 Okay, we're back here live in Silicon Valley. This is Silicon Angle in Wikibon's theCUBE, our flagship program where we go out to the events and extract the signal from the noise. I'm John Furrier, the founder of Silicon Angle. I'm Joe Michaels. Dave Vellante, the co-founder of wikibon.org and this is theCUBE. We'd love to go talk to anyone who has information and data and share that with you, extracting the signal from the noise, as we say. And our next guest here is Thomas Summers, CEO and founder of Rex Computing, 17-year-old whiz kid, fellowship from Peter Thiel's fellowship program where he encourages kids to pursue their passion, take a break or drop out of school. Congratulations. Welcome to theCUBE. Well, thank you. So you're a whiz kid. You know, we were just at the Apple 30th anniversary, WASNIC wasn't there, but we know WAS from Fusion IO. We've interviewed him and we've known him in the valley. You're the next WAS. You're out there building out the next stuff. Tell us a little bit about, one, what the story is around the fellowship, okay, the award, when did it happen and what are you building? Yeah, so I received the Thiel fellowship in May of this year, or not this year, of 2013. And basically how it works is there's a whole application process and finally rounds down to the 20 under 20 and so it's called the 20 under 20 Thiel Fellowship where you receive $100,000 over two years with the only requirement being is that you can't be enrolled in school. So I applied, this is actually my third time applying and I put in working in high performance computing and specifically tried to reduce the cost and the energy needs by computing by many orders of magnitude. And so I received the fellowship and started a company. Sometimes Charm. I guess so. So any VC funding yet? I mean we're in a bubble. I mean any funding yet from besides Peter Thiel? No, so thankfully I actually dropped out of my 11th grade of high school so I didn't have any debt and I've been able to bootstrap. So you dropped out of your junior year or you finished through junior year? So I dropped out one month before the end of my junior year. Okay, so technically you've got two years under your belt of high school. Yes, in terms of... Were you really bored in class? Kind of, yeah. I asked him what he got on his SATs and he said I didn't take him. But Dave, this brings up a big trend. Before we get into, we're talking about the kid, we call them the next waz because it is a lot of hardware geeking going on there. Dave, this is exactly kind of what we talk about the cue, you know. Information, the commoditization of journalism and publishing and information flow is the new normal and you don't need to go to linear school or you can just, you go online, multi-online mooks or multi-online classrooms, information's out there. People can move at such different paces. This is why I'm so passionate about Silicon Academy, the things we're doing that's going to be coming out this next year and props to Peter Thiel because some people are against this, like hey, you know, why disrupt the status quo? But hey, linear path in school is not. So I mean, did you have any reservations or you're like, thank God I'm out of high school? I mean, what was your thought process? It was pretty crazy at first. I found out I received the fellowship. It was kind of a surreal experience for the first couple of days after, week after I found out that I received it, I officially withdrew from high school. And it was kind of just weird, okay, this is just a whole new step into my life. I had already been working prior to receiving the fellowship. I worked at MIT for just over three years at the point and I kind of had some life expenses. You knew, you had a gut feeling, your compass was telling you, and you applied three times. So take us through the first two rejections. Like, were you pumped to apply? Were you not sure? Was the rejections hard on you? And then obviously you kept coming back at a third time's a charm. Take us through that. Yeah, so the first year I applied, I was 14. It was kind of, I've always just been a geek and working in many different areas. And so I wasn't really focused in my application and I was also pretty young. Like, I definitely wouldn't have chosen me at 14. What inspired you, let me ask you that question. What was the inspiration to have the passion to do the kind of geekiness that you're doing, certainly with hardware? You don't randomly see that. It's kind of like a black swan scenario where you get someone at that level. Did you have a relative in the computer business? Do you have your own workshop? How did you get into, what was the motivator, what was the inspiration? So my mom was self-employed. She had a desktop publishing business. She did layout and stuff for different newspapers. And so she had computers in her office and I got to use them from when I was like two until I started first grade when I was about six. And I just really loved them. And when I got about seven, my dad got me my first electronics kit. It was one of those things from the 70s that it has a bunch of wires and springs and a little light bulb and you can do a bunch of different electronics like kits and stuff. And I just thought that it was really amazing that you could take a bunch of components which don't really do anything on their own and you can make something entirely new. And at that point I was just going through a book and