FLOSS Weekly Episode 690 Transcript
FLOSS Weekly Episode 690 Transcript
Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.
Doc Searls (00:00:00):
This is FLOSS Weekly I'm Doc Searls this week, Jonathan Bennett and I are talking with Stephano Cetola about risk five. He was with us a little over a year ago, and so much has happened since then. And so much of it is exciting that you really need to get this one. There are things being done with risk five that aren't happening with arm. They're not happening with other CPU designs. They are really cool and really pretentious, and that is coming up. Next
Speaker 2 (00:00:33):
Podcasts you love from people you trust.
Doc Searls (00:00:39):
This is FLOSS Weekly episode 690 recorded Wednesday, July 20th, 2022. Risk five revisited. This episode of floss weekly is brought to you by collide that's collided with a K collide is an end point security solution. Built around honest security. You can meet your security goals without compromising your values. Visit collide.com/floss to learn more and activate a free 14 day trial today. No credit card required. Hello again, everybody everywhere. I am Doc Searls and this is FLOSS Weekly. I am joined this week by Jonathan Bennett. Who's looking, he's got his new haircut and he is looking all cool. There he is. <Laugh> Hey, what's the background there is that you is a giant screen in front of a man cave wall, sort of looking a thing,
Jonathan Bennett (00:01:39):
Something, something like that. It's a, it's a built in desk behind me and then a couple of monitors, but I've got a polarizing filter on the camera turned just right. So you can't see what's on my monitors.
Doc Searls (00:01:49):
That's true. <Laugh> oh, that's oh, that's very cool. Yeah. I, I, I, in my car, I have the car. I, I have an old car, so I don't have like the built in you know, screen in a dashboard. I just, I just have a phone that's clamped onto the thing that you shove it to this CD slot, because mm-hmm <affirmative> who does CDs anymore. And that didn't work anyway. And wearing, I always wear Polaroid Polaroid, sunglasses, but I can only see the phone if I twisted sideways. It's the that's deal
Jonathan Bennett (00:02:21):
Over tilt your head to the side as you're driving down bridge.
Doc Searls (00:02:23):
No drive like this. Yeah. Yeah. <Laugh> new thing, new thing. So say just to cut down the reflection on, you've got the leather chair there that makes you look shorter than you really are, and is that's reflective also, but you're not worried about the reflection. You're just worried about what's on the screen.
Jonathan Bennett (00:02:42):
Yeah. I, you know, once, once upon a time I was always very careful before the show. It's like, all right, let's put hack a day up on one side and maybe the TWI logo on the other. And I always, I always had this fear that I'm gonna forget and accidentally leave, you know, a document up that I don't need the world to see, you know, something like that. <Laugh> and eventually I just concluded, I've got this polarizing filter. I can just take care of it. And so I've done that ever since
Doc Searls (00:03:05):
<Laugh> you don't have up to you, the, you have been summon to appear <laugh> none of those <laugh>
Jonathan Bennett (00:03:12):
No, I mean, I'm, I manage it for a lot of different businesses and so I've got, I've got documents with their secrets on it and stuff. So yeah, I, I try to be real careful about that. So anyway,
Doc Searls (00:03:22):
So, so our, our, our guest today is Stephano Cetola, who we had on a little over a year ago. And you actually volunteered, you came forward, say, I wanna be on this show, which we don't usually get.
Stephano Cetola (00:03:36):
Well, it was just such a great conversation. Last time I was excited
Doc Searls (00:03:39):
Stephano Cetola (00:03:39):
Come back and do it again. It was <laugh>
Doc Searls (00:03:42):
There he is. Yeah. So I, I might as introduce you cuz we don't, we didn't start with an ad. So I'm just so so, so Ste for those of you are who are visually enabled <laugh> to, so come on is the director of technical programs for risk five international he's development managed numerous open source initiatives and software and hardware over the course of his 20 year career in technology. Before that he worked for Intel contributing to the Octo project, embedded Linux in various open source firmware efforts. So welcome to the show, dude. He's back. Thank you.
Stephano Cetola (00:04:19):
Thanks. So, so it's great. Great to be back.
Doc Searls (00:04:22):
Yeah. So, so why don't we just start by filling us in, on what's going on for the last year, year or so? I mean what's different this time because one of the reason we have you on is because it is different things are going on.
Stephano Cetola (00:04:34):
Yeah. Yeah. One of the reasons why I reached out was because we know we had a lot of a lot of work that got done last year. And so it's exciting to see the risk five community growing. It's exciting to see a lot of stuff getting done. And so I just wanted to come and chat about some of it. I wanted to start by talking about some of the extensions that we've ratified, but before I do that, let me pull back a little bit for your audience. For those who are relatively new to risk five risk five is an open source, ISA instructions at architecture. I'm gonna talk a little bit about some extensions that we've ratified, but before I do that, I should probably talk about why we care about extensions. So the risk five ISA, like I said, it's open source.
Stephano Cetola (00:05:13):
So opposed to, you know, Intel like city six and arms architecture. This one is developed in a community setting in an open source CA setting. And the base of the ISA is essentially the groundwork for a processor. So you've got as an example RV 32, I is the 32 bit integer base ISA. And this gives you things like 32 bit. It tells you whether it's 32 bit 64 bit, 1 28 bit it gives you the register layout. So for the integer, there's 32 registers it describes how operations are done. So what do the op codes look like and how would one do an and or an or so that's essentially the foundation for risk five, but then on top of it, we layer extensions that do useful things. So things like multiplication with the M extension or floating point with F and D.
Stephano Cetola (00:06:12):
So what you can imagine is this base, ISA has several extensions layered on top of it to allow you to do something even more useful than a calculator like boot Linux. So you get enough extensions piled on top and you can actually boot Linux and the beauty behind the standardization of these extensions is, as you might imagine, if we don't standardize around multiplication, everyone's gonna go off and do it their own way. Now with multiplication, that might not be such an issue, but when you start to get into the more complex things, I'm gonna talk about like vector or cryptography. We really want the community aligned behind specific standard extensions. That way we reduce the effort the community gets to work on it. We release one version of it and everyone can build their value. Add on top of that. Now that being said, we, we don't discourage custom extensions.
Stephano Cetola (00:07:00):
So if you wanted to go off and write some, you know, fancy extension that doesn't standardize and isn't common throughout risk five ecosystem, you can absolutely do that. We reserve op code space specifically for that to happen. So, so that's essentially the concept behind why we have extensions. And the power is really in the flexibility, right? When you're building when you're building a microprocessor or or a larger, you know, enterprise level chip you, you get to pick and choose which extensions you're going to add. And this is much this is much different than in the traditional setup we have today, where if you buy a platform, you're already getting all the stuff, regardless of whether or not you're planning to use vector, vector's already on your chip. So I mentioned vector a couple of times. That's one that we ratified.
