WEBVTT - Can Open-Source Semiconductors Upend the Chip Industry? 0:00:10.760 --> 0:00:14.240 Hello, and welcome to another episode of the Odd Thoughts Podcast. 0:00:14.360 --> 0:00:18.600 I'm Chacy Alloway and I'm Joe Wisental. So Joe, uh, 0:00:18.960 --> 0:00:23.600 we just can't get away from semiconductors. You know. I 0:00:23.640 --> 0:00:26.319 thought we were going to end it with our five 0:00:26.360 --> 0:00:30.480 episodes series, but the people are clamoring for more chips. Tracy. 0:00:30.560 --> 0:00:32.800 That's just a fact I can't get can't get away 0:00:32.800 --> 0:00:34.800 from it. Yeah, It's one of those like just when 0:00:34.800 --> 0:00:36.680 you thought you were out, they pull you back in 0:00:37.000 --> 0:00:40.920 kind of things. Okay, So, by popular request, this is 0:00:40.960 --> 0:00:44.640 a bonus episode to our semiconductor series. It's going to 0:00:44.680 --> 0:00:46.960 be a little bit different to the other ones because 0:00:46.960 --> 0:00:51.280 we're going to be talking about a new technology, a 0:00:51.400 --> 0:00:53.640 new sort of how do I explain this? Well, I 0:00:53.640 --> 0:00:57.960 guess it's an open source hardware technology that could have 0:00:58.040 --> 0:01:01.240 a big impact on the chips industry and the way 0:01:01.280 --> 0:01:04.280 things have traditionally been done. Right. So a lot of 0:01:04.280 --> 0:01:07.560 our discussion has sort of talked about a few like 0:01:07.720 --> 0:01:11.360 basic business models. I mean, we've talked about Intel and 0:01:11.400 --> 0:01:15.279 the sort of integrated UM chip design and fab model 0:01:15.480 --> 0:01:19.680 boundary versus fab and then the fabless semiconductor companies which 0:01:19.720 --> 0:01:24.040 have given rise to the contract manufacturer Tawan Semi, which 0:01:24.040 --> 0:01:26.560 is sort of the ultimate fab. So we've sort of like, 0:01:26.760 --> 0:01:29.520 um got this sort of like broad contours of the industry, 0:01:29.600 --> 0:01:32.600 but in terms of like interesting new technologies or different 0:01:32.640 --> 0:01:35.600 directions or totally different models of where how things could go, 0:01:36.160 --> 0:01:41.160 still more to be discovered. Yeah, that's exactly right. So, uh, 0:01:41.400 --> 0:01:43.840 someone told us that we had to talk to this 0:01:43.959 --> 0:01:47.000 particular company because they were one of the leaders in 0:01:47.160 --> 0:01:52.840 trying to commercialize this open source hardware idea. It's called 0:01:53.120 --> 0:01:56.680 Risk five. It's a let me see if I can 0:01:56.720 --> 0:01:59.760 pull it up, it's R I S C and then 0:02:00.200 --> 0:02:04.440 five Risk five. So what we're gonna do is actually 0:02:04.520 --> 0:02:07.760 learn about what the technology is and then dive in 0:02:07.760 --> 0:02:10.440 a little bit to what it might mean for existing 0:02:10.480 --> 0:02:14.880 chip makers and how the whole ecosystem of semiconductors actually works. 0:02:16.240 --> 0:02:18.200 I can't wait. I don't know anything about any of 0:02:18.240 --> 0:02:21.560 this stuff, so I just want to learn more. Yeah, okay, 0:02:21.600 --> 0:02:24.839 we're all learning together. So without further ado, let's bring 0:02:24.840 --> 0:02:28.440 on Chris Latner. He's the president of product and engineering 0:02:28.720 --> 0:02:33.200 over at SCI five. He's also a long time tech 0:02:33.240 --> 0:02:35.960 guy and quite famous in the industry for reasons that 0:02:36.040 --> 0:02:40.880 we will get in. So Chris, welcome to our thoughts. Hey, 0:02:40.960 --> 0:02:42.519 thank you for having me. I'm excited to be here. 0:02:43.120 --> 0:02:45.919 Could you could you maybe give us a snapshot the 0:02:46.280 --> 0:02:52.040 elevator pitch for risk five before we begin. Sure sounds great. 