 Live from San Francisco, it's theCUBE, covering Google Cloud Next 2018, brought to you by Google Cloud and its ecosystem partners. Welcome back to the Moscone Center in San Francisco, everybody. My name is Dave Vellante, and you're watching theCUBE, the leader in live tech coverage. This is day two of Google Next, big cloud show, about 20, 25,000 people here. Barat Rao is here, he's a PhD, PhD in ML and AI. We're going to talk about that. National DNA leader for healthcare and life sciences at KPMG. Barat, thanks so much for coming on theCUBE. Welcome. Thank you for having me. Pleasure to be here. I'm really excited about this conversation. Obviously, you're an expert in this field. It's a quasi-passion of mine, healthcare, just for many, many reasons, but it's an industry that's so ripe for transformation and with such massive potential. But tell us a little bit about your background. PhD in AI and ML, quite fascinating. Talk about that a little bit and what your role is at KPMG. Sure. So, I graduated in the early 90s with a PhD in machine learning from the University of Illinois, and I joined a large multinational company working for the research lab, and I basically founded the data analytics group, did research in several industries, started working on healthcare data and just fell in love with it around 95 and I've focused on that, then became the GM of a company doing that for about 10 years and joined KPMG. Healthcare data is the most complex, fascinating, rich data that is. And when I say healthcare, I mean it with a capital H. It includes the provider, the payer, life sciences, the physician's notes, images, genomics, et cetera. And it's so rich and the ability to process and transform it is going to change, not just healthcare but change our lives dramatically in the coming years. All the ingredients to make it happen are here. So, in KPMG, I'm leading the healthcare and life sciences analytics team. Our approach is that healthcare is going through a once in a generation revolution and it's going to be fueled by several aspects. One is economics. People have realized that you can't keep spending 15, 17, 19, 20, 25% of the GDP on healthcare. That curve has to bend and come down. The second is there's a rich regulatory environment which is forcing provider's payers' life sciences to adapt. The third is the rise of the intelligent consumer. You know, this thing didn't exist 10, 12 years ago. You can now go on the internet and figure out everything you want to know and show up in your doctor's office at the market. So how are we going to deal with this? The answer is technology and there are two things which I think are making this possible. One is the increased availability of digital data in healthcare, the high tech act in the US. Basically, what I call the e-modification of America has created lots of healthcare data. The same now with the provider's life sciences companies. The second is the cloud machine learning and AI. You now have the capacity to crunch unheard amounts of data and do things that are completely unimaginable even a couple of years ago. And it's the power of those two technologies which I think are going to fuel the revolution for a new healthcare system. A healthcare system that's driven by technology, not a better healthcare system, but a new one where physicians' decisions are impacted and patient experience is improved by the use of technology to make things easier, better, faster. Hopefully it is better, but so now the push to EMR and the incentives around that and meaningful use, has that in your opinion, and maybe you could talk about this, set up the data model so that we now can exploit the data to create a technology-driven healthcare industry? So it's a great question. So part of it was people said when they did high tech and they did EMR, we're going to collect all this data and magic will happen. Well, when you collect digital data and you start to put it together, you have a lot of digital data. It's the ability to crunch and process the data. And the problem is the data isn't collected for the purpose of analysis. The data is collected to treat the patient. Or the data is collected to conduct a clinical trial. Or the data is connected to do the billing of the patient. It's not collected to say, right? And so 1% of data or 0.1% of data is clinical trials which is data which is collected just for the purpose of answering a question. So I believe there's going to be a shift from what we call evidence-based medicine which is mining, you know, which is clinical trials where you have a careful hypothesis, take 200 patients, do a trial, do what I'm calling experience-based medicine. You know, where you have all this vast amount of data that's been collected lying in a data swamp. But now with the technologies we have, big data, the ability to process unstructured data, images, video genomics, that's what's going to take that data and make it actionable. And that's the revolution that we're in the middle of now and we are right at the start of what I believe will be the tipping point. And that will complement that brute force clinical. You see it as eventually replacing that but through inference and predictive analytics? So I don't think you're going to replace a doctor. Not in my lifetime, maybe in my kids. I don't see that