 The T2 Tile project is building an indefinitely scalable computational stack. Follow our progress here on T Tuesday updates. So check it out. This is the beginning of the next T2 demo. It's the 2D printer or at least it's the first 2D printer demo. The little green thing is the object to be copied. It's passing the swap lines through their white and black rows. You can kind of see them. Copy line by line lead behind the children as the parent moves away. And in the very first generation, if you look close, there's a mutation. I mean, on the one hand, I say, you know, this is obviously because I am a bad programmer. I'm old. I'm stupid and all of these things. But then, you know, the bigger picture, it seems to me is that getting perfect, you know, extremely high reliability copying is non-trivial. When you're doing this spatially distributed, you don't have, you know, at edges of memory. It's all happening in the same place. Whereas imperfect copying, which leads to evolution, is much easier to come by. So that's the motto for today. Evolution happens. So it's been two weeks. There's lots of news. Let's just get into it and see as fast as we can. We have two, well, we have two new recurring donors. I really didn't even think that was a thing. I mean, I knew it was a thing, but I didn't think it would be a thing for the foundation. Abhinav, who got a nerd number a while ago, has, you know, stepped up to recurring donation. And Brian, a guy from, that I knew from UNM days, took a bunch of classes with me. Did some of the, you know, early stuff with the Mobile Feast Machine has reappeared. It's a monthly donor. Hi, how you doing? Thank you. I know exactly how this is all going to work out, you know, it remains to be seen. But just the feeling of getting support from folks. Oh, and I haven't sent the emails, folks. So as far as getting your actual nerd number for you, Brian, it's just because, of course, I'm terrible. All right, moving on. Oh, yeah, I have been invited to give an invited talk at the eighth workshop on biological distributed algorithms. It's sometime in May, I think, not that far away. It's virtual. Invited by, you know, a friend and colleague who is, you know, on the committee for the workshop as often is the case. We talk about, you know, the old boys network or the just old people network. And there's a lot of truth to it, but that works true for me too. I mean, I would much rather work with friends. I work with people that I know. This is a little bit scary. I mean, biological distributed algorithms sounds a lot like living computation. And it is. But my understanding of, I haven't been to one of these is that, you know, it's biologists, real wet biologists or the spectrum of biologists at one end and computer scientists at the other. And then the computer scientists who are take part in this tend to be toward the more theoretical end of the spectrum. Pod-C principles of distributed computing is one of the big theoretical distributed computing conferences. Probably the biggest one, maybe as far as I know. And I'm not at the theory end of the computer science spectrum. I'm at the systems end of the computer science spectrum. So I don't know exactly how this is going to work out. We will find out. Stay tuned. Joey Castillo, someone I don't know, posted a tweet about, you know, machine learning problems and how do you like you stick a piece of paper on an Apple saying iPod and the vision system calls it an iPod. Put a little piece of tape on a 35 and the Tesla speeds up to 85 and so on. And this reminded me of a picture that I had taken way back in 2013 and posted on Google Plus. May it rest wherever all the Google apps that get tried and dumped go to rest. So I dredged it up and I reposted it as a reply to this. And since Joey's tweet did a 10,000 plus 10, I don't know, 20,000, lots of likes, I got a tailwind. And so this is it, robot trap. You know, I drove by this one day and it just struck me as funny. So I took a picture of it and, you know, it fits into the whole rule following frame of mind that we think robots would be that way. We don't really think robots would be that way. But that's the way to the traditional view of robots. And so then in addition to everything else that's happened over the last couple of weeks or not all of it, but the last week, particularly. I've been kind of keeping an eye on the notifications, watching the likes come in and so forth. And, you know, with my absolute newbie Twitter, nothingness. This was this was fun. This is more likes than I've ever seen. And as of this morning, it was like 395 likes. So I'm betting it's going to break 400. And among all those likes, there were a bunch of people that, you know, certainly I had absolutely no connection to who followed. And some of them liked and even retweeted some of the older stuff that I had done that was clearly in the living computation thing. So yeah, you know, me, the last person to get to it, of course, you know, the power of marketing. You got to get out there. Well, so I got out there a little bit. Yay. In that same spirit, last T Tuesday update 295, you know, with our club of 150 reporting in. Thanks, folks, on clicking in the YouTube comments is doing almost 200 views. It's number one in the last 10 videos on the T Tuesday updates. And so that's good for what it is. We'll see whether this one can match that as well. I kind of hate the fact that you're always competing with yourself and YouTube just keeps showing you how you're doing against yourself so that you cannot possibly win. Next week, next time the standards are going to be higher because of the success of this one. All of this video for both for the updates and the demos and all the grid recording and everything. We are running out of disk space. So we're shopping for a network attached storage. You know, if I was younger and more like I used to, I would be building one of these myself, you know, a little Linux box Raspberry Pi. Who knows what it is. At this point, I'd rather throw a certain amount of money at it and, you know, get a pile of disks and have somebody else maintain it for me. We have one of these synologies, a small one and it's been it's been fine. So we're looking at that. I'm thinking raid 10. I do not believe raid five, six. We have any raid proponents out there. What do you think I