Stephano Cetola (00:07:47):
A lot of folks have been looking forward to this one. It's actually been frozen. It was frozen for quite a while. So we had plenty of time for the community to review it. But we just recently ratified it in December. The idea buying vector for those who aren't familiar with vector processing it's a type of parallel processing. The concept has been around since the sixties and we use parallel processing in our daily life all the time. Your phone has tons of cores in it. Your desktop has even more cores in it. That's one type of parallel processing. Vector is a different type which uses multiple different pieces of data to operate on the same instruction what that all translates to in practice. One of the common examples, there's things like graphics. You can imagine on your screen, there's a whole bunch of pixels and there's set up in a grid fashion, and you may wanna perform some operation, some instruction on those pixels, but you may wanna do it to more than one pixel.
Stephano Cetola (00:08:45):
So you might wanna do it to a whole row or to several rows with a traditional processor, a traditional core you're gonna line up those instructions and each of those pixels, and one at a time you're gonna loop through 'em. And if you're particularly smart, you're gonna spin up eight cores and you're gonna do eight of 'em at a time, but there's an inherent limit to what you can do there. You're limited by the number of cores with vector processing. You're allowed to create special registers and special instructions that then act on multiple different pieces of data. So you can change say the pixels in your entire screen or in a set of rows would be the analogy. And there are multiple other places you might want to do. This sort of computation graphics is just a, a common one that people use, but there's also machine learning, artificial intelligence, lots of different aspects of computing where multiple data being operated on the same time is, is critical. So that's, that's one of the extensions that we ratified that lots of folks are excited about. As you imagine with graphics, you're gonna start to see this coming into play with things like graphics cards, which we all care about. But then also in things like neural network processing. So when you hold your phone up to your face to open it, it would be really nice if that didn't take 35 seconds. So that's one of the things that vector gets us is the ability to put our hand down quicker. <Laugh>
Jonathan Bennett (00:10:08):
Hey, I'm, I'm curious about this before we go on to the next one. I, I, I've gotta ask this almost seems like this is being implemented in risk five differently than some of the other platforms. So for example, you, you mentioned cell phones, so our cell phones, they have an arm chip, and then sometimes it's on the Silicon. Sometimes it's a separate chip, but they'll be say a, a power VR whether it's physical or, or logical, it's a discrete device. The, the way you're describing these vector extensions, it almost sounds like you're taking the GPU stuff and actually making it part of the CPU proper. Is that kind of a different approach than any of the other chip designs have gone with?
Stephano Cetola (00:10:51):
The approach is different in a subtle way. Excuse me, it's a little different than that, but let me touch on what you're getting at. So what you're getting at is with risk five, we've called out these extensions. So it's very obvious, you know, here's the part of the processor that's doing your traditional compute. Here's the part that's doing vector. And we are doing that in a very obvious way. Other chip manufacturers historically have done it, but have done it in a much more subtle way. So it's hard for you to know, is there vector processing on my chip or is it its own accelerator? The answer is most of the time, it's both. Most of the time there is a part of that Silicon that has some vector processing capability, but whether or not they want to talk about it is up to the person manufacturing that chip, what they will likely talk about is their accelerator that they put on there.
Stephano Cetola (00:11:37):
So when they wanna sell you a phone, they're gonna tell you that they have a fancy accelerator that's, you know, maybe on a special bus connected to the chip, that's gonna do neural network processing or image processing extra in an extra quick manner. So, so risk five, isn't doing it differently that way in that we are just being much more transparent. So we are saying, here's our vector extension. That's separate from the base ISA, and here's how you could implement it. And then we leave it up to the implementers, the folks actually making the chips to say, okay, well, I'm just interested in creating an image processing accelerator. So maybe I want a very simple core. Like I was saying, the base ISA has several extensions. You could layer on top, but you can keep it very small. So maybe a very small risk, five core, and then connect that to a specialized unit for doing vector processing. And that piece of Silicon would be considered an accelerator. And the beauty of risk five is that this is all very transparent. You're able to look at the specification and see how people are actually creating things. Even if they don't open up the source to you, you can still get an idea for what they're doing. Does that make sense?
Jonathan Bennett (00:12:43):
So it, it does. So someone could take the risk five ISA and build a dedicated GPU out of it. Now, obviously we're, we're, we're kind of heading the clouds with this idea, but it, it sounds like that's a, a possibility.
Stephano Cetola (00:12:56):
Absolutely. And it's not that far off. I mean, if you think about it if you, if you're out there building SOCs today and you're using risk five, you're most likely doing graphics processing. So there's going to be some, some graphics capability. So that chip and the chances are that you're offloading, you know, the amount of work you need. So if you're doing a general purpose compute, you need a decent amount of graphics computational power there. The idea that someone's out there creating a dedicated GPU is a big step, but the groundwork has already laid for that. So it's not that far fetched.
Jonathan Bennett (00:13:29):
Yeah. Very interesting. All right. I, I'm sure you wanna talk about some of the other extensions I see on our notes here, that there are also cryptographic extensions, which, you know, have a, a lot of potential in different places. Why don't we talk about that
Stephano Cetola (00:13:41):
Indeed? Yeah, no, we're excited about the crypto. So we did crypto scaler. So scaler, you can think of as just to simplify down, we have scaler operations and vector operations. That's a, of gross over simplifications. So <laugh> my apologies to those out there who actually study vector processing. But the idea is for, for crypto, we started with the scaler operations that are very common. So you can think of shot 2 56 a AEs, the algorithms that you're used to hearing, even if you don't know what they are, you hear about them because you use them every day. You use them every time you type HTTP S into your browser, assuming you're actually doing that. These crypto algorithms underpin everything we do in computing, we can't do anything in computing without them. The key is we're standardizing around them. And that's, that's why the extension is important. It's not so important in that it's doing something wildly different or <laugh>, it's, it's certainly not doing anything that cryptography scholars are going to dig into. What it does is it enables cryptography scholars, people who are doing security research to layer their ideas on top of a standard and not have to worry about that standard changing in the future. So they know that those algorithms, those instructions are gonna be available to them going forward. And it also opens up the door to more interesting things.
Jonathan Bennett (00:14:58):
You kind of want your cryptography implementation to be real boring as much as possible.
Stephano Cetola (00:15:03):
Exactly. Yeah, no, that is kind of the saying, you want it to be boring. You don't want it to be terribly exciting. And that gets, you know, it's funny that term gets bandied around a lot, because if you look at it from a software perspective, they don't want it to be terribly interesting. Nine times outta 10, they want it to be boring. They wanna be able to compile their code and run it on your machine. Just like they've been running their software on machines for the past 50 years. They don't wanna change the way they're doing things. So a lot of our stuff, I don't wanna say it's just coming up to parody, but what it's doing is it's bringing the cryptography capabilities that software developers want and standardizing that and making it readily available. And it's opening the door to what we want to do next, which is things like vector cryptography. So more interesting extensions that do provide advanced features, but again, giving folks the option of whether or not they wanna implement that in their, in their product. So not saying it's part of the base. I say, it's not something they have to do. It's some, it's an option. It gives them that flexibility to do it or not to do it.