0:02:52.200 --> 0:02:54.680 Let's start with what what is risk five? So when 0:02:54.720 --> 0:02:57.680 you start talking about these companies, companies like Intel or 0:02:57.919 --> 0:03:01.799 Arm or m D things like this, they generally make processors, 0:03:02.120 --> 0:03:04.799 and so the processor in your computer it's jobs to 0:03:04.840 --> 0:03:08.120 run software. And there's this complicated dance that happens between 0:03:08.120 --> 0:03:11.160 the software and the hardware and the thing that mediates that. 0:03:11.280 --> 0:03:13.480 The thing that lives in between these two worlds is 0:03:13.520 --> 0:03:17.000 a thing called an instruction set. That instruction set is 0:03:17.040 --> 0:03:21.240 often proprietor and so um. With Intel, for example, they'll 0:03:21.280 --> 0:03:24.520 use an instruction set called x A D six. ARM 0:03:24.639 --> 0:03:27.320 has a set of instruction sets that they named after themselves. 0:03:27.400 --> 0:03:29.720 They have the ARM Instruction Set UM. There are many, 0:03:29.880 --> 0:03:31.760 many of these things that have existed over the years. 0:03:32.040 --> 0:03:36.600 A company called MIPS remember, a company may remember ages 0:03:36.640 --> 0:03:39.600 ago called Sun had an instruction set called Spark, and 0:03:39.640 --> 0:03:42.840 many of these things existed in their designed for different 0:03:42.920 --> 0:03:46.360 different configurations. Now these have all worked and these have 0:03:46.400 --> 0:03:48.000 been a good thing, and obviously we have a lot 0:03:48.000 --> 0:03:51.960 of successful computers today. Um, but there's a problem. The 0:03:52.040 --> 0:03:54.840 problem is that the software and the hardware lasts longer 0:03:54.880 --> 0:03:58.400 than the companies do sometimes and so MIPS for examples, 0:03:58.440 --> 0:04:02.440 in bankruptcy these days, Sun kind of went away and 0:04:02.480 --> 0:04:06.320 Spark is in a different world. And so one of 0:04:06.360 --> 0:04:08.720 the challenges that you end up having as a device 0:04:08.800 --> 0:04:12.560 manufacturer as a software provider is that you need something 0:04:12.600 --> 0:04:16.760 that will survive. And you know, companies in general go 0:04:16.839 --> 0:04:20.000 through changes and um, you know, as an industry, it's 0:04:20.040 --> 0:04:23.040 always evolving. It's a fast paced world. Now there's another 0:04:23.120 --> 0:04:26.320 challenge with this, which is also when you have a 0:04:26.320 --> 0:04:30.040 company that's tied to a specific construction set that they create, 0:04:31.160 --> 0:04:33.640 you only have a single provider. This can sometimes be 0:04:33.720 --> 0:04:37.640 limiting if you're building a serious advices around one company's technology, 0:04:37.960 --> 0:04:40.520 both if they go away, but also you don't have 0:04:40.560 --> 0:04:44.520 as much market pressure pushing them to innovate and pushing 0:04:44.560 --> 0:04:48.120 them to you know, have reasonable charges and rates and 0:04:48.160 --> 0:04:50.960 things like this. So RISK five is a different approach 0:04:50.960 --> 0:04:54.520 to this. Instead of saying it's a proprietary interface between 0:04:54.600 --> 0:04:57.800 hardware and software, Risk five is an open standard and 0:04:57.880 --> 0:04:59.400 so it's you can think of it as an open 0:04:59.440 --> 0:05:03.760 source kind of instruction set which allows many different implementations 0:05:03.800 --> 0:05:07.080 of hardware to work with all the software. And so 0:05:07.240 --> 0:05:08.960 in this in this in this vein you can see 0:05:09.000 --> 0:05:11.760 to something like t c P i P, which is 0:05:11.760 --> 0:05:14.359 an open standard for networking, or you could see this LINUX, 0:05:14.360 --> 0:05:18.279 which is an open standard for for software UM in 0:05:18.320 --> 0:05:21.240 the unex world and things like this. And so that's 0:05:21.240 --> 0:05:23.920 really what RISK five is is it's an old idea 0:05:24.000 --> 0:05:26.920 about this interface between hardware and software. But the new 0:05:26.960 --> 0:05:30.120 thing about it is that it's uh an open standard 0:05:30.240 --> 0:05:32.400 and it also has a fresh new design that really 0:05:32.480 --> 0:05:34.800 learns from what has happened in the industry and it's 0:05:34.839 --> 0:05:39.360 just technically better in most ways. So just to be clear, 0:05:39.800 --> 0:05:45.320 your company SCI five, is it manufacturing chips or is 0:05:45.360 --> 0:05:50.839 it just creating this software layer from which others