 There's the unrelenting treadmill of Moore's Law. This diagram here, there's two curves. One is in blue, and the other one is in yellow, and there's that little merger that kind of looks in odd color because of Gidoo plot. And the blue line is sort of the goodness that someone working alone can impart to a system by, say, optimizing something with hand assembly or sort of improving the system architecture a little bit, and the yellow line is what Moore's Law ends up doing, right? So tomorrow you can take your system, write some assembly code, push a patch for it, and it runs a little bit faster, right? But then a year and a half later, someone will just throw away all of your hand optimization, recompile for this platform a little round twice as fast. And so to some extent, hardware's biggest problem has been that just sitting around has been more effective than innovating. All these people who have built these massive parallel machines on the left-hand side are now completely trunked by an innovative videographics card today, for example. So this too is coming to an end. No exponentials are forever, right? And we have seen this in the clock speed, clock scaling. It sort of started in the 70s, and I would say right around the turn of millennia, clock speed just stopped getting faster, right? Kind of around a two, two and a half gigahertz mark or something like that. I mean, you can buy extreme systems and overclock stuff and push it to four gig, but if you looked at the roadmap originally from like the mid-90s, from Intel, we're supposed to be on like 15, 20 gigahertz right now, we're actually on target from Moore's Law. So that went away very quietly, kind of until we did all the numbering schemes to be less megahertz and generation centric to more about cores, right? There are lots and lots of cores and most of them sit around idle most of the time. And so currently, Moore's Law right now, this is a really interesting slide that was released by Nvidia showing the wafer prices over time. So each of these lines here, so the blue line is the 80-nanometer node, the red line is 55-nanometer, the green line is 40-nanometer. So each of these are different process nodes and those lines represent the falling cost of those wafers over time. And every line represents sort of a 30% to 50% improvement in density. So there's a point at which, there's a prosperity point which you would just jump from one node to another. And you can see that these nodes are getting ever more and more spaced apart. And what you find actually is that, in fact there's almost no intersection point for the 20-nanometer node and the 80-nanometer node and so forth. So if you look at this chart, basically it shows that yes, we can make ICs in ever more aggressive processes, but at the end of the day, you're not gonna pay for them because there's no cost for a point in price. So what does this mean? It means that quite soon these statements will become more true than false. Next year, you cannot buy a faster computer. The computer you have here yesterday isn't the fastest computer you ever go to home. Next year, you can't buy a flash drive that stores more data. The flash drive in your pocket will be just as big as the flash drive tomorrow. And next year, your phone won't be smaller or more powerful. Your phone that you have today is about as big as the phone that you're ever gonna need to use. I say this is actually really good news, particularly for people who are in costs and car ranges, like me, or small individuals. If you look at that graph of Moore's Law that was showing me for the yellow line is still there, but just, you don't say Moore's Law isn't necessarily dead, but it's just slowing down a little bit. So the yellow line is Moore's Law plotted doubling every 18 months. The green line is Moore's Law doubling every 24 months. And sort of that sort of pink line on the bottom, which now turns buff as the overlay is 36 months, right? We take the graph, we go ahead and plot it on a log scale, so we can kind of see the linear part being a little more exaggerated. What you find is that if Moore's Law slows down to a doubling of once every three years, instead of once every year and a half, someone hand-optimizing machine code and constantly offering incremental improvement on a single process note has almost six years' opportunity extra for a product cycle than they had before, right? So before you had to be super fast and super on the ball and you had to essentially be employed full-time by a big company to make improvements to products that matter. Now you can go ahead every weekend and just keep on hacking at something. And several years later, it's still actually relevant. It's actually something you can make an impact on as an individual contributor. So here's some of the implications for this. I mean, there's many, many, many implications, but I only got 20 minutes left to talk about them. But one of them is that ARM is rising now. This was something that was originally used to control toasters and DVD players and, you know, they'd be in little, tiny boxes. Now it's incredibly ubiquitous. Everyone's got an arm, several arms of them, probably right now, I mean, two, but you have like many arms in your phone and there's an arm core and it's basically, everything's got an arm core in it. And it's actually becoming a serious contender to the X86. The Cortex A15 implementations are pushing well over two-gear hertz, or fours, or 64-bit starting servers, and so forth. So there's this new architecture that's coming out that has been sort of propelled forward and it's actually sort of becoming equivalent, performance