doing all the different projects that I could, you know, that were in there but it started making me think of, okay, but what if I apply these different things and start putting things together in a different way? And I just found out. So when you won the award, did your mom freak? Cause you're, I mean, you're still relatively young, moving to California, right? From Massachusetts. Yeah. So how'd that go? Was it supportive or reluctantly supportive or how'd you deal with that? So my mom has always been supportive of me through my life. The first time I applied, she thought I was kind of crazy for applying, but she, you know, you know, do what you want. But when I finally got into the finalist round for the fellowship, it kind of, okay, there's, you know, about a 50-50 chance that I'll actually receive it. And she thankfully believed in me and thought that, you know, if I actually set my mind to doing something, that I could do it. So the product that you've conceived, it's, I mean, big trends, right? Obviously low power, you got multi-cores going on and you're observing these trends. So how did it all come together? And you've got a prop here. What are you actually building? So I originally started this sort of line of research at the lab I was working at at MIT, the Institute for Soldier and Nanotechnologies. And so my mentor advisor there that kind of wrote me into all of this had a long history of working in high-performance computing. And so we got interested in some ARM stuff. So he was utilizing cell phone processors to do high-performance computing and just specifically clustering them to be able to do more computational work with them. And so this was about 2010 and we bought a bunch of TIO map development boards. So it was the first dual core ARM processor, commercially available. And we decided to get a bunch of them and start seeing what we could do. And so it took, going with a bunch of different designs and such, about nine months later, we got a full system and we built a supercomputer in a rubber-made trash can. And that was, at that point, we had actually built a, what you could almost call a supercomputer out of these cell phone processors. We had 48 of them, total 96 cores. And we thought, okay, this is pretty cool. We posted online. It got a bit of attention. And we started to think about, okay, we did this just as a fun project. And then we also wrote a paper on the research end. But then thought, okay, could this be possibly turned into a company or something along those lines? Okay, so it's a challenging area. We were at the original HP Moonshot announcement right outside of Bill and Dave's offices in Palo Alto. That was pretty cool. Calzada, one of the companies that was at that announcement is now defunct. What is, and John's mentioned the walls a couple of times. And it's actually quite interesting, John. I mean, the early floppy disk, right? The system on a board. We look back now, before you were born. Doesn't look like a big deal, but at the time it was really breakthrough. What is it that you're doing that you see as breakthrough? So in the evolution of how this just started as a fun research project to being a company, we were originally just fascinated with arm processors, what you could do with a bunch of really wimpy cores. And that's now involved to what we're doing today, which is utilizing co-processors to kind of get the performance boost against a conventional, say, Intel processor when doing high performance computing tasks. So what I have here is a development board called Parallella by a company called Adaptiva. And so this has a dual core ARM Cortex-A9 processor with an integrated FPGA. But the really fancy part about it is this 16 core co-processor, which can get about 30 gigaflops while using like two watts of electricity. So that's beating an Intel CPU and performance per watt. It's obviously not getting to its actual raw performance. But when you put 10 of these together, you're beating an Intel chip while it costs less and, you know, output starts to outperform it. So obviously the key to massive scale adoption is applications. I mean, you're certainly seeing it in mobile. Do you see, I mean, obviously you see it, how long do you feel like this will take to really essentially take over the data center? It's from a concept standpoint, it makes total sense. Everybody's complaining about, you know, heat density and space and power and cooling. So it makes so much sense. What are the headwinds? Obviously application support and how long do you think it'll take for this to take the data center by storm? Yeah, so the, what we came here to the Open Compute Summit with was a prototype we threw together in about three weeks. And it's a Open Compute Winterfell Torpedo chassis. And we filled it up with these boards and are using it as basically a development platform because we wanted to kind of test the software waters. And we're, there's the Open Compute hackathon happening here right now. And we're giving the different hackathon participants access to both these boards and the systems in our rack. And the idea, I mean, the big problem, both when it comes to ARM and any sort of co-processor is that it is difficult to develop for. Companies don't want to port their software and they don't want to just deal with that mess. And so I'm personally a hardware person but thankfully other members of my team are much more capable on the software side. And we're