could 0:05:50.960 --> 0:05:56.279 theoretically then build ships. Great question. So, if if I 0:05:56.320 --> 0:05:59.840 look backwards, Risk five started started at the University of Berkeley, 0:06:00.120 --> 0:06:02.719 and so it's about ten years old. It started in 0:06:02.760 --> 0:06:05.960 a research lab, and you know, initially they were just 0:06:06.040 --> 0:06:09.719 building an instruction set to support academic research, support some 0:06:09.760 --> 0:06:12.919 of the projects that they're interested in. And that that 0:06:13.440 --> 0:06:17.240 became a movement where a lot of people agreed that 0:06:17.279 --> 0:06:20.719 they needed a standardized system and that allowed people to 0:06:20.760 --> 0:06:24.040 do hardware research without having to do all the software work. 0:06:24.320 --> 0:06:27.279 Really was it was enabling new kinds of innovation coming 0:06:27.279 --> 0:06:30.760 out of the academic world. The founders of Risk five 0:06:31.520 --> 0:06:34.520 all came from that research lab in Berkeley and five 0:06:34.600 --> 0:06:36.520 years later decided to start a new company, and that 0:06:36.560 --> 0:06:40.040 company is called SCI five. So that's where that's where 0:06:40.040 --> 0:06:43.240 we We now are leading the Risk five revolution with 0:06:43.360 --> 0:06:46.320 the founders of Risk five, and we're building a new company, 0:06:46.360 --> 0:06:49.479 but we're also building and defining and driving much of 0:06:49.520 --> 0:06:54.400 the technology the powers this open cross ecosystem standard that 0:06:54.600 --> 0:06:57.440 Risk five is, and so SCI five as a company, 0:06:57.520 --> 0:07:01.240 we build Risk five implementations. And so you can think 0:07:01.279 --> 0:07:03.720 of as an analogy um T c P i P 0:07:03.880 --> 0:07:08.000 is an open standard for networking, but Cisco builds routers, right, 0:07:08.400 --> 0:07:11.760 and so we yeah, we're we're building implementations of Risk 0:07:11.840 --> 0:07:15.320 five and so uh we are the best place to 0:07:15.400 --> 0:07:18.200 go to if you want to build a product around 0:07:18.320 --> 0:07:20.680 Risk five technology. And we have a wide range of 0:07:20.680 --> 0:07:25.800 different implementations with different performance points, size area, power trade offs, 0:07:25.800 --> 0:07:28.680 things like this could you maybe give us a little 0:07:28.720 --> 0:07:31.400 bit of your background, So I teased it in the intro, 0:07:31.600 --> 0:07:34.360 But you are very well known within the tech and 0:07:34.880 --> 0:07:38.400 computer world, most notably well you might debate this, but 0:07:38.840 --> 0:07:42.000 for developing the Swift programming language, which I think is 0:07:42.000 --> 0:07:45.760 the thing underpinning iOS on Mac and Apple and things 0:07:45.800 --> 0:07:48.640 like that. How did you get involved in this space 0:07:48.680 --> 0:07:51.760 and what attracted you to sci five and the risk 0:07:51.840 --> 0:07:55.680 five idea. Yeah, so, so I've been kicking around the 0:07:55.680 --> 0:07:58.080 tech industry for a number of years now. Amazing how 0:07:58.160 --> 0:08:02.080 time flies. I spent eleven plus years of Apple, for example, 0:08:02.560 --> 0:08:05.400 but did a lot of work on fundamental low level 0:08:05.640 --> 0:08:10.200 software called compilers. And compilers are the thing that takes 0:08:10.240 --> 0:08:13.480 the application that a programmer rights and gets it to 0:08:13.480 --> 0:08:16.360 work with the hardware and so that instruction set that 0:08:16.360 --> 0:08:20.480 that that interface between hardware and software. Compilers are just 0:08:20.640 --> 0:08:23.200 on the software side of that and they handle that 0:08:23.720 --> 0:08:26.360 that making it work on the processor for you and 0:08:26.440 --> 0:08:28.920 so um. As part of that, I build a number 0:08:28.960 --> 0:08:32.559 of technologies that are all open source, including this thing 0:08:32.559 --> 0:08:34.960 called l VM and playing and a bunch of other 0:08:35.520 --> 0:08:40.199 compiler technologies that underpinn and enable a lot of that technology. 