 From theCUBE Studios in Palo Alto in Boston, connecting with thought leaders all around the world, this is a CUBE Conversation. Hello, welcome to theCUBE's coverage of HPE Discover 2021 virtual. I'm John Furrier, your host of theCUBE. We're here with CUBE alumni, one of the original CUBE guests, 2021 back in the day, her president, her chief architect of Hewlett-Packert Labs. He's also a Hewlett-Packert enterprise fellow and vice president. Great to see you and you're in Vegas. I'm in Palo Alto. We've got a little virtual hybrid going on here. Thanks for spending the time. Oh, thanks John. It's great to be back with you. So much going on. Love to see you guys having this event, kind of in everyone in one spot. Good mojo. Great to see HPE, you know, back in the saddle again. I want to get your take. You're in the action right now on the lab side, which is great. Disruptive innovation is the theme it's always been. This year more than ever coming out of the pandemic, people are looking for the future. They're looking to see the signs. They want to connect the dots. There's been some radical rethinking going on that you've been driving in the labs. You look back at labs, take us through what's going on, what you're thinking, what's the big trends. Yeah, John. So it's been interesting, you know, over the last 18 months, all of us had gone through about a decade's worth of advancement in decentralization, education, health care, our own work, what we're doing right now suddenly spread apart. And it got us thinking, you know, we think about that distributed mesh. And as we try and begin to return to normal and certainly think about all that we've lost, we want to move forward. We don't want to regress. And we start imagining what does that world look like? And we think about the world of 2025, 175 Zetabytes, 150 billion connected things out there. And it's the shape of the world has changed. That's where the data is going to be. And so we start thinking about was it like to thrive in that kind of world? We had a global defense research institute came to us and asked us that exact question, you know, what's the edge? What do we need to prepare for for this age of insight? And it was kind of like when you had those exam questions and I was one of those kids who you get to the final exam and if it's a really good question, suddenly everything clicked. I understood all the material because there was that really forcing question. And when they asked us that, for me it solidified what I'd been thinking about all the work we've done at labs over the last 10 years. And it's really about what does it take to survive and thrive? And for me, it's three things. One is success is going to go to whoever can reason over more information who can gain the deepest insights from that information in a time that matters and then can turn that insight into action at scale. So reason, insight and action. And it suddenly was clear to me everything we've been trying to push for in labs all those boundaries we've been pushing all those conventions we've been defying are really trying to do that for our customers and our partners to bring in more information for them to understand, to be able to allow them to gain insight across departments, across disciplines and then turn that insight into action at scale where scale is no longer one cloud or one company or one country, let alone one data center. Lot there. I love the data. Metadata and meta reasoning insights always been part of that. And you mentioned decentralization. Again, another big trend. I got to ask you, where is the big opportunity? Because a lot of people who are attending discover people watching are trying to ask what should they be thinking about? So what is that next big opportunity? How would you frame that? And what should attendees look for coming out of HPE Discover? So one thing we're seeing is that this is actually a ubiquitous trend. Whether we're talking about transportation or energy or communications, they all are trying to understand and how will they admit more of that data to make those real time decisions? Our expectation is the middle of this decade when we have those 175 zettabytes, 30% of that data will need real time action out at the edge where the speed of light is now material. And also we expect that at that point in time, three out of four of those 175 zettabytes they'll never make it back to the data center. So understanding how we will allow that computation, that understanding to reach out to where the data is and then bringing in that's important. And then if we look at all of those different areas, whether it's energy or transportation, communications, all that real time data, they all want to understand. And so I think that as many people come to us virtually now, hopefully in person in the future, when we have those conversations at labs, it's almost immediate, it takes a while for them and then they realize, oh wait, that's me. This is my industry too, because they see that potential and suddenly where they see data, they see opportunity. And they just want to know, okay, what does it take for me to turn that raw material into insight and then turn that insight into action? You know, storage, compute, never goes away. It gets more and more, you need more of it. This whole data and edge conversation is really interesting, you know, we're living in that data-centric, you know, everyone's going to be a data cover. Okay, that's, we know that, that's obvious. But I got to ask you, as you start to see machine learning, cloud scale, cloud operations, a new edge and new architectures emerging, and clients start to look at things like AI and they want to have more explainability behind it. I hear that all the time, can you explain it to me? Is there any kind of, is it doing good? Is there bias? Is it good, bad? Or, you know, is it really valuable? Experimental, experiential. These are words that