equivalent to the X86. Another implication of this is that repair culture is becoming more common. It used to be the case that if something broke, you wouldn't fix it, you'd just throw it away and buy the next thing. There's like, it doesn't even make sense to fix your old Nokia phone or whatever it is you buy the next generation on a Blackberry or iPhone or whatever type of person you were at the time. Now that things are kind of slowing down, it actually seems like maybe if I replaced the screen on my phone that it broke, it would be perfectly fine. Like the new phone isn't as compelling. I don't want to pay $1,000 for the new phone, I'll just pay $100 to fix the screen on my phone. So broken gadgets are now actually having more recycling value. And this also means that reverse engineering the current gadget you have has more value. For a while, if you would spend a couple months trying to work out the connector pinouts and build a adapter or whatever it is, you're like, okay, well that was kind of not worth it because six months later, the new Grace thing is out and they had a new connector for it. Now you can take a couple years to go and ping around and figure out what's going on inside these boxes and a couple years later, people might actually still be using that phone. So these reverse engineering is actually starting to get more value these days, at least in the hardware front. And so one place where you sort of see an evidence, a really strong evidence of this repair culture rising is in China. China, the emerging markets are a couple years behind on the tech curve. And so in those markets, what you find is that yesterday's phones becomes today's parts. This is sort of a picture, very common see that you see the people taking these phones and pulling out parts and recycling and reselling them because for those guys, they're actually quite happy to have tech that's a couple years old or that's kind of stabilized and so forth. And what this has led to is this sort of rise of an information ecosystem based on this sort of older technology. This technology had time to sort of marinate as well, people reverse engineer it and sort of publish now books of schematics on those Nokia phones and they share source on the hardware designs and the CAD and they actually have whole books about how these transceivers have written that work and all done in Chinese. And the result is that people are now able to take these ideas and incorporate them as this sort of backkeep ideas in their own zone. The left here is like this mouse that looks like an Apple mouse but actually has a phone on the inside. Inside the Apple icon is actually a camera and so on and so forth and the thing actually works as a mouse too. It's kind of fun and the right hand side is a phone that has only like five buttons on it that's meant for like kids, right? So you have a kid, you don't want them texting in class but you want them to call home. They just went ahead and wedged a phone into this sort of wacky form factor. And so it actually is a little plug for a workshop that Zobz had been giving later on. Tomorrow on Saturday, we're actually going to dive into one of these things and sort of show you I can hack one of them. We actually have reverse engineered the chip set and made all the hardware available for online for you guys to access. Tomorrow afternoon, I think at one o'clock we're gonna be... But at the end of the day, what this is is it shows us an opportunity for small innovators. The people who do this in China are called Shantai and they sort of demonstrate that you can build phones with very low capital investment in very small teams and not big corporations doing these things. There's small people, so wake up when, you know, I really want a phone that has a cigarette lighter in it. They go ahead and they kind of like draw it out and a small team of people, a couple months later, pops out a phone and has a cigarette lighter on the inside because they want it, right? It's not like they have to go through this huge corporate process and get lots of money to do it. And so, you know, there's sort of stable platforms and open ecosystems help do this, but these kinds of products don't happen overnight. One of the core things is that you find in open sources that open products take time. The red one stuff that comes up, the stuff you push to get up in the beginning is just kind of almost a sketch, right? And so, like for example, if our Fox started in 2002, it didn't really become usable until many, many years later. Linux was 1981, it took, I mean, I remember looking at Linux 0.9 back in the day and it was just not really, it was more like the something we did as a geek project to sort of show that we're just playing around operating systems and kernels, but there was nothing for it at the time, it took years and years. And the Libre office, you know, started a long time ago and only in the last few years did it really, in my opinion, become really usable. So one of the things is that with technology stabilizing, right, and small disruptive teams and time for organic growth to happen as long as it repaired DIY culture, do we finally have an opportunity for open hardware impact opportunity? And so sort of going by this equation, going by this theory, me and actually Sean, Sean's in the audience right here, you wanna stand up and say hi to everybody? That's what Sean looks like, he's a other guy. I've been working with me for the past few years here in Singapore to do different random and crazy projects that we've done. We decided we would actually try to build our own laptop because all the factors have conspired to