0:08:40.280 --> 0:08:43.000 Right at the hardware software boundary. There are many things 0:08:43.000 --> 0:08:44.920 that happened in Apple. We build a lot of really 0:08:44.920 --> 0:08:47.520 great things. The Swift programming language is one of them, 0:08:47.600 --> 0:08:49.240 and so I've done a lot of work and that 0:08:49.360 --> 0:08:52.040 that space. Since then, I've worked a number of other places, 0:08:52.040 --> 0:08:55.560 including Tesla and Google, workd on machine learning, infrastructure, large 0:08:55.559 --> 0:08:58.400 scale TPUs, whole bunch of things, and it's all been 0:08:58.480 --> 0:09:04.960 mostly this systems software working with the hardware. So when 0:09:04.960 --> 0:09:08.080 I was considering enjoining a SCI five, the really exciting 0:09:08.120 --> 0:09:09.960 thing to me about SCI five is is a number 0:09:09.960 --> 0:09:13.480 of things. Is this focus on open technologies, which you 0:09:13.600 --> 0:09:15.480 kind of have an unfair advantage when it comes to 0:09:15.559 --> 0:09:18.640 changing the world right because people love open technologies and 0:09:18.679 --> 0:09:21.760 a lot love the reach that you can have with them. Um, 0:09:22.120 --> 0:09:25.280 it's just the the idea of instead of being on 0:09:25.320 --> 0:09:27.959 the software side of the hardware software divide, be able 0:09:28.040 --> 0:09:30.800 to straddle that. And so SCI five is both a 0:09:30.840 --> 0:09:33.520 software company and a hardware company, and so we're able 0:09:33.559 --> 0:09:36.480 to innovate by spanning that that gap and really doing 0:09:36.520 --> 0:09:40.400 things that are quite next generation because we bring this 0:09:40.440 --> 0:09:43.680 combination of knowing how the transistors work and your silicon, 0:09:43.920 --> 0:09:46.679 but also how how to get programmers to use it. 0:09:46.920 --> 0:09:51.800 And that's something that UM typically UH companies have really 0:09:52.160 --> 0:09:55.680 bifurcated on. You have software companies, you have hardware companies. 0:09:55.760 --> 0:09:58.360 It's really rare to be able to span that gap. 0:10:13.320 --> 0:10:17.040 So give us an example of a theoretical customer, or 0:10:17.240 --> 0:10:20.880 if you want an actual customer who comes to Risk five, 0:10:21.040 --> 0:10:23.280 what are they looking for? What can they Why do 0:10:23.400 --> 0:10:25.680 they go to you as opposed to someone else. What 0:10:25.800 --> 0:10:30.240 can they get from you in terms of performance, in 0:10:30.360 --> 0:10:33.320 terms of a product that makes sense for them that 0:10:33.559 --> 0:10:37.880 from a business standpoint, UM can't get from somewhere else. Absolutely, 0:10:38.000 --> 0:10:41.800 So there's many different categories. This one example is a 0:10:41.880 --> 0:10:45.959 replacement customer. So somebody is using a product from, for example, 0:10:46.040 --> 0:10:48.839 Myths and they're saying, hey, wow, we're building our technology 0:10:49.440 --> 0:10:53.080 on a technology staff built by a company in bankruptcy. 0:10:53.120 --> 0:10:56.480 Maybe that's not a good idea, right. We also have 0:10:56.559 --> 0:10:59.120 many customers coming to us from arm designs, and so 0:10:59.200 --> 0:11:01.880 the arm arm company has a number of CPU designs 0:11:01.880 --> 0:11:05.079 in their portfolio, and for a variety of reasons, customers 0:11:05.080 --> 0:11:07.080 are not as happy with that as they used to be, 0:11:07.600 --> 0:11:11.280 and so a replacement of existing products typically they're looking 0:11:11.320 --> 0:11:14.600 for something that is um you know, solves the same 0:11:14.720 --> 0:11:17.200 kind of a job as what they were using before, 0:11:17.280 --> 0:11:20.320 but is better in terms of UM the area, the 0:11:20.400 --> 0:11:24.000 size of the hardware, or the power consumes UM and 0:11:24.160 --> 0:11:27.640 things like this, and so Risk five as a technology 0:11:27.760 --> 0:11:29.679 is better than many of these things out there. Is 0:11:29.720 --> 0:11:33.640 better designed and engineered, mostly because it's much newer. And 0:11:33.720 --> 0:11:37.720 so the technology industry has invented a lot of things, 0:11:37.800 --> 0:11:40.959 