 Computers keep changing the world, but their power and safety is limited by their rigid design. The T2TILE project works for bigger and safer computing using living systems principles. Follow our progress here on T Tuesday Updates. This is the 52nd T Tuesday Update. It's the last T Tuesday Update of a year worth of doing this, which is unbelievable. For the last couple of weeks, I've supposed been working on getting the locks and the cash updates done so we could do Real Inner Tile, which is sort of the last big hardware-software issue before being able to benchmark these things, and I've been completely clutched up. I've been procrastinating in every possible direction. This week I spent a bunch of time reading a book by Bunny Huang. I'm going to talk about that a little bit. I did work some on the mechanical form factor stuff, the power zone frames and so on. I'll talk about that, but exactly how I'm going to actually get myself to do the software redesign to get the cash update working into Tile. Who knows? Next week I want to do a recap of the entire first year sort of for the record and we'll see what kind of plans we can think of going forward. So we're ending season one, year one with more of a whimper than a bang. So be it. All right. So yeah, Andrew Bunny Wang is this famous hardware hacker guy, wrote hacking the Xbox when he was a grad student at MIT. I mentioned it earlier in this series when I did the cardboard prototype for the light sensor that is now on all of the T2 tiles. And I got this book of his which this week, The Hardware Hacker Adventures in Making and Breaking Hardware. Now I had managed to convince myself when I saw this come by in my timeline, whatever it was, that it was a relatively recent book. But then as I started reading it, it was all kind of older stuff and so it turned out actually that, you know, all right, August 27, 2019, that seems pretty fresh. But if you go to a different page in Amazon, you get no starch press first edition, March 15, 2017, and that's closer to what actually is the case. You know, it's the exact same ISBN, the exact same everything, August 27, 2019. The best I can guess is that Penguin, blah, blah, blah, the Kindle edition may have come out in just recently, but the book itself is two and a half years old. Anyway, you know, I wanted to get the book just because, you know, I'm a fan of Bunny Wong and I wanted to support him, although it does turn out a little bit that in this case there is some arguments to be in favor of getting the e-book instead. In particular, this is a page from the free sample of the Kindle edition on Amazon. This is a corresponding picture from my physical copy. The pictures in this book, I mean, it's matte paper. You know, if they'd wanted to get high quality res pictures, it would have cost a lot more and they would have had to use glossy, at least on the photo pages. And, you know, it just, it doesn't look great. And one of the other things that Bunny does, which I used to follow along a lot and I fell away from a little bit, is name that where, where he would show a picture each month of some circuit board and everybody, all the electrical engineers and fans would, you know, use their skills, reading the tops of circuits and try to just using their knowledge of electrical engineering to figure out exactly what it was. And each of these would be these incredibly high resolution, you know, luscious, you know, tons of detail sort of things, which is what you'd expect to be able to figure out what's going on, although he does strategically mask stuff that might make the puzzle too simple. So, oh yeah. And, you know, one of the things he does in one of the later chapters is sort of compare the original Apple II, I think. I'm schematic with the, the gene regulatory networks of Michael Plasma, Numuni, an actual living system, one of the simplest things, and, you know, and says, you know, that, you know, the pentagonal boxes in the diagram are enzymes, you know, and the thing is, you know, trying to read the pentag, find the pentagonal boxes in that picture, not that easy. But this particular chapter where this is from, chapter 10 in the book is great. You know, a lot of people talk about, you know, using computers as metaphors to understand biological systems and using computer technology, you know, informatics to study computers, to study biological systems and so forth. But Bunny does what I absolutely prefer, which is use computational systems as metaphors to understand specific biological stuff. And this is great, the description of how to understand, you know, the whole enzyme binding and stuff, you know, that's okay, that little lock and key, there's a lot of ways to understand that. But he does a great rundown of it, and in particular, the discussion of CRISPR-Cas9, which is this latest gene editing stuff, is great. And the fun thing about it is, you know, you know, as a hardware hacker, he ends up doing reverse engineering, where you're studying what a system is doing by, you know, sticking inputs in and seeing how it reacts and so forth, which is exactly how biologists, microbiologists study these things by sticking stuff in and seeing how it reacts and, you know, and break knocking something out and seeing what it changes and so forth. It's extremely clear, it's a wonderful description. So I like that a lot. Now, one of the other things that I had to read was the story of Chumbie, which was this crazy product from 2007-2008 that, in fact, Buddy, you know, sort of learned how to do electronics manufacturing in China and logistics and all that stuff, doing the Chumbie. The Chumbie was never actually successful as a product itself, but I actually bought one. And, you know, so Chumbie was a truly open platform, enabled hardware to tinker. Oh, yeah, and he's got a footnote. He was