Jonathan Bennett (00:16:04):
I, I'm curious. And we're gonna, we're gonna get into the weeds here for just a moment. I I'm real curious though about how, and I would have, I would've to really look into how the the, the, the, the documentation of this works, but you know, so you mentioned shot 2 56 and I don't, I don't know what the primitives look like, but I'm, I'm curious, so shot 2 56 at this point is just about old crypto and N you know, a couple of weeks ago just said, Hey, we are, we are now ratifying. We have selected some post quantum crypto which have some real fun geeky names, like crystals ki and crystals dial lithium. There's, there's some real sci-fi geeks in some of those groups they're getting, what, what does the, yes, what, what does the path look like to making those next gen crypto algorithms work in risk? Five? Are, are the, are the primitives that that you've implemented in, in these crypto extensions? Are they basic and flexible enough that it's just a matter of writing, you know, writing the code and hooking it up to the extensions? Or is this gonna have to be another round of extensions to support the next, you know, kind of the next ideas in crypto?
Stephano Cetola (00:17:19):
Yeah, no. So you sort of hit on it when, when you use the term implemented. So what we've implemented is really just a framework. So it's not, here's how you're going to here's how you're gonna build your algorithm. Here's the pieces you're gonna put together. It's just the requirement for here's what instruction we expect to be there. And here's how that instruction should work. Now, you go off and implement your shot 2 56, or shot five 12, or your, you know, next generation algorithm. However you want to integrate it or implement it. We are just telling you, here's the standard way that people are going to expect to be able to communicate to the processor. So here's the instruction they're gonna use. So getting to your point of how do we layer on top of that? So we have scale or crypto instructions today that create a base, a foundation for folks to build on.
Stephano Cetola (00:18:09):
We'll add to those in the coming year with vector crypto and your concept of you know, where we go with the really future looking stuff, the quantum stuff, that's where risk five flexibility really comes in handy because now we have a community that's willing to align around, you know, which of these futuristic for pH lack of a better term algorithms, are we interested in standardizing around, but where do we wanna stay away from to allow the industry to create innovative solutions that aren't standards that are their custom instructions? Because if you think about it from a value add perspective, as we move up that stack of cryptography closer to cutting edge, we're getting at the value add, right risk five wants to be simple. It wants to be basic so that companies can come in and innovate at that top level and do some really interesting quantum stuff. And then hopefully the community can look at what's going on and decide which pieces of that they wanna start standardizing, but there's no reason we have to standardize the stuff that's really cutting edge.
Jonathan Bennett (00:19:12):
And, and I'm sure there's kind of a a flow down where some of these things at various companies, cause we see this in other places in the tech industry, you know, AMD or Intel or Invidia, you know, whoever they, they do something really, really clever. And it's, it's just on their stuff for the first couple of years. And then the next thing you know, that that's been folded into the standard and now, right. You know, you buy the next generation of chip and everybody has it. Do you see that kind of flow down happen with risk as well, where the, some of these these extensions that start out as custom just from one company, the next thing you know, they're getting folded in officially.
Stephano Cetola (00:19:49):
Yeah, absolutely. Yeah. You can already sort of see that happening today. So the way risk five is structured as a, as a nonprofit is you have members who are, you know, at different levels at, at the organization. And you'll notice that the companies that are sort of driving innovation, the ones who are premier members are the ones that are slowly integrating some of the big new concept. So things like PAC SIMD is a good example. We have a group from Andy's technology that have clearly already implemented this, but they're trying to figure out, okay, how do we take the stuff that we have implemented and bring some of that into a standard, because at the end of the day, that rising tide is gonna lift all the boats and risk five will do better. And we'll still have our piece that we can layer on top of that.
Stephano Cetola (00:20:36):
So you're already starting to see that in in vector type processing. I imagine you're gonna start seeing it in cryptography specifically around different security stuff. So I'm thinking about things like open Titan that Google did. So that was one of the first areas where folks said, okay TPMS, trusted platform modules is something that everyone needs, right? We all, again, when you hold your phone up to your face, it needs a way to compare what it's seeing with some secured image and that needs to be secured. And so we need some trusted platform module to store things like cryptographic keys, but the way in which we're doing that is not something that companies see as their value add. It's much better if we all combine our efforts around that idea of a TPM. Now, if you think a little bit about that, that TPM today is operating on sort of the traditional computing standards.
Stephano Cetola (00:21:27):
As we introduce things like quantum, there will be places to innovate, even in a TPM. The question is, as these companies go off and do that, what are they gonna find that they think, okay, this is gonna be much better to pull down into a standard so that I can work across organizations with Google and with huawe and with Alibaba to come up with a solution that allows us all to layer on top of it. I think that's what you'll end up seeing is is those, those filters of, you know, we know where we're innovating, what can we bring in and work together as a group doc?
Jonathan Bennett (00:22:00):
Yeah. Excellent. I know, I know there is one other sort of extension or group of extensions that, that you really want to cover. And that's the physical memory protection. And I, I will probably have some follow up questions about this, cause it sounds real fascinating.
Stephano Cetola (00:22:12):
Absolutely. Yeah, no, that one's kind of near and dear to me. So trusted execution environments are something that most people have probably run into whether or not they know it from that, that acronym. But think about arm trust zone or Intel SGX. So I mentioned earlier you know, ways in which we log into our phone have advanced. But it even goes further than that. When you're opening up your banking app, there needs to be a way to securely open that application and know that the phone hasn't been temporary with. And so companies have been using trusted execution environments, which are essentially an area of computation that has a higher level of protection. So a way of quarantining off specific operations or specific tasks in an area that's more secure. Now this is old ha this has been around a while SGX and trust zone have both been around for over a decade.
Stephano Cetola (00:23:07):
The beauty of risk five is that not only do we get to learn from historically what's happened there, but we get to start from an open standard. And that's what physical memory protection is. It's the first layer in that open standard. So as you might imagine, let's just pick on trust zone. Trust zone is something that lots of chip manufacturers have to implement. And there are lots of layers to trust zone. It's not simply one layer of protection. It's, it's actually several layers that the application has to go through to execute its code. All of those layers were defined and there was really, you didn't get a say in that, that <laugh>, you got handed trust zone and told how to implement it, and you get to go implement it with risk five. We're starting with sort of a bottom layer of physical memory protection and layering on top of that until we have a trusted application platform.
Stephano Cetola (00:23:58):
And all of this is being developed in the open. So physical memory protection is sort of the first layer. You can really just think of it as a way to protect the actual physical hardware memory. Almost like a bitmap that sits above that, that the hardware is able to have control over. And what we ratified in December was the next layer up sort of an advanced or enhanced layer of physical memory protection. And we have a security task group that's working on, you know, several different parts of the stack now, and they're just starting work on sort of the application layer. So how does an application that boots through these protected layers of memory, then make calls back in to continue execution in a secure environment? And the part that's interesting is that we're all doing this in an open task group that anyone can join, that anyone can contribute to.