and as with any any space, there are mistakes that 0:11:41.000 --> 0:11:43.200 are made along the way, and Risk five is a 0:11:43.200 --> 0:11:45.240 fresh start that allows us to get those things right 0:11:45.280 --> 0:11:47.240 and be able to draw the best ideas from across 0:11:47.240 --> 0:11:50.280 the industry into one one system. I think that the 0:11:50.320 --> 0:11:53.840 other more interesting and exciting part of what sci Fi's 0:11:53.880 --> 0:11:58.319 business is propelled by is differentiated solutions. And so while 0:11:58.440 --> 0:12:00.120 many people come to us saying like, hey, I want 0:12:00.120 --> 0:12:02.880 to swap this thing out, that the most fun pieces 0:12:02.920 --> 0:12:07.120 are that we're able to enable products that nobody else 0:12:07.160 --> 0:12:11.239 can build. And that happens because we build our processors 0:12:11.240 --> 0:12:15.000 and our other technologies as kind of as software. And 0:12:15.040 --> 0:12:18.200 so we build hardware with the spirit of software that's 0:12:18.280 --> 0:12:21.040 very configurable, dynamic, and you can tune it for the 0:12:21.120 --> 0:12:23.600 use case. And this allows you to say, hey, well, 0:12:23.640 --> 0:12:26.120 I want to build, you know, an AI powered thing 0:12:26.280 --> 0:12:29.480 that does voice recognition and this stuff, and so we 0:12:29.520 --> 0:12:31.679 can give you the world's best implementation of that that's 0:12:31.760 --> 0:12:34.600 really tuned to the use case. And here we are. 0:12:35.040 --> 0:12:39.240 One of the macro phenomena that's going on today is 0:12:39.280 --> 0:12:41.360 the thing that many people would call the end of 0:12:41.400 --> 0:12:45.160 Moore's law. And so the process technology, the fabs are 0:12:45.200 --> 0:12:47.720 not not not moving at the speed that they used to, 0:12:47.840 --> 0:12:51.160 and the economics are really shifting. And what this means 0:12:51.240 --> 0:12:53.800 is this means that there's an increased desire and increased 0:12:53.800 --> 0:12:57.640 need for custom solutions. And custom solutions can be much 0:12:57.679 --> 0:12:59.840 more power efficient, they can be more cost efficient, they 0:12:59.840 --> 0:13:04.080 can be just better for the you know, being part 0:13:04.080 --> 0:13:07.360 of your life as a human and they fit in 0:13:07.400 --> 0:13:10.280 as IoT light bulbs are like all these other things. 0:13:10.720 --> 0:13:13.200 And so that's really driving customization way that we haven't 0:13:13.240 --> 0:13:16.160 seen before. And this is really what SCI five strength is. 0:13:17.400 --> 0:13:20.240 So I want to go back to that replacement idea 0:13:20.559 --> 0:13:24.920 because a big theme running through this entire semiconductors series 0:13:25.000 --> 0:13:28.719 is the politicization of chips. And of course we've seen 0:13:28.720 --> 0:13:32.400 the US put restrictions on China and things like that. 0:13:33.360 --> 0:13:36.160 I think one of your customers was reported to be 0:13:36.679 --> 0:13:40.320 Ali Baba, So I mean, is this a way for 0:13:40.559 --> 0:13:47.760 companies too, I guess accelerate a reduction in their dependency 0:13:47.920 --> 0:13:54.600 on things like TSMC chips or maybe Western chips from Intel. Also, 0:13:54.640 --> 0:13:58.360 I think that's a complicated topic and I'm certainly not 0:13:58.400 --> 0:14:00.360 an expert on that. I think there's many factors that 0:14:00.360 --> 0:14:03.920 are going on here. Risk five is an open standard, 0:14:04.160 --> 0:14:09.560 allows collaboration, and that collaboration happens across political boundaries, and 0:14:09.640 --> 0:14:12.720 so the US and sci FI, for example, is investing 0:14:12.720 --> 0:14:16.240 a lot in Risk five. China is also investing a 0:14:16.280 --> 0:14:18.920 tremendous amount into Risk five software in particular, and they're 0:14:18.920 --> 0:14:21.960 starting to look at Risk five hardware and so um. 0:14:21.960 --> 0:14:24.760 From a technology perspective, this just kind of flats all 0:14:24.800 --> 0:14:27.280 boats like this just makes things go faster, the technology 0:14:27.280 --> 0:14:32.280 curve go that way. In terms of the plutotas politicization 0:14:32.480 --> 0:14:34.920 of fabs and things like this, I think it really 0:14:34.920 --> 0:14:37.320 comes down to the business model of the individual companies. 