just a lead hardware guy, although he did the Linux kernel stuff too, knew from the time learning insides of Linux from boot to halt, you know, I feel like I'm following in Bunny's footsteps 10 years later. Of course, I have to deal with system D. Hacker-friendly platform, open and so forth. Now, you know, I had a different experience a little bit coming at it as a software person. I had read about all this openness and so forth, and you know, here's the picture of the original thing. It was like wrapped in leather. It was really kind of cool. You know, it's not exactly clear what you're going to do with it, but the idea was that software people, people like me, would write little widgets that would plug into this thing and provide information services so that Chumbie could remind you whatever good stuff you wanted in the morning and so on and so forth. My problem with it was that despite all of its openness, you know, fine print didn't seem actually open to me. So, oh yeah, he's talking about his total nerd stuff and that yeah, this is great. C-Tech astronomy, do you know this trick? And that's a movie reference that they used on the main board and this is another example where, you know, the pictures in the book don't really do it justice, but I went and found the picture on the internet. You know, there it is. So, it's a way to bypass the authorization just explicitly allowing you to do it. This was an example of how they were being very open even when they weren't. What got me was that there were some fine points in the license and so I read this and in fact, yes, he discusses it in the book. The license had a couple of other restrictions that were not truly open as if it wasn't a big deal, but in fact, that's what gave me pause as I was working through and saying, am I going to sort of make this one of my things and try to spend a lot of time doing it? And, you know, in the end I didn't. Burying that clause in the license meant it couldn't be called open source. Yeah, it couldn't be called open source, so Chumbie could say no, though it never did. Yeah, well, I wasn't going to take that risk. I actually returned my Chumbie. Oh well. You know, again, navigating all these things, but you know, people who investors were putting up a lot of money and so forth, it's one of the reasons why, you know, I try to make lemonade about the fact that the budget for the T2Tile project is like, you know, double digits or something like that. It's not actually that bad, but it's nothing like what goes on to do actual stuff. And actually, one of the things that's really great about this book is that it has all these little side stories about what manufacturing is like and how the big boys do it and how little boys can do it and so forth. I like this line. Merchants, this is about how pricing stuff for retail. Merchants tend to look at your product as so many grams of plastic and so many wires, and then they multiply that by how much plastic should cost, how much wires should cost, and then they say, okay, that's the base price for this thing. How much more than that do I actually have to pay you to get this on my shelves? Like that. And you know, that, I like that. That's sort of mips per milligram. How much computation can we get by the pound? That's the way we want to think about it in the future. We're certainly not there yet. Apple spends over a billion dollars a year on tooling, just, you know, molds to make things. Injection molding costs $40,000 and takes months. Apple is known to build five or six simultaneously and throw all but one of them away. Wow. Tossing to $100,000 in tooling to save too much time to market is peanuts to say that this is the power of scale. So the book is full of those sort of things. If you like hardware stuff, you know, it's fine. It's a good read. You know, he's got some stuff that I had read, you know, interviews and articles from Wired and so forth, stuff that I had read a lot. A lot was in the book was not actually all that new. He was taking time off. It was a year off that turned into four years off and so on to develop, rediscover and develop creative bits that have atrophied because of the business stuff. And, you know, and really what I want to say is, you know, if you know Bunny or if you have some kind of contact connection to him, you know, indefinite scalability, the T3 tile, you know, one of the other things I know about Bunny just from stuff I've seen over the years, that, you know, he was doing some kind of artificial life related stuff when he was doing some badge life type stuff, making, you know, conference badges that had electronics that could interact with each other. So he's got a life awareness. And I really think, you know, they've done a bunch of projects. They did open source Moore's Law is running out. You know, the CPU model is going out. No, it's definitely running out. The future is like FPGAs which is much more like the movable feast approach. He spent time with Sean Cross, his collaborator in a lot of these projects, trying to make an open source phone, which is a really wonderful thing to do, really thankless and difficult. And, you know, so they spent a ton of work, a ton of time, and they presented their results, but the week before we submitted, we learned that the company that made the chip had released a platform. And, you know, so in the end, there was a lot of enthusiasm, but not a lot of action. And it seems like, you know, it didn't really go anywhere. And I sort of think, you know, maybe you need to look a little further down the road, maybe not quite so close. You can still have impact. You could have bigger impact on society. T3 indefinite scalability. Think about it. All right. Anyway, so that was the Hardware Hacker Adventures in Making and Breaking Hardware, even if it's a 2017 book, rather than