Stephano Cetola (00:24:52):
And it's gonna be a completely different model than what we've seen before, simply because it's going to get to learn from past mistakes and take into account the direction that people are trying to go today. If you imagine, you know, 20 years ago, when they were looking at this concept, they were doing it for things like chip and pin computations completely different than the kind of things we're trying to do in an, in a, in an enclave or in a trusted execution environment today, today, we're trying to shove huge applications into these trusted zones of computation. So not quite chip and pin simplicity. So I think the, the thing that we're gonna win is the ability to look at something much more holistically and build a solution that's geared toward the problems of today.
Jonathan Bennett (00:25:33):
Does, does physical memory protection include things like encrypting memory physically?
Stephano Cetola (00:25:39):
So it doesn't, it sort of stays clear of how you want to implement that encryption. So what it does is it gives you a mechanism by which you allow or deny access to the memory. And then the next layer up essentially allows you to run, let's say, an operating system or a hypervisor and make calls back into that securely. So it really is the foundation. It hasn't even gotten up to the port of, okay, how do I encrypt something inside that memory yet? You will see that there will be instructions. I'm sure that talk about, all right, let's say I want to store something in the trusted execution environment and I wanna encrypt it. And here's where I store my key. That's the sort of problem that we're solving today is those sort of issues. And how do we standardize that? How do we, how do we get everyone to align around specific ways of accomplishing that?
Jonathan Bennett (00:26:29):
So I I've gotta ask how ha it has and how has risk five avoided the classic problem of designed by committee sometimes when you have a bunch of people working on these problems, the, the solutions end up unwieldy, let's say,
Stephano Cetola (00:26:48):
Right? Yeah, no, I was just talking to actually our CTO mark this morning about this, it's all about pushing the responsibility down into the task groups and then the task groups align around a committee. So that's kind of a complicated answer to the idea of we're not a consensus organization. We are an organization that works with task groups that whose job is to get work done. And so if something is standing in the way of getting that work done, it's our job at risk five international to solve those issues and to bring those to closure. But because the, the technical steering committee, the, the sort of high level committee that oversees all these task groups has pushed a lot of the responsibility down into the work group. What you end up is with experts in that field, making the decision. So the technical steering committee oversees it.
Stephano Cetola (00:27:40):
They will get a final say before we ratify specification. However, they have pushed the decision making down to the people who know that area of expertise. And while we have populated the technical steering committee with experts, so we make sure that there's someone on there who really understands security. For example, we under, we make sure that there are actually two different folks that understand different ends of software. So software from a sort of operating system level perspective, but also application level software perspectives. These are things that you don't often see in hardware standards, bodies. They tend to just leave software to the software folks, but risk five is aligning all those folks at sort of that top level committee. But like I said, pushing all the responsibility down for doing the work to the task group and leaving those decisions up to them. Now, the task groups do have to stand by their decisions and give reasons why they made those decisions, but it's, it doesn't become, every committee gets its own say, and everyone gets to throw into sense.
Stephano Cetola (00:28:36):
It really ends up being this tight knit working group of folks who are all aligned around a vision. I would actually pick on the crypto group as a great example of that. If you go to one of the crypto meetings, it's usually somewhere between 10 and 15 cryptography experts meeting to discuss what makes sense in that specification and why. And when they're done with their specification, you can just shy of guarantee that when they hand that to the TSC, it's gonna be flawless or, you know, there'll be a few things that, you know, a few Ts need to be crossed or whatnot, but it really is. It's an, it's a example of how, when you get experts in a room and you work with them to align on a goal, you avoid that problem of, well, the committee each has their own opinion. So you're gonna get 20 different opinions. You really end up with a much more cohesive work process.
Jonathan Bennett (00:29:27):
In this case, the metaphorical committee is made up of engineers instead of middle management.
Stephano Cetola (00:29:31):
Right? Exactly. Yeah. And, you know, we don't, we don't completely remove the management. There is still oversight. <Laugh> companies still have interests and we wanna make sure those interests are, you know, know, are we're aware of them and that we take them into account, but we let the engineers do their jobs and come out with a spec that, you know, when you bring that spec to a cryptography conference and talk about it, you can be proud of it. And you can know that you're, you're standing behind a bunch of cryptography experts who think that this is the right way to move forward. I'll also touch on public review. So staying on sort of this cryptography spec, before we ratified it, we put it out for a 45 day public review period on a public mailing list that really gave anyone a chance to comment.
Stephano Cetola (00:30:14):
So if there's, you know, dissent in the broader community, aside from just risk five international and people who are members, but the broader community, and that includes the Linux kernel community, and you know, other communities that you can imagine might have a thing or two to say about a risk five specification <laugh>. We give them that opportunity to voice their opinion. And we take that back and discuss that both at the TSC level with the technical steering committee and at the board before we ratify it. So you can be sure that if there is dissent somewhere out there, that dissent is rolling all the way up through our process of ratification and may even change what the spec looks like. If the dissent is it warrants that sort of, you know, conversation.
Doc Searls (00:30:56):
So I'm, I find myself wondering about how users or buyers in the actual world are gonna see and touch this. But before we go there, I wanna let everybody know that this episode of us weekly is brought to us by collide. It admins often feel like they have to choose between their commitment to site for security and their duty to protect their employees. Privacy. Naturally, you need to safeguard company data against hack and breaches, but you don't want to turn your workplace into 1984. Traditional MDMs, give the it team complete access and control over company devices. But since employees are inevitably going to use their work laptops for personal activities, these tools can saddle you with surveillance capabilities. You never wanted like access to photos and browser history. So before you know it, the end users are complaining about all the security agents slowing down.
Doc Searls (00:31:51):
Their laptops. Developers are frustrated by lack of autonomy. People start secretly working on their personal devices just to get things done. So it's easy to fall into the trap of top down security, but it's not the only option. Collide is an end point security solution built around honest security. Their philosophy is that employees, aren't your busy security risk. They're your biggest allies. And your relationship with them should be based on transparency and informed consent. So collide works by notifying your employees of security issues via slack, educating them on why they're important and giving them step by step instructions on how to resolve them themselves for it. And security teams collide provides the right level of visibility for Mac windows and Linux devices, and it addresses high risk issues that can't be solved through root force or automation. What's more, your end users can see exactly why and how every piece of data is being collected via collides user privacy center and their open source code base.
Doc Searls (00:33:00):
You can meet your security goals without compromising your values, visit collide.com/floss and find out how, if you follow that link, they'll hook you up with a goody bag just for activating a free trial. That's K O L I D e.com/flos. Okay. So, so as we're, we're looking at the real world here, I mean, I'm, I'm, I'm, I'm fascinated by all this stuff. You're, you're, you're working on with the engineers and the companies and all the different groups, but somebody wants to use one of these. I, I know you've got some stuff coming up, so where do we see this in the real world? Where can you see and touch and hold and buy stuff? Absolutely.
Stephano Cetola (00:33:43):
Yeah. <laugh>, that's what it comes down to is what can I get my hands out and
Doc Searls (00:33:47):
Actually boots sooner or later people buy this stuff. Right.