0:14:37.320 --> 0:14:40.000 And from sci Fi's perspective, we're very happy to work 0:14:40.040 --> 0:14:44.840 with UH, Samsung and TSMC and multiple different fabs and 0:14:45.360 --> 0:14:47.280 UH for that. For US it's really about what our 0:14:47.320 --> 0:14:49.720 customers are looking for and what the products they're trying 0:14:49.760 --> 0:14:51.440 to build are and trying to find the best possible 0:14:51.480 --> 0:14:56.040 answer for their their outcome. So the big FABS that 0:14:56.120 --> 0:15:00.360 everybody knows when it comes to the annual manufacturer, your 0:15:01.120 --> 0:15:04.960 their client. UM. Generally, the way it works is that 0:15:05.120 --> 0:15:10.200 we provide technology to other hardware companies and so UM 0:15:10.320 --> 0:15:12.920 let me give you an example of this. UM. We 0:15:12.920 --> 0:15:17.600 we recently launched. UH there's an FPGA company called Microchip, 0:15:17.640 --> 0:15:20.560 and they launched a product that incorporates our technology into 0:15:20.600 --> 0:15:24.160 their product. UM, they package it up, they work with 0:15:24.200 --> 0:15:26.120 the FABS, they do all of that work. We we 0:15:26.160 --> 0:15:29.280 are what's called a fabulist semiconductor company, and so we 0:15:29.400 --> 0:15:32.320 provide technology to other companies that then build build a 0:15:32.400 --> 0:15:35.680 larger product pulled together documentation and software and things like this, 0:15:36.160 --> 0:15:39.040 and then we enable that product to be built. UM. 0:15:39.080 --> 0:15:42.280 There are other cases where we actually are the the 0:15:42.280 --> 0:15:45.640 the device manufacturer, and so we do have developer boards, 0:15:45.640 --> 0:15:48.960 for example, and so we'll make little boards that developers 0:15:48.960 --> 0:15:51.040 can use to write software and work with risk five 0:15:51.160 --> 0:15:54.400 and that's been very popular. I mean, all of those 0:15:54.440 --> 0:15:57.720 sounds really great. UM. And obviously you're making a good 0:15:57.800 --> 0:16:01.880 case for the technology, but what's what's adoption like? Do 0:16:01.960 --> 0:16:06.000 you find that anyone is resistant to this new idea 0:16:06.160 --> 0:16:08.520 or is everyone um sort of jumping on board with 0:16:08.600 --> 0:16:12.680 the customization pitch. Well, So I think that there's different 0:16:12.680 --> 0:16:17.240 aspects of this. One is risk five as a industry trend, 0:16:17.440 --> 0:16:20.720 and there we see just a complete explosion in adoption 0:16:20.840 --> 0:16:24.360 from kind of the who's who of companies including Google 0:16:24.400 --> 0:16:29.680 and Video, Qualcom, Samsung, UM, Western Digital, also Ali Baba 0:16:29.760 --> 0:16:32.040 and all kinds of all kinds of companies are are 0:16:32.080 --> 0:16:34.640 working within doing things with Risk five, and so I 0:16:34.640 --> 0:16:37.040 think there's a question of when, when and if does 0:16:37.160 --> 0:16:40.520 Risk five take over everything, which I think is an 0:16:40.640 --> 0:16:44.000 entirely possible future where you know, you fast forward a 0:16:44.040 --> 0:16:46.440 hundred years from now, it could be risk five and 0:16:47.040 --> 0:16:50.240 um and hopefully much sooner than a hundred years um. 0:16:50.520 --> 0:16:54.080 With SCIE five we're in we're in effectively all of 0:16:54.120 --> 0:16:56.920 the big semiconductors products, and we don't talk about our 0:16:56.960 --> 0:17:00.960 customers specifically, but the adoption and the drive we're seeing 0:17:01.000 --> 0:17:04.800 is quite intense. It's very exciting time right now. So 0:17:05.080 --> 0:17:08.119 do you have any competitors in the Risk five space 0:17:08.240 --> 0:17:11.199 or is it too early for that and would you 0:17:11.240 --> 0:17:14.280 expect them to emerge at some point? Yeah, So, so 0:17:14.320 --> 0:17:17.280 there are other companies that build risk five processors. UM 0:17:17.480 --> 0:17:20.199 