a merely 2019 Kindle book, it was pretty good. All right. The other thing I did, I worked on, you know, I was talking last week about concerns about the robustness of these parts, and I went from these four little ribs, whoops, to three slightly thicker ribs and a little bit more reinforcements after last week's video. Andrew on the 3D printing room in the Gitter was saying, you know, you could try PETG, you know, it would be a lot more robustness and since these pieces are all just like flat and simple. And, you know, I bought a roll of PETG a long time ago, just on general principles, and I've never opened it. Andrew was the kick in the pants I needed. And so I went and I dug it up and found it in my closet underneath my glow-in-the-dark fractal pyramid, and I opened it up and I put it in, unloaded the PLA, you know, which is 260 degrees, that's 60 degrees on the print bed, which is plenty hot. These are setting great temperatures, but the PETG is, you know, 90 degrees. That means the the print bed is getting close to boiling point, you know, and it's way above the boiling point for extruding the plastic, you know, so this is what was scaring me to begin with. But, you know, it printed great. It was like no big deal at all. Andrew told me I'd have to worry about a little more stringing and stuff, and there was a tiny little bit and, you know, boy, it really sticks to the print bed, you know, I was going to get in there with my spatula but it was okay, although, you know, so this was the second round of it and, you know, there was sort of, you could see like these sort of little outlines. It was as if there was some sort of chalking coming off it or something like that, I really don't understand it, but it cleaned right off and didn't seem to matter, and boy with so much stick flatter, you know, having it not come up, having it not be clean enough to stick doesn't seem to be a problem. So I did a whole other power zone in PETG. It did the end pieces and so forth and it all seemed to be very nice and, you know, I've been playing with my little test pieces for, you know, several days and I've been getting increasingly mean to them and, you know, they're holding up. They're holding up well, so this is encouraging. Eventually, I said, okay, well, I'm going to tear down the PLA power zone that I made and move them all over to the PETG. PETG. See how it works. And I discovered well, this guy didn't come off right because in fact one of his feet got left in the frame because it got pulled out, it stripped its thread, so that's a little bit of a problem but okay, we'll deal with it. There's plenty of more feet where they come. The problem is going to be getting the feet out if they pull out. So I think maybe I hadn't screwed it down far enough and I was reusing them, you know, even if I could have grabbed a new foot and tapped it into fresh plastic, I was reusing the ones. I thought it was okay, maybe it's not okay. Here I was actually getting the core injector into the middle of the thing as I was building it up, tucking the cord down around the frame and getting the next guy in. Sorry about the focus on these things and this is what it looks like from the back and so you can have the power supply of the wall ward back there and so on. It was all looking fine, so I decided I was going to try to make kind of the last piece which is something to go along the top row and hook the whole power zone to a piece of conduit or pipe or something like that. So I designed this thing in the two directions, the flat piece would have the strength in that direction to interface with the existing struts and then the round piece would have its strength in the direction it needed and I printed up a bunch and they came off in PETG and they go together pretty nicely. So they have an asymmetric little set of things. They only goes in one way and you get it seated and it goes pretty nice. It doesn't go completely across because the bottom of the candy cane is closed to keep the parts from spreading easily and it seems pretty good and so I made a bunch of them. I clipped them on and they clicked in nicely and there it is. So now we've got on a single power zone you've got eight of these things 16 tiles in a power zone, eight of those things. It's basically carrying the weight of two tiles per for one tile zone. So it's under two pounds. The amount of load that one of these things has to hold and even if we go with three rows of power zones, which is what I'm imagining. I'm imagining we're going to have a three by three grid of power zones. Each of these things, since there will only be one row of them at the top, is only going to be holding about 10 pounds each like that. That's not so bad. I think that may actually hold. I'm more worried about the stress right around here than I am on the pipe itself. But we'll see. I think it's really pretty doable. There it was. I just hung it on the edge of the table just to see what it was like and then I went and got a piece of conduit. All the conduit I had was busy so I used a home handle. But there it is. It's hanging right up by the ceiling by its I don't know what to call them. I call them pipe interfaces but that PI has already taken for power injector so it needs to be something else. But it seems extremely strong and they're actually hanging quite vertically and they look pretty good. That's fuzzy in the dark. They actually don't look bad with the flash as well. So that is kind of encouraging. I think we can hook two more power zones on below that assuming that the conduit is tough enough to take it. This whole thing has got to weigh about a hundred pounds with the nine power zones plus the power supplies and so on. That's it. Alright. Next week will be a recap of the first year of the T Tuesday updates. Thanks for being here. Have a good week.