make it sort of the right time to do this project. And so we built Movinna, it's open hardware. So you can go ahead and download the schematics, the board layout, the 3D files for the system. It's open firmware, that's you can go to ZOB's GitHub repo, XOBS, and so you can pull out all the code required to build your Movinna kernel as well as things like our test programs and other fun frameworks and infrastructures that you may not actually often see associated with hardware products. And then the hardware itself is designed for people to mod and hack. So it's actually literally open hardware. Like the lid opens with a button and you can actually see the hardware on the inside. We built into it this sort of array of mounting bosses I call the peak array on the right hand side. So you can go ahead and just like, you know, screw in, I don't know if you want an Arduino board on the inside or Raspberry Pi inside of your laptop. If you want to go ahead and add some little wacky, you know, thing on top of it, there's an area for you to go ahead and hack that. And also the side panel that covers the ports on the right hand side of the replaceables with the idea that if you want to upgrade the motherboard, you don't have to throw in the whole case. You just replace this one panel on the side. In fact, if you want to just, I don't know, just use your case with the Raspberry Pi inside, that's also fine. You can go ahead and just throw in the motherboard, drill out this blank panel that we shipped with it, put the Raspberry Pi in there and then look up to LCD if you can. And then you have, you know, the Raspberry Pi in the case, that's totally fine, right? So it's meant to be open and hackable. We're also building a model called the heirloom model where there's this guy named Kurt Mottweiler in Portland, Oregon, United States, who's actually building these laptops out of wood. He's made this really cool wood composite that has a really nice feel to it and has the machine aluminum and all these bits and pieces to it. It looks a lot more conventional. You know, why are we doing an heirloom? The idea is that, you know, as hardware slows down and as repair culture becomes more common and people want to keep these things, it actually makes sense to build a laptop with exclusive materials and a lot of craftsmanship. The attention is that, you know, the case will be upgraded and used for years to come. It's not, this is not a static thing. It's like later on you can go ahead and pop another one over the inside. You still have this wonderful loom to do your work on. You can hand it down to people later on. And one of the cool things is actually building this laptop is actually very much enabled by the fact that we are open. So working with Kurt out there is super easy. We just sort of say, here's a website, here's our code, here's our link to the hardware models and so forth. We have a very easy to discourse because there's no NDAs, there's no contracts and crazy stuff around trying to get this all built. So looking back is, you know, how is it novena possible? Again, this is that slide I showed with the different costs of waiters over time. And on the bottom there's like the little, the x-axis is the sort of time plot up and quarter. So it starts in the first quarter, 2006, 2007, blah blah blah. So that little dashed green line represents the development trajectory of novena overlaid on this graph. So we started some time around early 2012, right? And I'm at six, which is the CT that we're using. So RMCP was found in 49 in the early years, about like six months or a year before that. And it took us a couple of years to do it. But as you can see through the conception, launch and delivery, the actual cost of that wafer and that node has actually been almost flat. There's been very little change during that node. If you look as a retrospective, if we were to actually try and shift that exact same green bar, even just like two or three years earlier, you can see that that same period of time during the peak of Moore's Law, which is just a few years ago, we would have gone through three process node generations. By the time we delivered, we would have been on a silicon process that is three nodes too old, no longer really interesting. We would have delivered an 80 nanometer system when 40 nanometers was hot. And so you would have had a system that was about half the speed of what you could buy today if you took your two years to develop it, right? And so you can see actually moving forward or look at this graph moving forward, there's actually lots of time for guys who are doing open hard work to go ahead and really learn a system, figure out the kinks of their EDA tool, go through a few revisions of the motherboard, go through all the time of building your supply chain and so forth because now Moore's Law has actually kind of reached this point where we've kind of not really grown to a halt but slowed down a bit. So where to go from here? Well, you know, open hard works are all about building communities around platforms, so please take our IP and visit our website, kosage.com. You can see all of our products listed there and you're more than welcome to use it, hack with it and if possible, contribute back to the community. So just as a recap, the experiment that we engaged in and also started my thesis is that as technology has come to stabilize, we can find small disruptive teams around the world who can have the time to organically grow communities and you combine this with sort of a rise and repair DIY culture and this results in an opportunity for an open hardware impact. So thanks for your attention.