 Live from Las Vegas, it's theCUBE. Covering Discover 2016 Las Vegas. Brought to you by Hewlett Packard Enterprise. Now, here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. We are here live in Las Vegas for HPE, HPE Enterprise Discover 2016. This is SiliconANGLE Media's flagship program, theCUBE, where we go out to the events and extract the signal from the noise. I'm John Furrier, my co-host Dave Vellante. Our next guest is Robert Young-Johnson, who's the EVP Executive Vice President and General Manager of HPE Enterprise Software Group at Hewlett Packard Enterprise. Welcome back to theCUBE, great to see you. Thank you. Thanks for spending the time. I know you're super busy with customers. You're doing a lot of customer visits. And now that you're running all the software, software is powering all the innovations coming out of HPE. We're seeing with IoT, certainly that is awesome. You've been kind of retooling the haven on demand now that's kind of growing, but AI is hot. Everyone's got some cool, shiny toy. Zuckerberg had virtual reality at Facebook F8 conference. Google I.O. was all about machine learning and AI. What are you guys proposing? What are you guys doing here? What's your big, shiny data toy that you guys are sharing? Well, we've got a lot going on in this particular world, but I think it's important to start by doing a little bit of definition of what people mean by AI, because the point we're going to talk to, they've got endless variants of what it actually means, ranging from the movie X-Machine, and the future generation of digital assistants right through to much more prosaic things. And I tend to be more on the prosaic side. I actually think machine learning and deep analytics are going to get embedded into almost every business process, every application that gets written out there over the next five years. In fact, I set up on stage, I believe that five years from now, almost every application is going to start with data. You're going to load an analytics framework on top of the data, and then the business processes are going to flow out of that. And that's, I think, a very much more pragmatic view than maybe some of the more fanciful views. I don't rule out the more fanciful views, but I'd like to focus on what we can do now and what we can do over the next five years. It's interesting, I was with my son on his 21st birthday this weekend, took him out to San Francisco with him and his friends, some of them are computer science students, and I asked them, I hear what they think of the computer industry, and I asked them to define AI, and they all had different answers. So I want to ask you, what is your definition of AI? What is AI, artificial intelligence? I'm going to switch the terminology, but news terminology I use quite a lot, which is I call it augmented human intelligence. So rather artificial and human intelligence, which sort of implies a distinction and a separation away from human beings. I think the real powerhouse over the next five years is going to be what I call augmented human intelligence. That's using compute power to help people make better decisions, faster decisions to assimilate vast ranges of data they can't assimilate today, and just think of the world in a different way. At one level, imagine walking around the world and everywhere you look, there's streams of information coming to you, we're giving you background to what you're looking at. You're looking at a store, it's telling you about what's going on in the store, it's telling you about where the store was created. You're looking at a person, you say, by the way, you met this person a year ago, their name is Fred, they've got three daughters, all that sort of constantly flowing at you. I need that app. Yeah, I know, I've said it already. But that's what I think of. That's why I think where the next phase of this is. And I think if you think there, you can suddenly see how all the tools fit. And I agree, I think the IAI term is just misstated, right? I mean, we don't call an industrial machine providing artificial strength. So the question is, is the intelligence that these machines are providing, is it real intelligence? You call it augmented human intelligence. Are the machines becoming intelligent? Well, it depends how you define intelligence. I mean, I think that are they finding patterns that we probably couldn't find for ourselves? Probably yes. But maybe that's because we're not capable of actually ingesting the vast amounts of information coming off a sensor or coming off a log file or whatever. So in that sense, you can call them intelligent. You know, the big breakthrough is going to become when they become, quote, you know, using the terminator terminology, self-aware. I'm not sure. I mean, maybe not in my lifetime. I would never say never to something like that. That's why I think this concept of, you know, taking the decision-making power that we have as human beings and augmenting it to allow us to make better decisions, faster decisions, more nuanced decisions is where the real power of this platform lies over the next three or five years. The humans are still the last mile here. So you called this a huge shift yesterday in your keynote, this machine learning, deep analytics. And you described the old days, you had a business process, you had a database. Yeah. And you talked about changing the paradigm. What's changed? I think the change is that people are saying that, you know, the core entity on which I need to work is the data I collect. And often that's random, often it's very ill-defined. What you need to do then build an analytics framework on top of that that's going to expose the value in that data. Having done that, you can start to build your new business processes on top of that analytics