Risk five as an open standard really enables this, and 0:17:20.240 --> 0:17:23.359 we encourage that. It depends on what you mean by competitors. So, 0:17:23.400 --> 0:17:27.760 for example, UM risk five is why is widely adopted 0:17:27.800 --> 0:17:30.639 within the academic context, and so there's many university projects 0:17:30.640 --> 0:17:33.200 where people are building risk five course. UM. I don't 0:17:33.200 --> 0:17:35.840 see these as competition. I see these as you know, 0:17:35.880 --> 0:17:38.920 the next generation of engineers and we'd like to hire UM. 0:17:39.000 --> 0:17:42.440 But you could say that they're playing in the same space. UM. 0:17:42.480 --> 0:17:46.320 Many companies are also building small risk five cores themselves, 0:17:46.480 --> 0:17:48.960 and I think that they're playing with the technology in 0:17:49.000 --> 0:17:51.640 many cases and they're starting to understand it. UM. There 0:17:51.640 --> 0:17:54.560 are a relatively small number of companies that are commercial 0:17:55.440 --> 0:17:58.280 productizing and commercializing risk five course the way we are, 0:17:58.880 --> 0:18:01.119 and I think that is also great. That's that's one 0:18:01.160 --> 0:18:04.240 of the great things about this standard is that you 0:18:04.240 --> 0:18:07.960 can get different implementations from different companies. UM. In comparison, 0:18:08.080 --> 0:18:10.200 SCI five has been at it for you know, our 0:18:10.280 --> 0:18:12.760 our team has been at for ten years. UH. SCI 0:18:12.840 --> 0:18:14.960 five as a company has been at it for five years, 0:18:15.240 --> 0:18:18.439 and so we're just much further ahead and than the 0:18:18.480 --> 0:18:21.119 pack in terms of the breadth for a portfolio, the 0:18:21.119 --> 0:18:24.560 maturity of the technology, um the established customer base. And 0:18:24.600 --> 0:18:26.960 so I'm very thrilled that our customers love our products, 0:18:26.960 --> 0:18:31.000 and that's really what I am We focus on. So 0:18:31.160 --> 0:18:33.840 one of the things that's come up a few times 0:18:33.840 --> 0:18:36.919 in some of our discussions has been, you know, you 0:18:37.000 --> 0:18:39.080 sort of hinted at it, which is the sort of 0:18:39.119 --> 0:18:41.920 death of Moore's law, you know, the sort of the 0:18:41.920 --> 0:18:46.800 the increasing difficulty of the arms race, how much it 0:18:46.960 --> 0:18:51.240 costs to improve speed size. And when we've heard from 0:18:51.240 --> 0:18:53.480 a couple of people, they say things like you asked 0:18:53.480 --> 0:18:57.600 the question like, okay, what could theoretically up end the 0:18:57.640 --> 0:19:00.840 dominance of a company like Taiwan Semiconductor. And the answer 0:19:00.920 --> 0:19:04.880 that a couple of people have given up is, well, 0:19:04.920 --> 0:19:09.960 maybe customers don't necessarily need the very fastest chip anymore. 0:19:10.000 --> 0:19:15.040 Maybe like the sort of traditional metrics of performance that 0:19:15.119 --> 0:19:17.439 we sort of think about with chips, certain measures of 0:19:17.480 --> 0:19:20.879 speed and so forth do not become the key things 0:19:20.920 --> 0:19:24.240 that people absolutely need. I'm trying to understand a little 0:19:24.280 --> 0:19:27.880 bit more like does risk five, does your model sort 0:19:27.880 --> 0:19:31.760 of create the avenue for a sort of I get 0:19:31.800 --> 0:19:35.280 maybe break in the arms race of chips such that 0:19:35.720 --> 0:19:38.439 the industry can go in a more direction, or in 0:19:38.480 --> 0:19:43.679 a direction that's not just about more expensive, faster chips 0:19:43.960 --> 0:19:48.000 and offer different value propositions. But I think that it's 0:19:48.080 --> 0:19:50.919 it's hard to paint the entire industry with one brush. 