as high as it is for men. And it gives you a sense of the huge gap that we have and the huge disparities we have when it comes to looking at women's health and men's health, and the big biases which are deeply ingrained into how we do things.
Do men and women medical professionals are they biased equally? In other words, do women doctors, women nurses also misdiagnose or underdiagnosed women.
That's actually a very interesting question. And if you think of what people learn when they get their training bit in nursing or in medical school, the typical default everything starts from in research, but also in the training in mid school is the man and the women. Sometimes is considered that's the abnormal version of the human being, and that leads to a situation that both men and women
may have that bias. At the same time. There is some data which shows that on average, women are more sensitive to some of these differences than men are, just maybe because they've also experienced it themselves.
It's funny even kind of preparing for this segment working with our producer Elizabeth Cedric, Like you know, even research that's been done is the baseline is off of what's going on with men, and it's just interesting, like trying to find data that everything is kind of off of that base, if you will. It was a little discouraging, to say the least.
Right, there is a lot of bias which is somehow built into the system in R and D funding. You mentioned that in your opening statement, but also in clinical studies. For many, many many years, women were completely underrepresented in the clinical studies, which then leads to medication being approved based on male samples only.
I'm wondering, Elizabeth, about the role of technology here, and we've had you on to talk about this in the past, but I'm wondering about bias when it comes to AI or when it comes to diagnoses, because if we talk about bias, it's something that's within every person and it's within doctors as well. Can we remove some of that bias if we use technology instead to make these medical decisions?
And that's actually a very interesting topic because if you think of bias in people and bias in physicians, there is a lot of bias in the data which is used to train AIDS.
I was afraid you were going to say this, and this is.
But the good news is there is a lot of ways of dealing with that. I mean, we enceemental healthy mes. We have more than twenty years of working on deep
learning AI based algorithms and dealing with patient data. We've built a huge database of more than two billion data points, and we can make sure as we're training our algorithms that we have a well balanced and diverse sample that goes into creating the algorithms and the prediction of the results that are based on the software what faith do There even ways of kind of adjusting for that, but you have to make a very conscious effort.
That's what I was going to say, what faith do you have in that happening? Because I feel like, you know, you know this, you know where I'm going to go. These are not new issues, concerns, problems in terms of certainly when it comes I feel like with women and men and health and data. So what hopes do you have that that you know that that effort will be made to make sure that women are incorporated into the data points.
I think there is a very strong link to how do we kind of ensure and also demand that the data that goes into the training of the AI does
have the right balance and the right mix. So regulatory authorities can help here, but also companies of course can make sure and we, for instance, we put a lot of effort and energy into making sure that we use high quality data which is vetted where we are confident and certain that the input we provide to the software and to the algorithms really reflects both the disease we want to see as well as reflects a good sample across not only JENDA but also other dimensions which matter.
So where do you go for the data to find unbiased data well, ultimately or do you just correct for it?
You essentially correct for it. So when you pull together the sample you want to use, you make sure, Okay, how many women do I have, how many men? What kind of Also do I have enough representation of Asians of India and of African of Caucasian patients in order to make sure that we do not kind of over index on any of those dimensions.
Yeah, fascinating if we don't get this right, I mean, what's at stake. It sounds like a lot in terms of health of women overall. And I wonder how much in terms of gaps between developed world and developing world, you know, where that really maybe has an impact or maybe it's everywhere.
I would say it is really everywhere. You know, women are not simply small men, regardless of where you go on this planet. And I think it's also reality that we cannot improve global health if we ignore half the world's population. And there is not only a risk speaking now a lot about the risk, but there's also a real opportunity because if we manage to remove some of these barriers and can improve the health of women in
the world globally. It will help everyone rise. You mentioned the one trillion of GDP by twenty forty that we could be tapping into if we did a better job at removing the barriers and the disparities in health.
Elizabeth, thank you for joining us an important conversation. We always love it when you take the time to join us, and again thanks for stand up late over there in a Germany. Elizabeth Stattinger is managing board member over at Siemens Health and Ears. It's a publicly traded medtech company sixty five billion dollar market cap