should do? And we have well, so there's the there's now two demos up on T2 demos. The third one on the evolution one that we saw a little bit here will be up neither today or tomorrow. Whenever I get through all of this and step by step, we now have some subscribers and so forth. Okay, yeah. One last point about YouTube. If anybody watches anything on YouTube, you've probably seen Neil deGrasse Tyson or the like in this ad. But you haven't seen this, you know, this is an ad on one of my videos. It's on the computer science, the history of computer science video, which, you know, I don't monetize and I don't really want into. I never really wanted to. I mean, I don't have enough traffic. They don't even this is a nice trick, right? I do not have an I have enough subscribers on the day back the channel, but I don't have enough watch time hours for YouTube to say I could be a partner in their advertising. But now, and this is a relatively recent change, although they pretended it wasn't a change. They are starting to put ads on videos of people that don't monetize like me. And so if the beginning of the video is like appropriate or whatever, they'll feel free to do this. I read about that a month or so ago, but I didn't really expect it to happen. So now I'm wondering, you know, do I need to make a new little intro bumper that we would put on it that would have stuff like this? You know, just a little montage. So to make the make my videos a little bit less advertiser friendly. You know, or am I supposed to relax? Am I supposed to say, yes, it's okay to try to make money, even though that's not what this project is about. And I always felt it was sort of a mark of honor that I wasn't trying to make money about it because I had a bigger purpose that the goal is to change society to change the computer science industry and not to make money for me. I start to wonder sometimes whether I'm going to have to try to make money in order to get anybody to take it seriously. But who knows? We'll see what happens. That did kind of aggravate me. All right, that's the news. Most of the time this week, I was yak shaving on the We Are Coders Hyperspace Lecture 2, which was supposed to be out in January of this year. And it's not out yet. But I'm going to declare that the next T Tuesday update, April 13th, is going to be this video. It'll be over on the Dave Ackley channel, but it's going to stand in as the next T Tuesday update. So it's now on the record. It has a specific date that it's got to happen. We are coders. You know, the idea is not just programming like programming regular computers, but also programming each other using natural language like English. Like it's happening right now. And if that means we are coders, that means we're also the machine if we're just exchanging programs with each other. So in order to understand the sort of We Are Coders Hyperspace point of view, we have to say, well, what kind of machine are we? What is the programming platform, the architecture that we are programming for when we are programming each other? And because the underlying hardware, the neurons and stuff are all extremely flexible. I mean, yes, they have, you know, they go where they go and they don't go where they don't go. We have optical illusions, all of this stuff that are limitations on our brain hardware. But still it's amazingly flexible so that you can implement many different kinds of virtual machines. And that fact that we have a choice about what kind of virtual machine to implement on our own wetware and on each other's wetware, each other's brain means that there's this opportunity to say, if we say, you know, let's shift our notion. If we shift our model of what kind of machine we are, that will shift what kind of programs we can run. It's like changing one virtual machine to another. The underlying hardware could run either virtual machine, but this virtual machine allows you to write these kind of programs. This kind of virtual machine allows you to write those kind of programs. So that's the idea. What kind of machine are we? The WeAreCoders approach says we are, we have four primitive operations. We can accept input from the world. We can perform steps, sequences. We can judge states of affairs, see whether we like them or not. And we can do things. We can make output. And this is us. Hyper subspace search and sequence limited. The sequencing I mentioned, the search, you know, all this, this is going to be the subject of the other video. But the limited is to threat to stress that number one, we have limitations and but number two, we are a corporation. Whatever else we are as a machine, we have to make a go of it. We have to keep the system running. And that is fundamental to everything about life and to everything about how we interpret programs that we are coming from other folks and so on. So input sequence, judge output, those four things. We have to keep putting ourselves. Don't forget them. We have to keep working on advertising them in a sense. We have to make them as simple as we possibly can. So I made a little symbol. It's got it's got a bunch of arrows, the input arrow. Where's my thing? Input coming into the center of the sequence. It goes around the arrow goes up around and comes back down the judgment arrow. Things are bad. Things are good. And the output arrow like that. So this I call the self image. This is the machine that we're a programming for it's a cartoon. It's an advertising slogan for it. It's a logo. And so yeah, so I made it and so here it is input sequence, judge negative judge positive and output. And so, you know, I made I did this several months ago in 2020 actually. And, you know, I've even I've even got some some couple little little one of these things printed up its shapeways and metal and stuff like that. So I've been living with the self image for a while here. And I like it pretty well. And I also said, you know, geez, this Prusament clear filament. And that's an example of it. It doesn't end up transparent because it's still 3D printed, but it lets light shine through. And I have this whole reel of these tiny little programmable multicolor LEDs. And I thought, hey, I could take a bunch of those LEDs, put them behind one of these transparent. Well, this is white, but one using the transparent filament, the clear filament and