Stephano Cetola (00:33:50):
Right. Well, so we've had a lot of interesting stuff happen. Sorry. The past year's been obviously a little rough with the chip shortage. So the chip shortage has affected folks who are ordering in small quantities. And as you might imagine, risk five tends to be ordered in small quantities, cuz these are all relatively new efforts. That being said, we have had some really interesting stuff happen. The, probably the most pertinent and the one that happened the most recently was the laptop that got announced. So I wanted to sort of touch on it because there's been already a little bit of chatter around, you know, what is this laptop? What's the point of it? So you have to realize with a new architecture, especially one that isn't being, you know, brought to you by Google or brought to you by large corporation X, Y, Z, there's going to be efforts to spin up products that may not be for everybody.
Stephano Cetola (00:34:38):
So this laptop is a developer laptop, but it's not meant for your average kernel hacker to go compile the Ker on. It's really meant for folks who are currently developing applications and wanna natively compile on risk five. So if you're, it doesn't matter what type of application, the, the idea being that it's not something that's gonna be a workhorse for you as a kernel engineer or as someone working on a large web application to regularly compile your app. But rather for someone who's interested in trying out risk five native compile on their app. So as an example, we're already trying to get these laptops in the hands of distro folks from SUSE and from red hat. And from ATU, we're hoping to get some shipped over to the folks at VMware, other other software vendors. We'd like to see them compiling their stuff on a risk five because that feedback that we can get from that native compile, even if it's not, you know, in a traditional C I C D setup, you know, it's not a server that I hand them to put into their server farm and run consistent builds on it still gives us feedback and is valuable for the community.
Stephano Cetola (00:35:45):
And so, you know, that's really the goal with this laptop is how do we sort of
Stephano Cetola (00:35:51):
Be a first mover and put a line in the sand of all right. Here's how one might build a laptop. And then what feedback can we get back from that into the community to help folks who are interested in getting involved specifically in software. So getting involved in hardware is pretty straightforward. I'm sure at some point in this interview, there'll be a link to risk five website and you can go join for free and be a member and will happily help you get involved and, you know, change how the ISA works and contribute there. Getting involved in software can be much more difficult, especially when you're, you know, relying on QMU, you're relying on some sort of emulation. So actually having a platform in front of you is key. Now we have a lot of developer boards out there. We've got a D one board and I think is under a hundred dollars now.
Stephano Cetola (00:36:37):
The folks from beagle board actually came out with a prototype. So they were hoping to go to mass production with that. But again, because of the chip shortage and other issues that was difficult. So hopefully this year or early next year, we'll see sort of the beagle board version two for risk five. But these smaller developer boards, or I don't wanna say smaller, but also more complex F PGA based boards. Aren't really ideal for software developers while the little dev board is great and you can compile stuff on it and play with it. It's not a traditional setup. A laptop is a much more traditional setup that you can sit down on, download your application, compile it and start to bugging. And it also gives us as risk five international and opportunity to seed the software community with with these laptops in order to get feedback for how we're doing, you know, what parts of our tool chain need work? You know, what parts of GCC need work? What, what pieces of that code are compiling and much, much larger. We need that sort of feedback so that we can do our job.
Doc Searls (00:37:37):
So in the video version of this show we just saw a story in ours, Technica that had this phrase in it, Silicon replacement, which, and I'd never thought about that before that, wait a minute, because the one thing you imagine is you're never gonna replace the, the CPU. It's like the head, you don't replace the head on somebody, you know, and they're, they're not somebody else, but, but the idea of replacing a CPU makes complete sense. Actually, you should be able to do that. Is that, is that a, a design, constru or design feature that came along in the first place that you wanna do Silicon replacement as like have modularity go all the way down, right. Replaceability to go all the way down. That's cool.
Stephano Cetola (00:38:19):
Yeah. Yeah. This is a conversation we have quite a bit in, in our community. Part of the reason is, and I'll, I'll use as an example, a company called framework it's frame.work if you wanna browse to their website. And I'm not being paid to endorse them. <Laugh>, <laugh> great people though. And they make some really interesting laptops. The laptops that they make are by their very nature serviceable, you can open them up, you can replace the, the motherboard inside them. You can replace the the screen. So this concept of, you know, replacing your processor in your laptop is something that we're really interested in. And now that's twofold. There's two reasons for that first off, because we know that as risk five gets spun up, there are going to be less powerful processors than something you can go buy from Intel.
Stephano Cetola (00:39:06):
So this first laptop that you buy is not going to be as powerful is the thing. You can go grab off the shelf from Lenovo that's for sure. And so that upgradeability, the ability to swap out that CPU is something that we'd be interested in, and it is sort of what risk five is built for is that modularity, right? So I've been talking a lot about extensions at the beginning. The whole idea that risk five can be completely modular means that you can decide what all of your peripherals look like. And so the chip can be completely sort of replaceable. You can sort of remove that module and put it in a different module with a different chip. That's something that flexibility isn't something that was historically possible. Mostly because vendors were sort of locked into different levels of support, right? They were going to sell so many million of these units. They needed to be sure that they could support those units over a certain number of years. And that those units would be replaceable by the next platform, whatever they came out with next, with risk five, we're trying to be much more flexible in how people build laptops and build their products. And so the idea that you might have sort of a system on a module that's able to be replaced, and you could upgrade to your, you know, next, next version of your risk processor, something we're definitely interested in.
Jonathan Bennett (00:40:18):
So it's interesting. You mentioned that I happen to have a well, it's a raspberry PI four cm, but it is on a little carrier board to make it pin out compatible with the Nvidia Jetson stuff. And this is actually for putting various devices in a, in a new device called a touring pie too, which I did a review of over Ron Hackaday, it's a really interesting concept. And, and what I really wish is that more people would get on board with a single pin out, whether it be, you know, the, the Invidia Jetson pin out or something from raspberry pie or even something from you guys, I don't care who comes up with it. But there's no reason why this laptop that you guys are putting together would need to be limited to a risk five processor. Right. You know, if somebody wanted to come along and build an arm, processor uses the same pin out that that should work too. Is, is that something that's, that's also kind of been rolling around in, in y'all's minds.
Stephano Cetola (00:41:19):
I can say that there's been a lot of discussion around the actual pin out of the chip, but yeah, from sort of the system on module concept, I think you're hitting at the right idea there. I think it's if, if you check out what framework is doing, right, framework allows you to sort of take out the, the part of your motherboard that you would wanna upgrade after so many years. And so if you think about that from the standpoint of, well, today, you can do that with an Intel chip. Is there any reason you couldn't do it with a risk five chip and simply plug everything into that board? The answer is absolutely that that's something, one of the things that was really exciting in working with these folks that are building the laptop is when I mentioned to them framework, I said, Hey, I'm, I'm excited to see you guys doing this laptop.