framework. It's a different way of thinking about how we create workflows and applications. Because in the past, you might have innovated, but it's essence you're always tying it back to an existing process. And if you look at some of the new age compute companies today, they're not doing it that way. They're saying I'm collecting this corpus of data, I'm not actually sure what's in it necessary or whether it's valuable. I'm using the analytics framework to expose the value and structure and then I'm writing my business processes on top of that. And that's allowing them to create breakthrough approaches to being competitive in a particular industry. Robert, last time we chatted in Boston, you were on the soap box, really kind of talking about the approaches of data. I think now we've come to, it's not about, you know, why should I use the big data or data, but how? So I got to ask you a question, and then get your thoughts on this. The pattern we're seeing is the early approaches are, what happened? Look at the data and find out what happened. Look to the future and figure out what's going to happen. But now it's about what's happening today. So the now moment is really now emphasized with the speed of data. So you mentioned speed in your keynote, fast data. How is the now, what's happening now? Cause that's a data problem that requires some streaming, that requires data sources to be accessible and be surfaced. What is that about and how challenging is it and what approaches are people taking for real time in the moment data and analytics? I think I'll make a little distinction. I think real time is important, but perhaps not as important as people think. Sometimes near real time is good enough. It doesn't have to be the microsecond, it could be within five seconds or 10 seconds. But what's much more important is the access to the analytic capabilities and making sure those are widely accessible as opposed to being the province of a few data scientists who generate algorithms. And that's what we've been trying to do with our platform, is to create a lot of embedded intelligence but allow people to access it easily through Stanford straightforward APIs, which means people can write code quickly to solve business problems. So be operationalized and scale. Yeah, be operationalized and scale on a cloud infrastructure. That's what we're setting out to do. A lot of approaches we're seeing to big data in the industry are taking a slightly different approach. I need a PhD, I need a highly paid consultant who's going to come and help you analyze your big data problem. It's then going to use some tools to develop algorithms that can help you tackle that problem. And I'm not sure that's the way this is going to spread out in time. Well, it's kind of been the only way to do it in the last five years, right? I mean, the industry has sort of caught up to the vision that you're putting forth. Yeah, that's right. So how are you putting that vision into action? Well, this is what we talked about yesterday with our Haven on Demand platform, which isn't new, but we've actually augmented it very substantially over the last six or seven months. We've extended the range of APIs. Originally it was very much textually based, extracting value and content out of text, which is actually pretty difficult. We've extended that to image, we've extended that to voice. One of the things we showed on our little video yesterday was the ability to do real-time transcription of broadcast so that you could index it. And so if you're going to want to go back and look for it, you can find it. Now you may think it's good or bad. You can use that job. They'll find out what you said three months ago. We have the transcripts. So on the big data platform, I noticed that we had Chris Selin on yesterday, and I noticed his title change. It's big data platform. Doesn't say vertical anymore. I know it's all kind of come together. You've done the consolidation of the brands and also the products. Right now there's a platform war going on after the big guys are fighting it, but also the entrepreneurs are developing tools. And we're certainly seeing in Silicon Valley the softening of the market where those can't be standalone companies, but they want to join an ecosystem. So we've been hearing and seeing startups and or growing companies building, you know, up to a hundred million dollar business on building a great tool. But the key thing now is to join an ecosystem. So what's your message to those folks out there? Why HP platform? How does a tool developer fit into your vision? Well, we're building a platform that is open, extendable, easy to access. We're making the platform available to developers at no charge. I mean, there'll be, eventually, you know, if we think people are running real world applications and making money, we make a, hey, we'd like a little bit of a share of that. But we started that way because my vision initially was that, you know, this had to get traction. This had to get adoption. That's how you build out a community. You know, if you start with the business model and then worry about your technology, you may never create that community. So we started to start with the technology and frankly, we'll build the business model down the track. But, you know, building an extendable API structure that other people can add APIs to. The nice thing about an API structure is that you don't hide anything. If someone says, I've got a better image detection algorithm than you guys do, bring it in. Expose it the same way. And that's good because it's going to end up creating a better class of functionality. Okay, so you made some comments in your keynote about Watson, this is great. So, because it's a good reference point for everybody. But there are others. You