0:19:51.560 --> 0:19:55.880 So I think that if you look at UM, for example, 0:19:55.920 --> 0:19:58.960 the Internet of Things, IoT space for example, there are 0:19:59.040 --> 0:20:01.320 time to market is worth a lot, right because a 0:20:01.320 --> 0:20:04.320 lot of it is finding product market fit, and so 0:20:04.440 --> 0:20:07.520 spending three to five years to build a chip is 0:20:07.680 --> 0:20:12.600 really kind of an untenable place because five years from 0:20:12.600 --> 0:20:14.600 now the industry will be changing, and you know it 0:20:14.600 --> 0:20:16.399 doesn't it doesn't really make sense to do that. And 0:20:16.440 --> 0:20:18.520 so I think that that that is a pretty big 0:20:18.520 --> 0:20:20.720 shift and how people think about things from back in 0:20:20.840 --> 0:20:23.959 the Windows PC days, where until it would come out 0:20:24.000 --> 0:20:26.119 with a chip every every so often and it was 0:20:26.160 --> 0:20:29.000 just the steady, say, march up and performance. On the 0:20:29.040 --> 0:20:30.800 other hand, I think you look in the server space, 0:20:30.840 --> 0:20:34.320 for example, and there there you're kind of limited by power, 0:20:34.680 --> 0:20:38.119 and so each each socket in a large scale cloud 0:20:38.240 --> 0:20:41.720 data center can dissipate roughly two hundred fifty watts, and 0:20:41.800 --> 0:20:44.080 so the question is how much how do you get 0:20:44.200 --> 0:20:46.359 how much can you get out of that? Right? And 0:20:46.440 --> 0:20:49.879 so to me, uh, we we as an industry have 0:20:50.200 --> 0:20:56.040 solved that problem by pushing, pushing process technology and chasing 0:20:56.080 --> 0:20:57.959 from one fab node to the next, which is very 0:20:58.000 --> 0:21:01.280 expensive and as you say, it's it's not a particularly 0:21:01.320 --> 0:21:04.439 scalable approach. The other way to tackle this is to 0:21:04.680 --> 0:21:08.320 innovate in the hardware. And so what we're seeing now 0:21:08.400 --> 0:21:11.440 is we're seeing a rise of GPUs, and we're seeing 0:21:11.440 --> 0:21:13.639 a rise of machine learning accelerators, and we're seeing a 0:21:13.720 --> 0:21:18.199 rise of other other accelerators, and to date they're primarily 0:21:18.840 --> 0:21:22.879 purpose built. So I want to accelerate video and cooding, 0:21:23.480 --> 0:21:25.440 you know. And so if I want to accelerate video 0:21:25.440 --> 0:21:27.800 and cooding, I can build a box that does video encoding, 0:21:27.840 --> 0:21:30.199 and I can build a chip for video encoding. The 0:21:30.240 --> 0:21:32.399 far extreme of that is I want a fully general 0:21:32.440 --> 0:21:34.560 box that can do anything. And that's what you see 0:21:34.560 --> 0:21:38.879 with typical CPUs. There's a big space in the middle 0:21:38.960 --> 0:21:41.880 where you say, if you're a cloud provider, for example, 0:21:42.680 --> 0:21:45.520 I want to be able to run a wide range 0:21:45.600 --> 0:21:48.760 of different specialized workloads. And so I want to do 0:21:48.800 --> 0:21:50.520 a little bit of video encoding, I want to do 0:21:50.520 --> 0:21:52.199 a little bit of signal processing. I want to do 0:21:52.200 --> 0:21:54.639 a little bit of machine learning. I want to do 0:21:54.640 --> 0:21:56.040 a little bit of graphics. I want to do a 0:21:56.080 --> 0:21:57.960 little bit of this. And so what you can do 0:21:58.000 --> 0:22:01.919 is you can build what are called heterogeneous computers. And 0:22:01.960 --> 0:22:06.760 these heterogeneous computers can have accelerators for different domains and 0:22:06.920 --> 0:22:10.159 assembling those together into one package is something that you know, 0:22:10.240 --> 0:22:13.760 Risk five is standard and a lot of the ecosystem 0:22:13.880 --> 0:22:16.480 that we're helping drive well, I think really changes the 0:22:16.560 --> 0:22:19.080 nature of compute. And so I think that we as 0:22:19.119 --> 0:22:21.919 an industry or shifting and it's kind of hard to 0:22:21.960 --> 0:22:25.400 see the end state from you know, we're halfway through 0:22:25.400 --> 0:22:45.119 this transition, but it's a really exciting time. So going forward, 0:22:45.359 --> 0:22:49.040 how do we know if people are using Risk five? 0:22:49.160 --> 0:22:52.080 So my understanding is that you know, if you're using 0:22:52.119 --> 0:22:54.960 something from ARM, you have to sign like some sort 0:22:54.960 --> 0:22:57.080 of licensee agreement and we can kind of figure out 0:22:57.080 --> 0:23:00.639