Stephano Cetola (00:42:00):
And we talked about the fact that, you know, it's, it's only gonna have four cores. It's only gonna have 16 gigs of Ram. So we're not talking about a rocket ship just yet. But the first thing I said to them was, would you be interested in talking to companies like system 76 and framework and their answer was yes, because that's sort of what risk five is about. If they wanted to get into generating a proprietary system that only they could buy into, they can do that today within Intel, but with risk five, they can work with system 76, work with framework, sort of start raising the bar for what we can do with laptops in a way that they couldn't before. And I was really excited to hear that I was also, I sort of tried to pitch the idea, being an open source person, tried to pitch the idea of, you know, if you opened up your CAD files and opened up your bomb, it would be easier for folks to build stuff.
Stephano Cetola (00:42:45):
And I just did it as sort of a nod in the right direction. Assuming I would not get any traction there, but I have actually seen traction. So I've been chatting with them about, you know, what would go into that? How would they be able to do that? You know, because obviously they're working with vendors, so they have to check all their boxes and make sure that they're not stepping on anyone's toes, but they were immediately interested because they realized that, you know, it's not, we're not talking about competition, right? Someone who goes out and buys this laptop is not gonna go out and buy a framework, laptop. Those are two different concerns. People are solving different problems with those things, just like if they were to go buy a system, 76 laptop, it's a different concern. And so by enabling all of these laptop manufacturers to build with risk five hardware, they're opening up the door for, well, it doesn't have to be risk five, really any chip can be sort of picked in, placed in that, in that way. And I think it would be a great, you know, personally, that's something I'd love to see. I'd love to stop, you know, buying a new laptop. And luckily I have kids that I can recycle my laptops too, but for those who don't, you end up with older laptops that are pretty much not serviceable and I'd love to see that change.
Jonathan Bennett (00:43:47):
Yeah. So I, I first off, I would love to see somebody from risk five have a conversation with the touring machine guys who are the ones building the touring pie two. I think that would be, that would be cool to be able to put a risk five chip in with the other arm chips on one of those. Yeah, absolutely. Also, if you guys wanna send me one of these laptops, I'll be glad to do a review on it. <Laugh> excellent.
Stephano Cetola (00:44:14):
As you would imagine, a few of those
Jonathan Bennett (00:44:16):
<Laugh> right, right. There's there's the Turton PI V2. It's, it's a, it's a neat, it's a neat idea. It's it's you, you can put four different essentially system on a chip system, on module boards on one and you know, play with Kubernetes or what have you. Right. You know, when I first, when I first read about the, the laptop that is gonna be risk five, I was reading the article and then I got to the end of it. And they said, oh, by the way, this is, this is also an NFT. And I'm like, is this actually legit? Because right. I'm at the point now to where anytime somebody talks about NFTs or that side of crypto, it's like, I, I just don't know about this. You're
Stephano Cetola (00:44:53):
Immediately suspicious. No, I mean, you have to think about it this way. This is the first laptop that anyone's even thought about building for risk five. So the company involved in doing this is obviously going to be something a little different. You're not going to see your, like, it's not gonna look like in the Novo. That's not what the process you're gonna be going through. <Laugh> and you can sort of think about this as a prototype in its, in and of itself right there. I think the, the number of units we're talking about getting up to if possible would be 10,000. So we're not talking about huge numbers here. So when you're talking about that small amount of numbers, there's going to be a lot of marketing buzz around it and different ways to try and spin the laptops to make it interesting. What I can tell you is at the end of the day, when this actually starts shipping and you go and look at the specs, it's gonna be pretty straightforward. You're gonna look at it and think, okay, yeah, this is this is a laptop that I probably wanna boot Linux on. And I probably wanna pick one of the Linux distros that gives me, you know, more bang for my buck in terms of performance.
Jonathan Bennett (00:45:52):
All right. And we're, we're getting close to running outta time, but I've gotta ask about horse Creek that this is kind of follows the Intel naming scheme for their processors. But this is, this is something different. Give us the give us the quick version of this.
Stephano Cetola (00:46:08):
Yeah, absolutely. So, I mean, you know, being an alum from Intel and having worked on risk five, even when I was at Intel, I'm really excited about this because it does show that, you know, Intel is, is, is changing gears. They're, they're shifting the, the direction their ship is sailing, and that's gonna take a while, but when it does shift, I think when you start seeing someone a company of that size, being able to put their weight behind generating a platform, I think you're gonna see something really impressive because one of the problems I'm trying to solve today is how do I get hardware into the hands of distros right. Linux DROS, that's going to be a key to risk five software success. We need to make sure that folks from Debbie and from ATU, from SSA, from red hat, all of these Linux diss need to have hardware to actually do their job.
Stephano Cetola (00:47:00):
And while it's a lot of fun working with some of the smaller companies, folks like BELE bone are sort of BELE board and and this new laptop company, it's fun working with them. And we get to create some really interesting products. That's not what these D destroys need. They need something that's much more that looks like a server. And I can tell you from having worked on these platforms before it is a huge lift, but it's not a huge lift for Intel. That's something they do every day. So if anyone can rock the boat, it's them, they can come out with something that I can then go hand off to folks over at, you know, open SSA and at fedora. And you'll see sort of the, the C change will begin when the software starts natively compiling on risk five, and suddenly you can go off. And when you're, you know, hitting your download page for your software, the options are Intel probably M two processors for Mac, and then risk five. I think when you start seeing that, it'll be because of companies like Intel that have put their weight behind producing platforms that software engineers can use to compile their, their code.
Doc Searls (00:48:05):
I, I, I wanna ask some, some more future questions, but first I wanna tell us about club TWI club TWI is great. It's another way to support our network. As a member, you get access to ad free versions of all the shows on TWI, as well as other great benefits as a, a bonus twit plus feed. That includes footage and discussions that didn't make the final show edit as well as bonus shows, we started such as the, his fizz ask me anything and fireside chats and Jonathan's own show <laugh> on the weekends. And you know, with your favorite twist TWI, I was to say, twist TWI guests and cohost as weekly plus listeners, you may also be interested in getting to it. <Laugh> checking out the unsettled Linux show. That's Jonathan's show, we're doing a set of order.
Doc Searls (00:48:59):
That's available only. This is important to club TWI members. So sign up for club TWI for just $7 a month, head over to twi.tv/club TWI and joined today. We thank you for your support. So getting back on my small screens that I'm unaccustomed to, so I can bring our, our images back up front. I, I, I wanna ask about the far future here. And I, I, I asked this in with, with this perspective I, I started covering as a journalist, the, the chip business in the eighties, right? And, and this is when when risk first showed up as a topic reduced instructions, set computing. We haven't even visited that acronym yet. And and the big players were Motorola versus Intel and, and and AMD and a few others that were, you know and and spark had come along.
Doc Searls (00:50:01):
I spark international was a phone that rang on my desk for a while, cuz I, I was working for sun as an outside operation at the time is now said contributor. And and it had the, the complete disinterest of all of the companies who were working on spark in cooperating with each other. So there's two things going on here that I'm observing. One is the Linux foundation has done a really good job with all of its operations of pulling together these giant companies, which of course frightens a lot of people saying, oh, the giant companies are running everything and you've covered a little bit how you keep that from happening cuz kind of basic things and engineers that are close to whatever needs to be done. But the, the important thing to me about covering this in the eighties is who won arm did arm was acorn back then and nobody cared about them, right?