I'm hearing more and more of. Not so much a speeds and feeds game, but these are outcomes. So you got the core data, you got a new architecture and you're hearing things like explainable AI, experiential, customer support, new things happening. Explain what this all means. You know, and it's interesting. We have just completed creating an AI ethical framework for all of Fula Packard Enterprise. And whether we're talking about something that's internal, improving a process, something that we sell, our product, or we're talking about a partnership where someone wants to build on top of our services and infrastructure, build an AI system. We really wanted to encompass all of those. And so we, it was, it was challenging. It actually took us about 18 months from that very first meeting for us to craft. What are some principles for us to use to guide our team members, to give them that understanding? And what was interesting is we examined our principles of robustness, of making sure they're human centric, that they're reliable, that they are privacy preserving, that they are robust. We looked at that and then you look at where people want to apply these AI, today's AI, and you start to realize there's a gap. There's actually areas where we have a great challenge, a human challenge, and as interesting as possibly efficacious as today's AI's are, we actually can't employ them with the confidence in the ethical position that we need to really pull that technology in. And what was interesting is that then became something that we were driving at labs. It gave us a viewpoint into where there are gaps, where, as you say, explicability. As fantastic as it is to talk into your mobile phone and have it translated into another one of hundreds of languages. I mean, that is right out of Star Trek. And it's something we can all do. And frankly, we're expecting it now. As efficacious as that is, as we tackle some other problems, it's not enough. We actually need to be explainable. We need to be able to audit these decisions. And so that's really what's informed now our trustworthy AI research and development program at Hewlett-Packard Labs. Let's look at where we want to play AI. We look at what keeps us from doing it, and then let's close that technology gap. And it means some new things. It means new approaches. Sometimes we're going back, back, back to some of the very early AI that things that we sort of left behind when suddenly the computational capability allowed us to enter into machine learning and deep neural nets, great applications, but it's not universally applicable. So that's where we are now. We're beginning to construct that second generation of AI systems, where that explicability, where that trustworthiness, and we're even more important than you said, understanding that data flow and the responsibility we have to those who created that data, especially when it's representing human information, that long-term responsibility, what are the structures we need to support that ethically? Yeah, that's great insight Kirk, that's awesome stuff. And it reminds me of the oldest new again, right? The cycles of innovation. You mentioned AI in the 80s, it reminds me of Desting Off, and I was smiling because, you know, the notion of reasoning and natural language, it's been around for a while, all these other, for a lot of AI frameworks have been around for a while, but applied differently becomes interesting. The notion of meta-reasoning, I remember talking about that in 1998 around ontologies and syntax and data analysis. I mean, again, well-formed, you know, older ways to look at data. And so I got to ask you, you know, you mentioned reasoning over information, getting the insights and having actions at scale. That doesn't sound like an R&D or a labs issue, right? I mean, that should be like in the market today. So I know there's stuff out there. What's different around the Eula Packard labs challenge? Because you guys are working on stuff that's kind of next-gen. So why, what's next-gen about reasoning more over information and getting insights? Because, you know, there's a zillion startups out there that claim to be insights as a service, taking action, outcomes. And I think there we're gonna say a couple of things. One is the technologies and the capabilities that got us this far. They're actually in an interesting position. If we think of that twilight of Moore's law, and it's getting a little darker every day, there's been such a tailwind behind us, tremendous. And we would have been foolish not to take advantage of it while it lasted. But as it now flattens out, we have to be realistic and say, you know what? That ability to expect, anticipate, and then plan for a doubling in performance in the next 18 to 24 months because there's twice as many transistors in that square of silicon. We can't count on that anymore. We have to look now broader. And it's not just one of these technology inflection points. There's so many. We already mentioned AI. It's voraciously devouring all this data. At the same time, now that data is all at the edge is no longer in the data center. I mean, we might find ourselves laughing, chuckling at the term itself, data center. Remember when we centered all the data because that's where the computers were? Oh, that's 2020 thinking, right? That's not even 2025 thinking. Also, security, that cyber threat of nation state and criminal enterprises, all of these things coming together. And it's that confluence of discontinuities. That's what makes a lot of problem. And the second piece is we don't just need to do it the way that we've been doing it because that's not necessarily sustainable. And if something's not sustainable, it's inherently inequitable because we can't afford to let everyone enjoy those benefits. So I think that's all those things. The technology confluence of technology disruptions and this desire to move to really