happening because there's a lot of, you know, I think what will happen though is a lot of the opportunity, things like second opinions. Did you miss stuff? There are a lot of mistakes that machines don't get tired. Machine learning, Alex, look at all the factors. The other thing that I think machine learning and AI do, it will actually tell us when not to treat, right? Because right now we, you know, everyone heard the stats, 40% of costs in the last year of life, 20% in the last two weeks. But you don't know when you're, but using machine learning and AI you can figure it out and you can then offer it as a choice to the patient, right? In terms of easily explainable ways. So I think that there is going to be, so I think that transformational comment too is first advising the physician. And my vision is, and I don't know if it's five years, 10 years, or 15, the liberation of the data where the patient has their own data. That's my vision of healthcare. And it's being advised by a number of electronic savants and the physician is basically coordinating that information and delivering the best care, but it's around the patient. And the patient is orchestrating and somewhat, you know, maybe more involved, taking more responsibility for those outcomes. Absolutely, right? Because that gives you the ownership and you understand immediately the consequences of missing your medicine, the consequences of sneaking that extra brownie in if you're a diabetic, right? And it's pretty clear. It's not just about healthcare data, right? Today our smartphones know how much we travel, how much we sleep, how often I'm on the phone. How many steps we take. How many steps we take, you know, what my stress level is. How many, am I triple booked three times, you know, throughout the day. And how long before now you start to integrate what I or my spouse bought at the grocery store or which restaurant I went to and ate. Did I go to McDonald's or did I go to the True Greens Eat Free, right? And all that information can be put together to really become an electronic life record, not just an electronic health record, which is ripe for mining in this new age mobile, social, and so many different ways, right? The patient journey is going to change from an isolated visit to a physician to a continuous seamless integration of systems. It's sort of like Google right away here, feel like one system, the great user experience, regardless of whether you're dealing with the doctor, the insurance company, your pharmacy, or your retail store. And I think that's the transformation that will happen. I love that vision, Bruh, because it's really not about, you kind of referenced it before, and I think I now get it, not necessarily creating a better healthcare system, it's creating better health. That's right. You put it, that's great. Can I use that line? Please, I mean, you just gave it to me. I appreciate that insight. No, but that's really what it's all about. I mean, that link between, you know, you're the old saying you are, what you eat, the link between, say, for instance, diet and other stress factors in a holistic system, creating better outcomes. That's a fantastic vision. And the cloud is, and you know, the enablers, we're sitting here, as I said, it's the cloud, machine learning AI, and data. And what's happened now is we collect data, we have walking sensors, you know, people talk about variable sensors, our iPhone is a variable sensor, our smartphone, or Android, where we collect, as you said, the steps we take. And how much longer before we're starting to collect other information, right? Everything, and I know there's a big brother aspect, but just setting that aside, is it pure scientist? If it can be used to make me healthier, better? That's one thing we all want, right? If we're sick, we want to get better. If we're well, we want to stay well. Yeah, I mean, the big brother aspect, we could talk all day about that, but at the end of the day, are you willing as an individual, patient, to make a trade-off between that big brother, you know, watching in your health, and I personally would say, I'll take the latter. Well, I'm going to say this, and I'm going to irritate a lot of my friends who are advocates of patient privacy. I haven't met a sick person who cares about their privacy. If you sat next to a sick person on a plane, you know their entire life history before you get up. People want to get better, right? So I think that's overblown. You don't want to misuse, you need safeguards. But yeah, if my data can help me, or someone else. Especially if I'm more in control of that data. Absolutely. That's a big piece of it. I want to come back to the data model. We do a lot in theCUBE with chief data officer events, and CDOs were early on in the healthcare industry, certainly financial services and government as well. And they came out of sort of a back office role, a lot of focus on compliance and governance, and certainly data quality. Are you seeing proactively healthcare organizations with a capital H, as you defined it? Really being more proactive about getting the data model right. Not withstanding EMR and some of the other, you know, magic expectations there. But is there a proactive effort going on to really get the data model right so that we can apply technologies like cloud, AI, machine