 We're back at the Red Hat Summit 2022. The Cube's continuous coverage is day one. We're here all day tomorrow as well. My name is Dave Vellante, I'm here with Paul Gillan. Matt Hicks is here, he's executive vice president of products and technologies at Red Hat. Matt, good to see you. Thanks for coming on. Nice to see you face to face. Thanks, thanks, Dave. Thanks, Paul. It's good to be here. So you took a different tack with your keynote today. Had an homage to Ada Lovelace and Srinivastur Ramanujan, which was kind of cool. Your point was they weren't noted at their time and nobody was there to build on their early ideas. I mean Ada Lovelace, I think it was a century before. Ramanujan was a decade plus, but you tied that to open source. Give us your kind of bumper sticker of your premise there. Yeah, I think I have a unique seat in this from Red Hat where we see new engineers that come in that sort of compete on a world stage in open source and the best, which is easy to track just in contributions, are not necessarily from the background you would expect them from. And for me it's always really inspiring. Like you have this potential in people and open source is a great model for getting that out. We told the history story because I think when you look over history just some of that potential that's been ignored before. Sure it's happening right now, but getting that tied into open source models, we think can hopefully let us tap into a little more than we have in the past. We're thinking about innovation and specific to open source. Is it a case where, I wonder, I don't really know the history here of open source, maybe you can educate me. Is it the case where open source observes a de facto standard let's say or some other proprietary approach and says hey we can build that in open and that's the inspiration or is it an innovation flywheel that just invents? I think it's both at this stage. So in the early days if you take something like Linux it was a little more of, there was the famous memo of like this was gonna be a hobbyist project we're just gonna light up x86 hardware and have an operating system we can work with. That was a little more of like this standards were there but it was can we just build a better operating system with it? Better than Unix because you live up to the promise of Unix. That's right. In Unix you had some standardization to models but it wasn't open in that same sense. Linux has gone well beyond a hobbyist project at this point but that was maybe that clone model to Unix. These days though if you take something like Kubernetes or take something like Ansible that's just more pure innovation. You didn't necessarily have a Kubernetes model that you're building a better version of. It was distributed computing and how can we really make that tick and bring a lot of great minds into that to build it. So I think you see both of them which is one of the things that makes open source fun like it has a broad reach at this point. There's one major area of software that open source has not penetrated yet and that is applications. There have been sugar CRM there have been open ERP applications and such but none of them really taken off and in fact tended to be drawn back to be proprietary. Why do you suppose open source has been limited to infrastructure and hasn't branched out further? Yeah I think part of it is where can you find a model where lots of different companies are comfortable contributing into. If you have one solution and one domain from one company you're going to struggle more getting a real vibrant community built around that. When you pick an area like infrastructure or core platforms you have a lot of hardware providers the use cases span from traditional apps to AI you have a lot of places to run that and some massive companies. So it's a volume really. It really is you just have an interest that spans beyond companies and that's where we've seen open source projects really pick up and build critical mass. How about crypto? Dows. I mean that's right. Isn't that the form of open source? I mean is that the application? Really exactly what you're talking about it. True or? It would if you look at cryptography, encryption algorithms even go to quantum going forward. I think a lot of quantum access will be driven in an open source model. The machines themselves will be machines but things like Kizkit that is how much people will access that. So it is a powerful model for getting into areas that are pretty bleeding edge on it as well. We were talking before Andy mentioned that hardware and software are increasingly intersecting. That was the theme we heard at the keynote this morning. Why do you believe that's happening and how do you see that, how does that affect what you do? I think the reason that's happening is there is a push to make decisions closer and closer to users on it because on one side like law of physics and then on the other of it's just a better experience for it and so whether that is in transportation or it's in telecommunications you see this push outside the data centers to be able to get at that data locally for it but if that's the draw I think also we're seeing hardware architectures are changing. There are standards like ARM that are lower power that lets you run pretty powerful compute at the edge as well and I think it's that combination saying we can do a lot at the edge now and that actually benefits us building user experiences in a lot of different domains is making this pull to the edge really quickly but it's an exciting time to be seeing that happening. And pretty powerful is almost an understatement when you think about what the innovations that are going on, right? I mean in particular at the edge I mean you're seeing Moore's law be blown everybody says Moore's law is dead but you're seeing the performance of when you combine the GPU and the CPU and the NPU and the exciting it blows away anything we've historically known. So you think about the innovations in software that occurred as a result of Moore's law what are the new beach heads that we could potentially see in open source? I think when you start taking the AI patterns on this and AI is a broad space but if you go even to like machine learning of optimization type use cases you start leveraging how you're going to train those models which gets you into CPUs and GPUs and TPUs in that world and then you also have the how am I going to take that train model put it on a really lightweight device and efficiently ask that model questions and that gets you into a different architecture design but that combination I think we're going to see these domains build differently where you have mass compute training type capabilities and then push that as close to the user as you can to make decisions that are more dynamic than traditional codes. So a lot of the AI that's done today is modeling that's done in the cloud and what you're talking about at the edge and you think about vehicles is real time inferencing and that's massive amounts of data it's a different architecture and requires different hardware presumably and different software so when you guys Linux is obviously there. Yeah that is where we get excited about things like the GM announcement you're in the square in that aspect of running compute right at the end user and actually dealing with sensor and data that's changing there to help in this case like drivers assistance capabilities with it but I think that the innovation we'll see in that space will be limitless on it so it's a nice combination of the two and you'll still have traditional applications that are going to use those models but I think of it almost as it's like the new