sustainable, really inherently equitable systems, that's what makes it a lab's problem. I really think that's right on the money. One of the things I want to get your thoughts on because I know you have a unique historic view of the trajectory arc. Cloud computing that everyone's attention, lift and shift, cloud scale, gray, cloud native. Now with hybrid and multi-cloud clearly happening, all the cloud players are saying, no, it's never going to happen. All the data says are going to go away. Not really. The data center is just an edge, a big edge. So you brought up the data center concept and you mentioned decentralization. There, it's a distributed computing architecture. There is no line anymore between what's cloud and what's not. The cloud is just the cloud and the data center is now a big, fat edge and edges are smaller and bigger. Their nodes distributed computing now is the context. So this is not a new thing for Yulia Packard Enterprise. I mean, you guys have been doing distributed computing paradigms, supplying software and hardware and solutions since I can remember, since it was founded. What's new now? What do you say to folks who are saying, what is HPE doing for this new architecture? Because now an operating system is the word that they want. They want to have an operating model, dev ops, dev sec ops, all this is happening. What's the state of the art from HPE and how does the lab play into that vision? And it's so wonderful that you mentioned our heritage because if you think about it, was the first thing that Bill and Dave did, they made instruments of unparalleled value and quality for engineers and scientists. And the second thing they did was computerize that instrument control and then they networked them together and then they connected those network measurement sensing systems to business computing, right? And so that's really, that's exactly what we're talking about here, you know? And yesterday was HPE cables, but today it is everything from an Aruba wireless gateway to a Green Lake cloud that comes to you to now our Cray Exascale Supercomputing. And we want to look at that entire gamut and understand exactly what you said. How is today's modern developer who has been just steeped in agile development and dev ops and dev sec ops, how can we make them as comfortable and confident deploying to any one of those systems or all of them in conjunction as confident as they've been deploying to a cloud? And I think that's really part of what we need to understand. And as you move out towards the edge, things become interesting, a tiny amount of resources, the number of threats physical and cyber increased dramatically. It is no longer the healthy, happy environment of that raised floor data center. It is actually out in the world, but we have to because that's where the data is. And so that's another piece of it that we're trying to bring with the labs or distributed systems lab, trying to understand how do we make cloud native access every single byte everywhere from the tiniest little edge embedded system all the way up through that exascale supercomputer. How do we admit all of that data to this entire generation and then the following subsequent generation who will no longer understand what we were so worried about with things being in one place or another. They want to digest all the world's data regardless of where it is. You know, I was just having a conversation and you brought this up. It's interesting around the history and the heritage embedded systems is changing. The whole hardware equations changes. Software driven model now. Supply chain used to be constrained to software. Now you have a software supply chain, I mean hardware. Now you have software supply chain. So everything's happening in these kind of new use cases. And edge is a great example where you want to have compute at the edge not having pulled back to some central location. So again, advantage HPE, right? You got some solutions there. So all these like memory driven computing is something that you've worked on and been driving. The machine product that we talked about when you guys launched it a few years ago. Looks like now a good R&D project because all the discussions I'm hearing whether it's stuff in space or inside hybrid edges is I got to have software running on an embedded system. I need security. I got to have memory driven architectures. I got to have data driven value in real time. This is new as a kind of a new shift but you still need to run it. What's the update on the machine and the memory driven computing? And how's that connected dots for this intelligent edge that's now super important in the hybrid equation? Yeah, and it's fantastic you brought that up. It's gratifying when you've been drawing pictures on your whiteboard for 10 or 15 years and suddenly you see them printed and on the web and you're like, okay, yeah, you guys were there. We're there because we always knew it had to be bigger than us. And for a while you wonder, well, is this the right direction? And then you get that gratification that you see it repeated. And I think one of the other elements that you said that was so important was talking about that supply chain. And especially as we get towards these edge devices and the increasing cyber threat, it's so much more about understanding the provenance of that supply chain and how we get beyond trust to prove. And in our case, that proof is rooted in the silicon. Start with the silicon, establish a silicon root of trust, something that can't be forged, that physically unclonable function in the silicon and then build up that chain, not of trust, but a proof of measurable confidence. And then let's link that through the hardware, through the data. And I think that's another element, understanding how that data is flowing in and we establish that provenance, that provable provenance. And that also enables us to come back to that