learning to that data? So the two things happening, and one is exactly what you said, right? People are saying if we get a lot of data, we want to make sure it's high quality so we can trust the conclusions we draw. And certainly, and to me, I think, and then there's the other group that says, wait a minute, we're sitting in Google. Is there a noisier, messier piece of data than the internet, right? I mean, it's filled with dubious information. Most of it is lies, mistruths, falsehoods, and yet it's turned into the single greatest source of knowledge in human history. So you have both camps, and how do you reconcile those? And I think there's an important distinction people miss. If I want to do learning and use machine learning and AI, it's fine if my data is noisy. If I have a lot of it, and that's what Google has shown, and Microsoft and Bing have shown, that you can take incredibly noisy data and make it worth billions of dollars, right? With the internet. So the learning of machine learning of machine learning and models that you want to use, that data can be noisy as long as there's a lot of it. And obviously the noise is not what we call systemic, right? And it's random. You can learn very accurate models and very useful, actionable models. And they have the value that you can apply them to noisy data, right? Because they will learn. But then when you're actually making life and death decisions on an individual patient, or you're trying to make decisions about at risk, going at risk on an insurance contract repair error, is this drug the right drug for your second choice? You need accurate data. So you need both, right? It depends on what the use is, right? So one is my point of experience-based medicine. We have this swamp, this ocean of data. We need to mine it to get the information. We can't worry about cleaning it because who cleans the internet? Nobody. The machine learning algorithms do it if there's enough. But then when you apply it to one patient to make one decision about with life and death implications, you need really, really good data for that patient. You need their genome. You need their lab values. And you need to make sure they're correct. So there's the, you need both. And I think Chief Data Officers understand that their times when data quality is critically important. Come back to your question. And their times when we just want to get a lot of it and get the signals out of the data. I want to ask you about, you know, the leaders in this space. If you look at the, and you know, this might not be true with Facebook anymore because of the whole fake news thing. But generally speaking, if you look at the top five companies measured it by market value and granted the stock market will fluctuate, but they're data companies. It's Apple, it's Google, it's Facebook, it's Amazon, it's Microsoft. And they're all like approaching, and again, notwithstanding Facebook's fake news problems, the trillion dollar market cap. Now they've surpassed financial services. They've surpassed oil and gas. And I think it's the intrinsic recognition that data is the core asset. And so are those the companies that are going to lead this charge and this transformation from a technology standpoint? It's a great question, right? The traditional model. If you look at each of these companies, you know, maybe with the exception of Apple, that transformation came, they started as startups and then they grew at scale and, you know, knocked the established players of the, you know, Microsoft took an IBM, Google took on Yahoo, right? And Apple was going along and then reinvented itself as the device machine, Amazon, right? So is the new disruptor, the question I guess you're asking is the new disruptor going to be a startup or is it going to be one of these four? Obviously I don't have a crystal ball on this, but I think that to have impact, you're going to have to do stuff at scale. So there are two answers to the question. I think to really do stuff at scales, to say we're going to take on the US healthcare system or we're going to take on Singapore and we're going to take that to, you know, and we're going to help drive costs and help the 15 million patients there or all the patients there. I think you're going to need one of the big players, like, you know, like Amazon and Berkshire Hathaway and J.B. Morgan, that really interesting combination to say, let's build an experimental system with a million patients, right? That's the kind of things right now you have these big players doing and they are very well positioned to actually cause that revolution to happen. The question however with the big company is, you know, speed, are you willing to cannibalize your own business models? You know, Microsoft is an infrastructure company, let's take that as example. Well to be a healthcare company, it requires more than being an infrastructure company, right? Google's a search company that does ads. Healthcare doesn't run on ads. You need to change your business models to do healthcare. Same with Amazon, right? So are they going to do that? That's a question of vision leadership and will. That's one. The second piece is they have also, by creating the cloud, by making it possible for a guy in a gal in a workshop to do the things which earlier you could only do if you had a