have, you know, animations of here's the input happening. Here's the sequence happening and so forth. So I decided to go for it, put a whole bunch of them around the thing and so on. So that's what I spent a lot of time about, which meant I'm back to hardware. It's been a long time, but I put together a quick schematic. It's really just, you know, a chain, a daisy chain of these little circuits, these little LEDs. You just run three wires from one to the next to the next to the next to the next and put down a capacitor every solvent, maybe. And I took the 3D model for the self image and flattened it out and made it into a circuit board. Eventually I went back and I shrunk it and rounded it off so that the circuit board was a bit smaller than this thing. So hopefully it'll fit in behind and it'll all work out. And I laid out the pixels, the LEDs on it and did the traces and so forth. I sent it off to PCB Way, the place that I'm most comfortable with because I did, I used them for the teach you tile boards and they just marched through all the steps. Sautermask, routing, delivery ready, DHL and wham, here it is. You know, here's the box and, you know, it's too much time to go into, but there's a, I had a freebie. I had leftover credits at PCB Way because they give you these little, you know, bees, they call them. And that allowed me to actually get two of these Ratticebury Pi Picoes, tiny little, there's a picture of one. You know, they're not full Linux boxes and so forth, but who needs that for the purposes of drives and some LEDs. So I got a couple of these free thrown in the box. And I'm thinking of trying to figure out how to use that to drive the self image, the animated self image we'll see. And there it is. Indeed, it fits inside the footprint of the real thing. And I've been working on a controller, which I will spare you the details, but you know, so here it is for running the thing, saying we're doing input, we're doing sequence and so forth. So hopefully that ends up being good, but that's what it is. All right, I'm going to skip over most of this because it's nerdy. And, you know, basically it's, I, you know, while I was not looking on Twitter to see how many likes I've gotten while I was not hitting refresh to see where the circuit boards have been nailed to me yet. I was trying to implement this 2D printer that I mentioned last time and I am now on something like my fifth generation approach. And the one that's in the T2 demo that we saw at the beginning was generation three, I think, three, maybe it's four. And I've learned a ton and I'm really kind of excited about it, but it is pretty nerdy. So, you know, it's all just piles of text. I'm just going to skip through it because it's piles of text. I don't know what that was. So the basic idea, again, I took the, the, this thing called grass from a UNM computer science class project. I had a life project that I used to do called frog world that had grass and things that would eat it and so forth. And the grass grew in a particular rule to say, you know, as long as you don't have two neighbors, if you have, if you have only one neighbor, you can grow. So this one, everybody has two neighbors except the very tips in this particular case and I modified this grass so it can only get so big. So this thing doesn't grow any further, even though according to the original frog grass rules, it could grow until it had two neighbors. And I use that as the subject for the thing to be copied and it surrounds it with this plating and it uses it to measure the thing and it sends out a copy of the plate, the build plate for the 2D printer, which is basically saying trying to reserve space saying, you know, if I, I'm going to need to move the parent this far in order to leave a copy behind in the kid. And if that's not all empty, then things might go wrong. So it figures out how much space it needs and sends out the build, it builds the build plate and only once the build plate has been successfully constructed, does it actually start the copying process or at least that was what was supposed to happen. The version that's in the T2 demos still has some bugs, but again, the bugs are kind of interesting. You know, it's like the Morris worm, one of the earliest computer security failures actually has a bunch of bugs in it. But Morris, the guy that wrote it, lost control of it before he finished debugging it. You know, he was trying really careful to, you know, these things are dangerous in principle. They, they have gain. They, they say, you know, show me a computer, just let me talk to it. I will use my hypnotic seance on it. Now all of a sudden you're enslaved to me and you start saying the magic seance to somebody else. So to think that we're going to end up with a perfect copy that hasn't got any bugs that won't cause evolution to happen. It seems kind of silly and I'm continuing to work. I think my fifth generation thing is going to, it's going to, it's the one that's going to actually get it right. We shall see what we shall see, but we're not there yet. And so there's some pictures of it after it's all gone wrong. I mean, if we look at this, you see several different shapes of squiggly grass is in there, even though it started from one piece of grass. So clearly something has changed. All right. And so that's, that's it basically. And the rest of it you can see in the T2 demo that'll be up. I'm kind of excited about it. I think it really, you know, it's, I think it's really fascinating to look at. There's just so much stuff going on everywhere all over the place. So that's progress, I think. So in two weeks, not a T Tuesday update, but lecture two of the hyperspace Academy 101 introduction to classical hyperspace hypersubspaces. We'll see how it goes. And then I will be back for a regular T Tuesday update two weeks after that. We are closing in, by the way, on T Tuesday 3 100, which will be the 100th episode. And according to, you know, the visa con and all the stuff on the internet. And I'm supposed to be good at this after doing 100 of them. What do you think? How are we doing? I hope you're okay. Things are going all right here. I finally got my first shot. Well, finally, I was lucky enough to get my first shot of the virus, the vaccine for the virus. Step by step. Hope you're doing well. See you in a couple of weeks.