know, Google, Facebook, Amazon. Talk a little bit about how you guys see that augmented human intelligence. What's the offering there? Can you compete in that space? What's your intention? Well, I'm totally confident we compete. But what I did on stage was try to draw out a, maybe I was a little not nuanced enough. A little bit of suckering. The distinction of approach that's being taken. I think IBM and, you know, I admire the company hugely with Watson are starting, they're collecting data. You know, they're buying companies like Weather.com and things like that. They want to create corpuses of data. And then they want to be able to sell the value of that data onto their clients. And I think it's a sort of hybrid platform consulting sell with data. Our approach is going to, we need to provide a platform that can get it threaded into almost any solution, any application that anyone wants to build. And that's a different approach. So that's why we've really focused heavily on exposing these APIs and making them available. So would you say IBM is more pre-packaged? You're more enabling? I don't think it's pre-packaged. I think it's more consulting and business process land. Custom, you're saying. It's custom. Your client is your argument there. Because IBM would bristle at that, but. I'm sure they would. But it's been a criticism of IBM. They've made a great business out of it, but you can't just shed your services-led value proposition overnight is really what I'm saying. I'm sure. And again, I've got huge admiration for what they're doing. They've certainly produced a huge amount of focus on the section, on this particular business segment. This is good for everybody. This is great for everybody. So we've been seeing a trend kind of boiling up. And I haven't seen this in my lifetime since I was right out of computer science and undergraduate was the developer market and the enterprise is hot right now. So we saw this in the 80s and 90s, early 90s and kind of kind of went south a little bit, kind of flattened out. But now we're seeing an explosion of DevOps. Your IoT announcement today was pretty significant. Okay, obviously part of the GEE. But now you're seeing real energy in what we're calling enterprise developers, whatever that means. Basically non-consumer developers. It's hard to do in the enterprise. What's your take on that? Because do you see the same thing? Do you see more action on enterprise developers? And how would that compare vis-a-vis standard web or mobile developer out in the wild? I think it's huge energy right now. And I think we're trying to elevate the developer right up our priority stack. And also recognize developers don't work the way they maybe did 10, 15 years ago within the enterprise. 10, 15 years ago there was very much a center of expertise that said these are the tools you'll use. This is the techniques you'll use. And that's what developers did. Today developers tend to choose their own tool sets. And you've got to respond to that. So we made a big announcement earlier this week of our new generation application lifecycle manager product, ALM Octane. And it was really designed to actually recognize that developers will choose their own tool chain. And you need to actually incorporate and integrate their tool chain into the way you think about application lifecycle management. But at the same time you need to actually have a workflow that collects information and ultimately will allow you to predict how successful applications might be in production based on how they go through that workflow. I want to drill down on that. This is really kind of a nuance but kind of not yet mainstream discussion point what you just said. And we're seeing similar thing with open source. Now a tier one citizen, everyone's using open source. The tools or the tool chains are out there. GitHub, we use it, we're a small company. But as enterprises start adopting these I'll call shadow IT like kind of pockets of developers, it's going to be a management nightmare. We're seeing that potential around the corner. Do you see it? I mean, is it a nightmare? How do people make sense of? I think it's an opportunity at the risk of being repetitious. I mean, this is where analytics starts to thread in again because whatever tool chain you use, you're going to be creating information out. Did this piece of code succeed in this test cycle? You know, what's his dependencies? All that sort of stuff. It's going to, whatever you do, it's going to create that. And we think there's a huge opportunity for an analytics framework along that to start to give people a more common view of what's going on. And our dream ultimately is better predict before an application even hits production whether it'll be successful in production. And that's what we're trying to do. I think that's a great opportunity. I think that's not really been identified yet in terms of critical mass yet, but that brings up the whole algorithms, machine learning trend. So the magic that's out there which is called machine learning is just magic. You put machine learning on anything, it's magic happens. Well, now you've got to have the skill to develop the algorithms. And even a lot of work in that area. For example, in the security area, which I've talked about for a long time, I said security becomes the ultimate big data problem. And so what we've been doing is putting a huge amount of effort and dividing specific algorithms to detect certain types of threat. Like when one we announced was DNS malware, just by looking at the pattern of interaction of your device with the network, we can determine whether you've been infected. You don't need antivirus software, we just know. If your PC starts talking to a website in North Korea, it's probably infected. Now, easy on that one-off