working on some new ways to both ease development on highly scalable systems and utilizing co-processors. So that's part of the project is actually a development platform on which people can build applications. Yeah, so the system we're showing here has 256 of these epiphany cores. And we have that running as a MPI cluster and we're trying to get other people interested in developing with it. But our actual product plans are to utilize the next generation epiphany chip with each chip has 64 cores, the epiphany four. And having 64 of those chips in a single, one third width 1U torpedo. So you can have 1,024 cores inside that or 3,072 cores in a 1U open compute blade. Thomas, is this a certainty in your mind that this type of architecture eventually will rule the world? Or do you feel like whether it's ARM or Intel has Adam, whatever. Some type of low power processor, obviously ARM has the big lead. Or do you feel like existing designs, X86, let's say, can be brought down to scale? What's your thought on that? So to first address the X86 point, X86 just has a lot of baggage with it. I mean, it was originally CISC and then obviously Intel brought in the whole microcode concept and they've actually have a risk core which a lot of people don't realize that X86 is risk. And they've basically just abstracted all of it. So they have the same X86 instructions and then a decoder which actually has it and it's native. And so you hit some limits with this. I mean, for the past 40 or 50 years almost, we've had Moore's law with us and so we get that basically free performance or energy boost depending on the way you look at it every 18 months or so. But with that starting to run out, we're going to be hitting some physical barriers pretty soon which you can't really contest with using silicon. We're not going to have those free improvements. And so X86, how it is having backwards compatibility all the way to 386 and Intel's focus on just trying to cram as much features and to throw in as many instructions as possible is just a bulky instruction set and processor that just isn't needed. And that's why I think that the most simple thing is the best. Thomas, talk about your plans now. So how far are you in the fellowship? And the question is, what do you do when the money runs out? Is there a follow on? Do they give Moore's or Milestones? So the fellowship money is given out over two years and so we receive $4,400 or so a month. And crazily enough, I've somehow been able to live on that while I was starting a company in the Bay Area, it's kind of... But it's going to dry up at some point so it's a two-year run, right? So it's almost like a really kick-ass internship with cash co-op assignment like I had when I was in college. But what are you going to do next? Is there a program extension? Have you been communicated? Do they communicate like, hey, if you do well or hey, you're on your own, make it or break it? So like another program that Teal Foundation offers or has that is not directly connected to the fellowship in any way, but besides it being run by the same group is Breakout Labs, which a number of Teal Fellows have talked with at least. Gone into, yeah. And so there's things like that, but I still have about a year and a half left. Yeah, you got plenty of time, which I was just curious. People might be curious about what the headroom's going to look like, the ladder, if you will. I have been bootstrapping up to this point, but would at some point like to raise some money just to be able to scale out. And so... How about a team? Have you talked about like a team, like recruiting a team? Have you looked at trying to find more teammates? So we currently have a full-time team of three. And then we have a couple different contributors and people kind of working part-time. And so... But it's a for-profit venture. Yes. It's not like science project. No. The server, let me see. Yeah. What's this called? So that's two parallel boards. And so the unit that we have is having 16 of those, but the really cool part about it, if you forget about the Xilinx dual-core arm is the fact that it has that 16-core core processor, which is currently one of the most power-efficient processors on the market. And what's the plans for it? So this first system that we are going to be shipping this year is going to have 1,024 with the Epiphany III. So 1,024 Epiphany Quars. And then with the Epiphany IV having 4,096 Quars in a single torpedo. Is being an entrepreneur what you thought it would be? Or not? And what's different? Yeah, it is different. Well, what did you expect in the mind of looking at what you might be embarking on? And what was the reality? Share the folks that pre-conditioning view and then what happened inside now that you're an entrepreneur? I thought that there would be a lot more parties. Yeah, I mean the social network has kind of given a bad name to the valley in that respect. Bong hits and funnels and beer bongs. And no, none of that, huh? Just sitting and we're making a circuit board. It's an ugly business. Basically, and I'm honestly happy with that. We call it eating glass. Yeah, no, we call entrepreneurs, they eat glass, they spit out nails, spit out nails. Eat glass, spit out nails, that's what we call it. So you're saying it's not glamorous as people think it is? Not at all. But I would say just as or I think even more fulfilling being here, I was out to the Denver Supercomputing Conference back in