Doc Searls (00:50:54):
It was, they had nothing going on <laugh> but in the, the a and R stands for acorn and, and they're sort of the prevailing one at this point. So where is risk five or whatever numeral it will have <laugh> you know, in the, for future, you know in this, in this game, whatever this game is because I mean, part of what I wanna see is this complete modularity, you covered a few minutes ago because I think that's, that's so much better and approach to everything than the complete verticalized world that we see to be living in now. Right? Where apple, especially they're from Silicon on up, but I've even heard rumors from no less than an apple genius who I will not identify saying, no, we want to be based. We want, we want this thing to be more modular. We want these things to be more fixable. You know, and not just by us, but by anybody cuz we wanna, you know, we want like a normal marketplace working here. So where do you see this going? What's the evolutionary path?
Stephano Cetola (00:51:58):
No, I think so. You asked me a similar question last time about, you know, where I saw risk life going. And I think my answer back then was, you know, we're trying to define how does one do open source when it comes to an instructions set architecture because that problem is still to be solved and we are in the process of solving it. I think that still stands. I think that is still a problem we're solving. And you know, in our day to day at risk five international at the Linux foundation, those are the problems we're addressing of, you know, as we layer on top of risk five, you know, each layer requires some oversight. How do we do that in an open source manner? And what does that mean? Cuz you can't just copy and paste what we did for Linux. It's not directly relatable.
Stephano Cetola (00:52:41):
There are different problems to be solved and different concerns that need to be addressed. So that's still happening, but I think, you know, you, you touched on modularity what I think the modularity and what the flexibility of risk five is gonna get is first it it's gonna encourage innovation. It's gonna allow folks to go off and create custom extensions where in the past they weren't able to. And that's going to start to enable aspects of computing that we haven't thought about yet. So that's, that's sort of piece number one piece. Number two, I think is the cost savings. That's sort of inherent in all of that. That isn't really there yet, but that I think will be at the heart of what risk five gets us because you're no longer it's, it's not just the licensing cost, right? It's not just that, but it's the R and D costs behind spinning something up.
Stephano Cetola (00:53:32):
So if you think about, you know, let's say I, as a human being in the eighties wanted to create an operating system, what would be my chances for success and now let's take it forward to today. And I want to make my own operating system based on UTU. What are my chances for success? I think that you're starting to see the, the cost behind doing something like that came down hugely because of the community effort around rising, rising the bar for everybody, everybody gets to sort of sit on the back of Debbie. And so we all get to go a lot further. I think the same thing is gonna happen in, in hardware. You're gonna start to see that that rising tide happen and that's gonna drive cost down because suddenly you are not trying to spin something from scratch. It's not your concern. What the op code looks like.
Stephano Cetola (00:54:20):
They've already figured out the op codes. They've already figured out how you can layer your custom extension. On top of the base processor, all you have to do as a company is concentrate on your custom extension. And so even though that veers a little away from the standardization, I talked about what it does veer toward is that innovation, the ability to do something completely off the beaten trail that no one else is interested in because they don't wanna standardize around it. But if you can hit on something interesting, something different, that'll change the industry. And I think that's, that's really what we have to look forward to. And yeah, I'm excited to be part of that.
Jonathan Bennett (00:54:59):
All right. So I I've, I've gotta jump in with a totally different question and I've kind of got a reputation on floss of being the one that acts the really technical questions and I'm going to continue to live up to that. I think RET bleed, which is one of these speculative execution problems it's the idea that in Intel and AMD and even arm processors, you can manipulate the processor to where it will speculatively execute out in a direction and then, you know, execution catches up with it. It rolls it back so that it does, it does the right thing, but it, it changes the way essentially the cash, it changes what's sitting in cash. And then if you're really clever about it, you can figure out those changes and read things from memory that you're not supposed to be able to read. That's that's the 32nd explanation of <laugh> of specul execution problems.
Stephano Cetola (00:55:53):
That was a good high level right there.
Jonathan Bennett (00:55:54):
<Laugh> there you go. Is this a problem in in risk five, it did risk five get hit by re bleed or any of these others, or is it just architectured in such a way that it's not an issue?
Stephano Cetola (00:56:06):
So I'll say no with a caveat <laugh> it's not an issue, but with a caveat. So one of the great things, so I was actually at Intel when spectrum and meltdown happened. So this is a subject that I've listened to and contributed in many conversations around risk five has the sort of the benefit of hindsight. So when risk five was designed they already had this concept of, you know, speculative execution and folks were already trying to poke holes in it. That was common. So that's kind of the first piece that that gives risk five, a little bit of an edge. It's it's the second mover principle, the idea that, you know, stuff has already come before us, we cannot make the same mistakes. The other key behind risk five, that sort of that gives us an edge here.
Stephano Cetola (00:56:56):
Risk five, especially risk five international, but as a group, what we're doing is not implementation. We don't do we don't build CAEs. We build, you know, simple implementation CAEs that are examples, but we don't fully flush out CAEs. What we do is we create interfaces that allow implementers to go off and create those CAEs. So what we can do is provide them with guidance, given what we know about the history of speculative execution, such that they can go off and build products that are less likely to have issues. Because if you think about it, as soon as you start building products, you're building bugs. The, the second you start producing something, whether it be a hardware product or a software product you're engaged in building bugs. So all you can really do is set folks up for success and then take the feedback that you get and implement that in your interfaces. And that's what we're doing. We're making flexible interfaces so that folks can go off and build their CAEs and then keeping those interfaces up to date such that the next generation of risk five will be able to, you know, take into account anything we've learned in the intern. So, so the answer is no, but with those caveats, I'm sure you can go out there and find an implementation that is probably substantial to this. However, moving forward, we've already integrated in these concepts into how do we keep it from happening again in the future.
Jonathan Bennett (00:58:20):
So it's, it's actually going to be a kind of a very different approach that, that things like the kernel is gonna have to take to it because in the kernel you can simply say, all right, anything from Intel from this generation to this generation, we know is going to have the red bleed problem, anything after it is not. Whereas with risk five based architecture, it's going to have to be essentially product by product
Stephano Cetola (00:58:44):
It, well, we are gonna align around two concepts, profiles and platforms. We don't have time to talk about that in three minutes, but the concept of being, you can think about things like R V eight as being our profile. So we will have a sort of line in the stand that folks can go, okay, if you comply with, you know, RV 22 RV 22, a has these features, and maybe one of those features might be vulnerable against some attack. So we will have ways to draw a line in the stands. We wanna make software people's lives as easy as possible, but we also wanna make it easier for innovators and implementers to create unique products. So, so it will be a balance somewhere between those two areas.
