 So here the Lenaro Connect and you the CTO? Yeah, I am lucky enough to have the best job in the world. Hello again, Charvex. And so there's been five years now, lots of things happening with the Neros? Yeah, it doesn't seem like five years. They gave everyone who's been at this five years a nice award, so I've got to really, I just put it away otherwise I'd show you, but it's just really heavy bricks, and my bag is a lot heavier on the way home, but it's been a wild five years. But there's still a lot of things to do with this. Oh, yeah. And the thing I've been looking at recently, because I get to look at Newster, is embedded in IoT. So IoT is causing a huge amount of disruption in the industry. It'll cause more. Giving everything the ability to connect to the web or to a gateway to the web means that security of the cloud, it opens up a whole bunch of possibilities. I have a Nest, I have Philips Hue lights. That's just the start of it, but the potential for all of your devices being connected and talking to each other and exchanging data and being linked in some way to services that you want, it's just amazing. So it's taking the embedded world apart and will cause lots of stuff to happen. It also coincides with this whole maker thing where people are taking devices, building new products, new types of products. Very quickly, the whole Kickstarter kind of hardware, small production run hardware, as you can prototype, you get products out so fast. So there's a whole exciting vibe going on around them. 96 boards is a good part of that. It is. I mean, that's kind of was solving, well, it has many, many aims. One of it is to solve the hardware problem, get more hardware available into our hands, into the community's hands, particularly ARM 64-bit based hardware. But that just wakes up a whole bunch of communities who can then take these and do wonderful things with them. So all these guys around here, are they going to be able to contribute in IoT? And what are they going to be able to do with embedded in IoT? Well, there's two ways of contributing, right? There are engineers and developer communities and what needs to be done. And there's also the NRO, the organisation. The way it collaborates is to get a bunch of companies together who care and do some joint engineering work that enables the whole process for ARM in that ecosystem. So there's a bunch of stuff that we'll be doing as we form a group around the A-class cores, the Linux kernel being able to be the IoT gateway or client or whatever. There's a bunch of stuff about integrating with the cloud. And there's loads of standards there that are just, you know, just every day I see a new consortium start and companies announce something. So it's kind of an explosion. So as an organisation we're forming a group, the role of that group, the companies in that group, is to define the standards that we want to follow and support, figure out the core engineering work and collaborate together to help enable that space. So there needs to be an open, kind of like a collaborative platform in there to make it a really huge success, the IoT. There's going to be billions of devices. Yeah, billions, trillions, who knows? And Linux is a good platform for that. And a lot of this stuff is about polishing Linux and making it more applicable. As an example, back in 2.6, Linux was smaller. It was easy to put in a smaller device, but, you know, it had more features added over time. And so now, even if you build a small kernel, it's not that small. So Nicholas Petra, who works in Lunara, has been looking at tinnification where you take the system calls out of it in an automatic way. If you don't want them all, you pull a whole load of code out of the kernel and it shrinks it down a bit and shrinks the data footprint. So there's engineering work like that that needs to be done. So it's making Linux smaller and less everywhere? Yeah, smaller and more applicable to the embedded space. A lot of the protocol work to support the various IoT sort of RESTful protocols is done, if you look. All of those standards are kind of in there. But it's really about looking at enabling the system. So things like compilers, right? So traditionally, in the embedded space, you don't have an MMU. GCC is still highly applicable, obviously. And for ARM, it's some two instructions there. So, you put, you know, there are some options around linking into a flat space. FDPic is one thing where basically you allow dynamic linking into a flat space. So let's enable that so that ARM-based systems can do that in the MMU-less case. Let's think about how Linux, the kernel, has also got a whole bunch of services and code and drivers that's really good. Can those be used in an MMU-less system but without too large a footprint? And then there's the system end-to-end aspects of this. So there's going to be lots of Cortex-A CPUs in the IoT. There's also going to be lots of Cortex-M. Are you going to work with the embed stuff? Yeah, absolutely. What they're doing with embed is enabling the embed space. So they're embed OS stuff where they move basically from a set of services to an event-driven OS, inter-driven OS. I mean, the M series is basically an event wheel. And so they're producing software for that and libraries as well that integrate that device with the cloud. Those libraries run on Linux. Some of that, they release a Java-based version, etc. So the members want to support that effort from ARM in the M space. There are also a lot of embedded RTOSs. And again, the compiler work would probably help them. The other thing about the embed space is incredibly wide. It goes all the way from, I call it, light bulbs to motor homes. If you imagine a motor home as an instead of things thing, which you can be. And so this group is a little bit different from the one with the other groups. We want it to be able to grow. We want it to be able to go, OK, we need to do this in automotive. So let's pull together that part. We need to do this in, I don't know, cameras. You're pointing a camera at me. A lot of cameras are based around ARM. They've got accelerators for face recognition, doing advanced thing with images. They're all using Linux. They've all got different frameworks for acceleration, how they interact, maybe ODP plays there, don't know. So the idea is to form a little special interest group and look at that problem. So really, the embedded group will be all about many different market segments as one way of putting it and common technologies across them. And just trying to ensure that all the bits are there so that companies can innovate and make great products, right? So Linarra can make a big contribution to the future. I believe so. I mean, it's a bit of an unsung hero thing. Some of the stuff we do is deeply unglamorous, like UEFI. I mean, no one buys a product because UEFI is the boot agent, right? They buy it because it's secure, because you've got the built-in secure trusted firmware and you've got the opti-e work. You don't buy it for UEFI, but it enables people. So it enables product. So that's the main thing that we do. I have no idea how many Cortex A's and how many Cortex M's are going to be part of this future of the IoT. I've seen various numbers in various places. I won't try to remember them or quote them because I'll get them wrong. But the ARM architecture is playing really, really well here. It's a very good architecture. And both the A and the M and the R are in between for the more real-time. They're very good solutions into the appropriate bits of the market. Of course, they run the same instruction sets. So I think you'll see huge amounts of activity. And so you're talking with all the guys around here and you're basically asking who's to be part of this? Yes, no, exactly that. If they're interested, they can just start trying to ask if they can be part of it, right? Well, get their companies to ask if they can be part of it. Yeah, working through the technical steering committee and the NAROs board, I have been proposing this. As part of it, I've been having these one-to-one conversations with various companies. And they're really enthusiastic. And I'm going to pull in my timeline a little bit because we normally form a steering committee and then we do an announcement. I think I'm going to pull that forwards. And I'm getting a lot of companies say, yeah, yeah, we need to do something. We need to work together and please the NARO can we join in. And suggesting what we do and getting involved is driven by the members. It's getting them together to a point where they can collaborate. And they see it working in other segments and they recognise the need in here and they see NARO as a good place to do that. It's a fun segment to play with. Well, from a playing point of view, I mean, Grant likely has been building little bits of hardware and running around with sensors. He's got an Arduino-based board that sits on 96 boards and it's got all these controllers and actuators and thermostats. I'm going to use one to automate my greenhouse irrigation system. You can measure the temperature at the water level, you can turn the tap on and get more in, more water in the system or open the windows. So I'm looking forward to playing with proper toys here. And it's going to be awesome to have your work in the 50 billion devices per year. Well, that blows your mind, isn't it? In terms of how many places. I don't write a lot of software that actually gets anywhere these days but I'm really proud of the software that's written in and around NARO that isn't. You're right. If you think about just about every mobile device it's got some bits of work that NARO did three years ago. It's worked through the kernel so yeah, it's really good to see.