middleware and we have our traditional middleware techniques that we know and patterns they will actually be augmented with things like machine learning models and those capabilities to just be more dynamic so it's a fun time right now seeing that. A lot of data too and again I wonder how much of that is even going to be persisted probably enough because there's going to be so much of it how much will come back to the cloud a lot maybe not most of it but it's still massive amounts relative to what we've seen before. It is and this is you've heard our announcement around OpenShift Streams and those capabilities and Red Hat what we do we will always focus on hybrid with it because a lot of that data it'll be dropped at the edge because you won't need it but the data you act on and the data you need you will probably need at your end device and in your cloud and maybe even on premise and capabilities like Kafka and the ability to pick and stream and stay consistent. We think there's a set of really exciting services to be able to enable that class of development and we're hopefully we'll be at the center of that. You announced today an agreement with GM to build on their Altify platform. Auto industry very proprietary historically with their technology do you think that this is an opportunity to crank that open? Absolutely I think in I've been involved with open source for a while but I think all of them started in a very proprietary model and then you get to a tipping point where open source models can just unlock more innovation than proprietary models and you see them tip and flip and I think in the automotive industry and actually in a lot of other industries the capabilities of being able to combine hardware and software fast with the latest capabilities it'll drive more innovation than just sticking to proprietary models so yeah I believe it will be one of many things to come there. You've been involved in open source for a while like how long of a while people must joke about when they look at you Matt they must say what did you start when you were five? Yeah you get that a lot. I do, it's my children I think age me but yeah for me it was the mid 90s that's when I started with open source so it's been a long run. You made the statement in your keynote that software development is messy I presume part of your job is to make it less messy but now we talk about all these new beach edge this new innovations. A lot of it's unknown and it could be really messy so who are those, is there a new breed of developer that's emerging? Are they going to come over from the cloud developers or is it the IOT crowd and the OT crowd that's going to be the new developers? I wish I knew but I would say I think you, I do think you'll get to almost like a laws of physics type challenge where you won't learn everything you're not gonna know the depths of 5G implementation and Kubernetes and Linux on that and so for us this is where ecosystem providers are really really critical where you have to know your intersection points but you also have to partner really well to actually drive innovation in some of these spaces the domains themselves are massive on it so our area is we're going to know hybrid we're going to know open source based platforms to enable hybrid and then we're going to partner with companies that know their domains and industries really well to bring solutions to customers. I'm curious about partnering because Paul Cormier mentioned that as well as being critical, do you have sort of a template for partnering or is each partnership unique? I think at this point the market's changing so fast that we do have templates of who are you going to embed solutions with? Who are you going to co-sell with and co-create? The challenge in technology though is it shifts so quickly. If you go back five years maybe in 10 years public cloud probably wasn't as dominant as it is now. Now we're starting to see the uptick of edge solutions probably being having as much draw as public cloud and so I think for us the partnership follows the innovation on those curves and finding the right model where that works for customers is the key thing for us but I wish there was more of a pattern we could say it stays stable for decades but I think it changes with the market on as we do that. But you know it's funny because you see every 15 years or so the industry gets disrupted I mean we certainly saw it with mainframes and PCs and then the internet and then the cloud. You guys kind of been there, well Linux throughout. I mean okay it built the internet, built the cloud, it's building the edge so it's almost, I don't want to say you're disruption proof because that's just going to jinx you but you've architected the products in a way that they're compatible with these new eras of industries. Everything needs an operating system. Everything needs an operating system but you've seen operating systems come and go and Linux has survived so many different waves, why, how? You know I think for us when you see open source projects they definitely get to a critical mass where you have so much contribution, so much innovation there that they're going to be able to follow the trends pretty well. If you look at a Linux, whatever the next hardware innovation that comes out is Linux has enough gravity that it's open, it's accessible, you're going to design to it the capability will be there. I think you're seeing similar things in Kubernetes now where if you're going to try to drive application innovation it is a model that gives you ton of reach. You have thousands of contributors. That's been our model though is find those projects, be influential in them, be able to drive value in life cycles, but I think it's that open source model that gives us the durability where it can keep changing and tracking to new patterns, so, so. Yeah, there's been a lot of open source that wasn't able to sustain so I think you guys obviously have a magic formula. That's true, there is some art to picking. I think millions of projects but you've got to watch for that critical mass. Yeah, open source is also a place where failed products go to die. So you have to be sure you're not in that corner. Look at Kubernetes. I mean the fact that that actually happened is astounding to me when you think about it. I mean even Red Hat was ready to go on a different path. What if that had happened? Who knows, maybe it never would have, maybe to your point about Able Lovelace, maybe it would have taken a decade or a revolution. You know I think in some of these you have to, you have to watch really closely. We obviously have a lot of signals of what will make good long term health and I don't think everyone looks at those the same. We look at them from trademark controls and how foundations are structured and who the contributors are and the spread of that. And it's not perfect but I think for us you have to have those, that longevity built in there or you will have a spike of popularity that has the tendency to just fall apart on it. So we've been doing that pretty well. We've been growing conditions for a long life as something that's a, maybe it's an art form, I don't know if it's a data form, it's a culture maybe, maybe it's culture. Yeah, probably a combination. Some days I think I'm like this could part art, part science, but it's certainly a fun space to be in and see that happen, it's inspiring to me. Matt Hicks, great to have you back on theCUBE and good job in the keynote, really interesting angle that you took. So congratulations. Thanks for having me. You're very welcome. All right, keep it right there. Dave Vellante for Paul Gillin, Red Hat Summit 2022 from Boston. You're watching theCUBE.