Doc Searls (00:59:26):
So this has to be almost a record setter for how fast an hour can go. <Laugh>, I, I think the number of seconds and minutes in it is no different, but the amount of information it contained is enormous, and this has been great. I, we always close with a, a couple questions. <Laugh> somebody's, <laugh> there's the back channel <laugh>
Stephano Cetola (00:59:51):
I, I knew I was gonna hear about NFT is when I, when I saw that on the laptop page, I figured I was gonna be in interviews for several years where that still came off
Doc Searls (00:59:59):
<Laugh> yeah. It, it, it happens.
Stephano Cetola (01:00:02):
I appreciate them keeping my life interesting. I don't want things to be boring.
Doc Searls (01:00:05):
<Laugh> so, so so is there anything that we, we haven't asked that we should have asked that you'd love to answer in any case?
Stephano Cetola (01:00:15):
Yeah, I think so. What, the question I get asked the most from people are just, you know, when am I gonna be able to buy the next risk five insert thing here? We have a program called the developer board program that I highly recommend your audience. Go take a look at. So as hardware becomes available, risk five international is gonna make that hardware available to engineers to do interesting projects. It's a fund that we've gotten from the board of directors down. And the goal is to encourage folks to work on risk five hardware. Cause we know it's hard to get a hold of. So I would highly encourage folks to go take a look at that. We do have a limited supply boards, like whenever we purchase, we are only allowed to purchase so many. So we have to be kind of choosy about, you know, which projects we we give the green light to, but it's a great way for someone to get involved in risk five, get hardware on their desk and start playing around with it. I think the, the feedback loop too to the community is really valuable because the work that's done gets posted publicly. We require that it be open source. We require that you give us feedback. And so we're actually seating both engineers who want to do work and the community who wanna see work being done and be able to build off that work. So it's a really exciting program. I would recommend folks go out and take a look at it.
Doc Searls (01:01:32):
So for the last two are standard questions. What are your favorite text editor in scripting language? I know you answered it before, but maybe that's finished. I dunno.
Stephano Cetola (01:01:41):
No, it hasn't. It's still been <laugh> them is UN unfortunately my gift from being an engineer in a former life. So it's still my texted of choice. And I do have to, I think Python was my answer before, and I'd have to stick with Python if I'm still <laugh> choosing a
Doc Searls (01:01:56):
Stephano Cetola (01:01:57):
No, I have to say that those skills are getting dustier and dustier. Every time you ask me that question. So <laugh>,
Doc Searls (01:02:03):
Well, I, I expect that given that the, the pace of things happening with risk five we'll have you back, you know, we always say, we'll have to, like, you're living proof that we do that. We actually have people back <laugh>,
Stephano Cetola (01:02:16):
It's great to be
Doc Searls (01:02:16):
Here. So thanks so much for being on. Absolutely. So Jonathan <laugh> lot of, lot, lot of stuff on the show.
Jonathan Bennett (01:02:29):
Yes. I, it sounds like I did not successfully talk myself into free laptop, which, you know, I'm kind of bummed about <laugh> that works every once in a while. Well,
Doc Searls (01:02:38):
Maybe one of our listeners is is privy of some of that. And there
Jonathan Bennett (01:02:42):
You go, you
Doc Searls (01:02:43):
Know, somebody with one of the, one of the OEMs out there working on this can help you out.
Jonathan Bennett (01:02:49):
Yeah, no, it's, it's pretty cool. I, I've not actually, I've not gotten to get my hands on any risk five hardware yet, so I need to pick up one of these developer boards just to be able to play with it and say that I have. But yeah, it's, it's a, it's a very, it's a very cool idea. And I'm excited that they're doing some of these different things like the, the vector extensions. So you could, you could feasibly put together a GPU using risk five, which then leads you to the idea of an open source GPU, which is something it's been the dream for, you know, certain people for the longest time. They're just kind of marching forward with all these things that you would expect and require to have from a modern CPU. And you know, you already see risk five, you, you already see risk five in things it's just not in places where you actually see it.
Jonathan Bennett (01:03:37):
So like there are some hard drive manufacturers that put risk five in their stuff. Intel uses risk five internally in some places in your CPUs. It's just, none of it are in places that you, you have any visibility. I guess I'm kind of looking forward to the first time that somebody shipped something that is obviously running a risk five CPU in it, you know, whether that be a, a, a phone or, you know, the laptop, but that that's a limited run and, and kind of a unicorn. You know, I mean, one of these days HP is gonna make their HP dev one and put a risk risk chip in it instead of the AMD. And, and I, I kind of think that that's when risk five is really gonna make the jump from a very niche. Certain group of engineers know about it too, maybe a little, a little wider public knowledge about it, and that that'll be fun.
Doc Searls (01:04:28):
Well, it, the one of the things that intrigued me is that when he said about, you know, what succeeds in the long run maybe is gonna be something does not in anywhere else that is too tempting. I don't think you said this exactly, but that was by take on it that, you know, there'll be, so maybe it's this vector thing and and, and make, roll your own GPU. That would be very cool. And there's, you know, I expect it'll succeed. I love the energy in it. And I think that the energy says a lot and the fact that our, there are in fact big companies in it that while they don't have their figures on the scale exactly of the development of it will, are interested enough in it to make something happen. So I'm encouraged about that.
Jonathan Bennett (01:05:10):
Yeah, it's, it's always fun. So we are all, all the time, just throwing things at the wall and trying to see what sticks. And it's always neat to see something kind of radically new start to work and start to really gain momentum. And I feel like that's where we're at with risk five. It's something totally different from anything we have in the CPU CPU slash GPU world right now. And it's, it's getting to the point of getting critical mass and getting some momentum. And so I think, I think we could be of the risk five guys could be in for a really interesting say next five years could be quite the ride.
Doc Searls (01:05:50):
Yeah. Yeah. And so we'll, we'll stay on that horse along with, along with the, with Stephano I have to let you know next week really interesting show maxi Reynolds of subsea cloud is gonna be on, this is, this is cloud running under the ocean. Like the fiber you're probably watching right now also does, but putting like a data center under the water is a really interesting thought. So that's coming up next week. Max Reynolds in the meantime, what do you got to plug there, Jonathan?
Jonathan Bennett (01:06:28):
All right. Hacka hackaday.com, make sure every Friday morning catch my security column bear and the occasional review and coverage of other things from me. And then you already mentioned it, but get on club TWIT and go subscribe to the untitled Linux show. And we have a lot of fun there and make sure you can come join us.
Doc Searls (01:06:47):
Yeah. And, and as is my custom, I did things a little bit out of order, but that's cool. Next week again, maxi Reynolds and subsea cloud. I'm Doc Searls. We will see you then
Speaker 5 (01:06:58):
Did you spend a lot of money on your brand new smartphone? And then you look at the pictures on Facebook and Instagram, and you're like, what in the world happened to that photo? Yes, you have. I know it happens to all of us. Well, you need to check out my show hands on photography, where I'm going to walk you through simple tips and tricks that are gonna help make you get the most out of your smartphone camera or your DSLR or mirrorless, whatever you have. And those shots are gonna look so much better. I promise you, so make sure you're tuning into twit TV slash hop for hands on photography to find out more.