instance, but put that across 200,000 PCs or other devices in a network, hundreds of billions of DNS queries, it's a really tough thing to do. And we've developed algorithms to do that with zero false positives. I want to talk about your business a little bit. You've de-levered some, tipping point sold to Trend Micro. It's a very important component of HP's business, the software business. I asked Meg several years ago, what's more important, internal R&D or M&A, and she said nothing until we pay down this debt. So now you have HP, the balance sheet, looking much better. I'm sure you got your shopping list, but talk about your business, the objectives of that business. How do you get that humming again to really be a driver of profitability for HP? I mean, the first thing I had to do was clean it up, which is why we ended up looking at very, very hard at every component of our business that, did it fit our overall value proposition or not? And if it didn't, we should really dispose of it, which is why we sold tipping point, great product, intrusion protection. I think Trend Micro will do a great job with it. Great acquisition for them. Yeah, great acquisition for them. We had some legal workflows and stuff, didn't we? And now I think we have a very, very coherent portfolio. And the theme I'm bringing to that portfolio, and again, I'm being repetitious, is my mission, my vision, our vision of HPE Software is to bring machine learning, deep analytics to three areas where we have specific content competence. That's IT operations, the world of the developer, and the world of security. That's what we do. And our products all need to fit into that category. Now, in some cases, it's going to be an evolution, but that's where we're headed. Service management, it's going to become an analytics framework that's designed to stop service tickets being created in the first place, rather than manage the workflow associated with them. Operations analytics, the security stuff, that's our big differentiating theme going forward. You mentioned the use the term democratizing machine learning, is that? I hated it in city, you know what I mean? It's overused, but you know. It's overused, but what does that mean? Because you're essentially, you compare it to IBM in a way, that was kind of what you were saying. That means moving it down from the PhDs and making it much more of a utility for everybody. Is that, would that be the definition? No, absolutely. And I mean, anyone could get an idea on havenondemand.com, they don't need to be a developer, they can actually just play with it. I mean, the way the website's been set up is allow people to try the functionality without having to be deep PhDs on statistical methods or whatever. And that's what I meant by it. Well, I always ask you this question because you're so awesome on thought leadership and knowledge for the folks watching, the younger generation, computer science students or folks in the career looking to really get trained up and or immersed in the data IoT software world that we're in. What advice would you have for them in terms of approaches, mindset? Give us the end of year, graduating speech. It's funny you say this, sunscreen or? Wear a lot of sunscreen? I was meant to be asked this very specifically by some 20 year olds recently, your family, you're connected to our family. And they were sure I go to business school, should I do this? I said write some code, get a project on GitHub, make it open source, write some code. That's the best way to get into this. And so what should I choose this language? I don't care, just write code. Whatever is best. Whatever you think's best. Then start to try to use some of the capabilities we have, like some of the APIs we're exposing, write some cool code that's going to do something that someone never even thought about doing. We had on stage this morning, you know someone who came to one of our hackathons used our code and pushed it at PDFs issued by governments on the performance of those government departments and the way they spent money. And they were able to do that, take it out of the PDFs, put it into Excel and then on that do some analytics that sort of expose some really scary things. Like 60% of the money was going on office supplies. That's a boondoggle. I think it's phenomenal. But it starts, you've got to write some code. Yeah, translate that service. Final few seconds, I want you to take a minute to share with the audience what is Haven on Demand and your vision for HP software? Okay, so for Haven on Demand, Haven on Demand is an exposure of the very deep analytic capabilities we have in both our Vertica and Idol platforms as a series of easy to use, easy to consume APIs, free to developers, free to anyone who wants to just sign up and use it. Also available to enterprises to deploy on-house that's what they choose to do. For overall HP software, our vision is very much rooted in this idea that every app, every solution becomes a machine learning and deep analytics. So you start by building a great platform. I think we have that. You then deploy that platform as far as we're concerned of those three areas where we have really deep competence which is IT operations, it's the world of the developer and it's security. And then we make the platform available to anyone else who wants to use it in their use cases. That's what we do. Robert, thanks so much for spending some of your valuable time here on theCUBE. We really appreciate the insights and the data and we will certainly put this data to work for you and you're out there, write code. So we'll be right back with more media's code videos here at HP Discover, coding it all for you. We'll be right back. I'm John Furrier with Dave Vellante. We'll be right back. You're watching theCUBE.