November and just being able to go to these events and actually feel like I'm contributing something to society by building something and creating value of some sort is a lot more than I would be doing if I was still in high school or if I would go on to college. So now we turn the interview onto John and Dave's personal focus group test. Since you're the young gun coming up, 17, obviously connected to a computer since you were two, so it's great. Now, but now digital natives are all connected. I have kids as well, 117, 118, same-year age. And what do you think about this discussion around education? I mean, there are many people including myself who think education is completely broken and think there are better ways than say government the way they're handling it. But you're a great example of someone who's in that linear track and now the virtual space, the internet, the connections are resources. So what's your vision and what are your opinion around education? I'm not going to hold it to you. So just it's your opinion. So there's no wrong answer. Yeah. What's your opinion on the future of the best path possible for education? I think one thing that is not developed in school, especially at the early ages where it's really needed is to find one how to learn. Like actually, and that how to learn I think is everyone's born with it. It's just not nurtured and that is curiosity and- The ability to learn, love to learn? And love to learn. And so I see that a lot of schools try to, creativity is good, but they really try to structure it in some way either through drawing or having an art class but not offering other programs and trying to fit everyone into some sort of mold. And I just think that that's the wrong way to go about it. And while that may have been required to a small extent by society 150 to 100 years ago when industrial revolution, there weren't, the society was actually trying to mold people into going into industrial jobs and such. That is I think detrimental to society now. And we're not adapting the education system. What do you think about these Gen 1 MOOCs? M-O-O-C, multi online courses. Massive online online courses. Essentially like video games basically. Are they too linear? Are they not enough data? Is the application tooling not there? Or is it good? What's your opinion? Have you kicked the tires on any of that stuff? I'm sure you use, I'll do a lot of learning online because you're a curious, smart kid. But I mean, you tap the resources available to you. What needs to get done? What are these MOOCs? What is this all about? What's this new online situation like? M-O-O-C-I haven't actually been involved with any of them. I haven't taken like an organized class. But from my knowledge of it, I think that it's a good idea for people who already are motivated and people who want to truly learn. And I think that sadly, many people who have the opportunity and the ability is to do many different things are discouraged at a young age and thus lose the love of knowledge. Let me ask you a different question. Where do you go to learn? I read a lot of papers. I enjoy, sadly, I don't have academic, like actual paper access, any research journal access through a school anymore. But just reading whatever the cutting edge research is from different labs from across the world. And you do that online, right? Yes. And so. Talk about academic papers. Yeah. I think that there's a lot of cutting edge research, which is like you see a lot of these things in the news all the time. But I find those to be the ones that make the biggest claims that typically are far from truth. And then the ones that are really interesting get published and nobody else sees it. Nobody talks about it. So and then we see them in 20 years. But that's a real art to be able to sort of squint through that text and figure out which ones really matter. So what are some of the things that you're tracking to these days? I mean, obviously low power, multi-core servers caught your attention. What else intrigued you? Ever since I saw Iron Man, I've been very, very interested in augmented reality and wearable computing. I see that's now a big thing now. But I think that that isn't going to be successful in this first generation with Google Glass and some of the other things that are going to be on the market. I think that there's going to be a lot of potential in some of the new photonics and laser research, which is taking place right now, but industry hasn't quite gotten directly involved with it. And so I think that academia is actually ahead right now with it, but it'll be getting into actual production in a couple years. Okay, Thomas, thanks so much for coming on theCUBE. Very impressive individual and gutsy to drop out of high school. Looks like you're well along the way, but you might want to just go mail in University of Phoenix, might have a high school GED program. It's like an hour of your time, knock it out. I don't even know if that's even an option, I just made that up, but sounded good. The future is what you're doing, and I think that's what I find truly interesting. You're a founder of a company, and that's awesome. So thanks for coming on theCUBE. It's theCUBE, we'll be right back. 17 year old genius here in Southern theCUBE, Peter Thielfellow, $100,000 grant to drop out of school, start his own company, and we'll see how that goes. We'll be keeping track of it. We'll be right back after this short break.