thousand computer data centers, right? You now have made it possible for the insurgent to have remarkable impact, right? So they've almost created the tools that allow smaller companies to come in and be disruptors because of the scale they've made available. I think that is a great insight because you know, you reframe my question, is it going to be the big five or the startups? And the answer is both, right? You just gave me the answer. You said both because of cloud. Cloud is not just about scale. It's not just about economics. It's about the ability to give these tools to startups and attract that innovation. And that's what, well certainly Google, Amazon, Azure have. I would presume Alibaba and some of the other China cloud guys have it as well. And then disruption is another thing that you struck a chord with me. It seems that Silicon Valley broadly defined if I can include Seattle clouds in there, have a dual disruption agenda. Maybe it's not their agenda, but possibilities. Horizontally with technologies like cloud and AI and machine learning, but also vertically now into these industry stacks, financial services, manufacturing, automobiles, certainly healthcare, ripe for disruption. You mentioned, you know, JPMorgan Chase and Berkshire Hathaway and Amazon getting together and formulating an alliance to essentially disrupt healthcare. Who would have thought that possible two years ago? Digital enables that. Yep, right? I mean that's a fascinating matrix that we just described, your thoughts. So, I mean, I think you said it best, right? I mean, you have the tools for disruption made available, but I think the big players we'd be, they have the opportunity. All three of them, plus Apple, let's not forget Apple. Apple has the device, right? Which you walk around with. They potentially have more information about you that anyone else on earth should they choose. They know which store you go into. I mean, they choose to, soon they'll, with Apple pay, they'll know what you buy, you know what. So, they all have unique access to data and unique ways of communicating the patient. To me it's a matter of, I think they could do it with the right business model and the right will. It's getting that business model right. How do, each of them have different strengths, right? So there isn't a one size fits all that would work for anyone, right? For some it's a partnership like Amazon partnering with Berkshire Hathaway. Let's not forget Berkshire Hathaway started off with an insurance company. Mr. Buffett knows something about that space which drives economics and healthcare. And, you know, J.B. Morgan Chase not only is financial, they're experts in blockchain which is going to be important, right? So there's this picture. Then you've got Microsoft, which is everywhere. Amazon, which seemingly, you know, with Whole Foods and Bill, just these fascinating changes that are happening. And of course Google, which, you know, is now a verb in our language. You know, my children cannot imagine a world where they can't answer a question tomorrow. And now with the intelligent home assistant they have being able to, they each have different ways I see them being able to transform healthcare. It requires the right business model and it requires the right, the will in some sense because part of it requires doing something different from what has got you to be so good. Because that same business model which you've refined to become a trillion dollar company isn't going to enable you to build a second trillion dollar healthcare business. Which I think is there, right in that grasp. Okay, last question we've got to wrap it. Will machines make better diagnoses than humans? Machines and humans together will make significantly improved higher quality diagnoses and treatments. The main value I think also comes in treatment. So not just a diagnosis of more treatment than humans. And it's through the elimination of mistakes. It's finding out the right drug that works for me versus what works for you for my condition versus your condition. And it's providing that information to the doctor and the patient. Let's not forget the patient. This revolution is going to change the doctor's oracle from sort of the doctor as coordinator. I mean, there will still be the expert. There'll still be God. They come from a family of doctors. Doctors don't think they're God, they are God. But they, I think that's the fundamental change. This is what my old computer-aided diagnosis. It's not computer diagnosis. And it's the physician and the machine working together with the patient to create a better, healthier, healthier patient. I love that answer. You know, I had the pleasure of interviewing Gary Kasparov, the chess master earlier this year. And when he lost to the IBM supercomputer, he didn't just give up. He said, you know what? I'm going to beat that supercomputer. And the way he beat the supercomputer is he got humans and machines together to beat the machine. Today, the greatest chess player in the world is not a machine. It's not a human. It's a human plus a machine. Exactly. A fantastic interview. Thanks so much for coming on the show. Thank you. I really enjoyed it. I appreciate it. All right, keep it right there, buddy. We're live from Google Next. You're watching theCUBE. Stay right there. Thank you.