equitable question. How do we deal with all this data? Well, we want to make sure that everyone wants to buy in and that's why you need to be able to reward them. So being able to trace data into an AI model, trace it back out to its effect on society. All these are things that we're trying to understand at the lab so that we can really establish this data economy and admit the data that we need to the problems that we have that really just are crying out for that solution, bringing in that data, you just know, where's the data? Where's the answer? I get to work with, I've worked for several years with the German Center for Neurodegenerative Disease Research and I was teasing their director, Dr. Nakatra. I said, yo, in a couple of years when you're getting that Nobel Prize for medicine because you cracked Alzheimer's, I want you to tell me, how long was the answer hiding in plain sight because it was segregated across disciplines or across geographies and it was there but we just didn't have that ability to view across the breadth of the information and in a time that matters. And I think so much about what we're trying to do at the lab is that, that's that reasoning more over more information, gaining insights in a time that matters and then it's all about action and that is driving that insight into the world regardless of whether it has to land in an exascale supercomputer or a tiny little edge device, we want today's application of open teams to feel that degree of freedom to range over all of those that infrastructure and all of that data. You know, you bring up a great call out there. I want to just highlight that because I thought that was awesome. The future breakthroughs are hiding in plain sight. It's the access to the people and the talents to solve the problems and the data that's stuck in the silos. You bring those together, you make that seamless and frictionless then magic happens. That's really what we're talking about in this new world, isn't it? Absolutely, yeah. And it's one of those things that, you know, sometimes my kids ask, you know, why do you come in every day? And for me, it is exactly that. I think so many of the challenges we have are actually solvable if the right people knew the right information at the right time and that we all had that, not again, not trust, but that proof, that confidence, that measurable confidence, back to the instruments that HP was always famous for, it was that precision and they all had that calibration tag. So you could measure your confidence in an HP instrument in the same way we want people to measure their confidence when data is flowing through Hewlett Packard enterprise infrastructure. It's interesting you bring up the legacy because instrumentation, network together, connecting to business systems. Hey, that sounds like the cloud. Observability, modern applications, instant action and actionable insights. I mean, that's really the same almost exact formula. Yeah, for me, that's that constant through line from the garage to right now is that ability to handle and connect people to the information that they need. Okay, great to chat, you're always an inspiration and we could go for another hour, talking about Exascale, Greenleg, all the other cool things going on at HPE. I got to ask you the final question, what are you most excited about for HPE and its future and how can folks learn more to discover and what should they focus on? So I think for me, what I love is that I imagine that world where the data is out there at the edge and we have our Aruba team, we have our Greenleg team, we have our core enterprise infrastructure business and now we also have all the way up through Exascale compute. When I think of that thriving business, that ability to bring in massive data analytics, machine learning and AI and then simulation and modeling, that's really what, whether you're a scientist and engineer or an artist, you want to have that intersectionality and I think we actually have this incredible, diverse set of resources to bring to bear to those problems that will span from edge to cloud, back to core and then to Exascale. So that's what really, that's what I find so exciting is all of the great innovators that we get to work with and the markets we get to participate in and then for me, it's also the fact, it's all happening at Hewlett Packard Enterprise, which means we have a purpose. If you ask, and they did ask Dave Packard, Dave, why HP? They said in 1960, we come together as a company because we can do something we could not do by ourselves and we make a contribution to society and I dare anyone to spend more than a couple of minutes with Antonio Neary and he won't remind you and this is whether it is here at Discover or in the halls at labs that remind me, our purpose at Hewlett Packard Enterprise is to advance the way that people live and work and for me that's that direct connection. So it's the technology and then the purpose and that's really what I find so exciting about HPE. That's a great call out and Antonio deserves props. I love talking with him. He's the true Bill and Dave, Bill Hewlett, Dave Packard spirit and I'll say that I've talked with him and one of the things that's as resonant to me and resonates well is the citizenship and it'd be interesting to see if Bill and Dave were alive today that now it's a global citizenship. This is a huge part of the culture and I know it's still alive there at HPE so great call out there and props to Antonio and yourself and the team, congratulations. Thanks for spending the time, appreciate it. Thank you, John, it's great to be with you again. Okay, global labs, global opportunities, radical rethinking, this is what's happening within HP Hewlett Packard Labs, great, great contribution there from Kirk, having him on theCUBE and always fun to talk, so much to digest there. It's awesome, I'm John Furrier with theCUBE, thanks for watching.