 The T2 Tile project is building an indefinitely scalable computational stack. Follow our progress here on T Tuesday updates. Hey folks, it's been two weeks, tons of stuff to talk about. Let's get into it. The top story this time is I did not submit to the Artificial Life Conference this year. As always, I had great ambitions, but once I settled down to actually start working on it with like three weeks, two weeks left, it didn't come together and I wasn't happy with the way it was coming out and I decided to let it go. The deadline was March 7th, but we knew it was going to be extended to March 14th. It was. It's now March 16th, it looks like the submission window is still actually open, but I didn't write anything. I mean, I wrote tons of code, but I didn't write a paper and that's sad in a way because it's been sort of a tradition that each spring there's a bunch of stuff that's happening on the T2 Tile project that I'm not talking about because it's kind of the obligation that I sort of think to if you're going to submit something new to a conference or journal or whatever that you're not supposed to really talk about it until the paper comes out. And I never really liked that that much, but that's the way it was. So for the last two years in 2019 and in 2020, we put up the wall of science between the spring like around February, March and like July on a bunch of work. And I didn't really like doing that, but of course it just is what it is. And what I would feel like trying to do this year is I want to develop this stuff in the open. I'm going to talk later sort of the feature story for this update, a little mission status review to try to reset and put things in context, but what we're actually trying to accomplish here or at least my vision for what we're actually trying to accomplish and see how far we've gotten and what the next steps might be and the stuff that I was trying to produce for the A-Life conference was some of that. That said, there is another call for the A-Life 2021 conference which is now fully online. They were hoping at first to have it be hybrid in Prague and online, but you know, world being what it is, it's virtual. The deadline for art and visualizations is April 25th and you know, we'll see. It's not out of the question and you know, there's the logo robot person without the mask on and the art and visualizations call all online and so forth. And you know, in the 2020 A-Life art competition, which I was one of the judges for, there was a prize for sort of science related art visualization as well as sort of art for art's sake visualization. And you know, it's possible, you know, we'll see. We might have something to say in a month. Talk about that more in a minute. So in the last two weeks, I started a new YouTube channel, the third one now, which currently has no subscribers. You can be the first subscriber. The purpose of the MFM, the T2Demos channel is to just collect the little, you know, two to three minutes or the pop song length visualizations of it doesn't have to be purely art or jokey or funny. You can have a scientific purpose as well, but it just needs to be really boiled down, you know, maybe absolute tops five minutes. And so far, I've briefed demonstrations, brief video demonstrations and I put up the the Eugene video, which is now at the moment on both, but and I'm going to put up the Don't Cross the Streams on T2Demos and in the future, they will go there. So if people are interested, and maybe someday people will, you know, a wider audience, you know, we've got our core of 150, 150 people that seem to check out the updates, you know, in tempo and, and, you know, thank you folks. I mean, it's not by, you know, winter to a ton standards by Colin first standards by all the YouTube stars. It's nothing. It's a rounding error. But it's 150 people that I feel are sort of part of the team and are staying along with the plot. So I need to feel an obligation to number one, advance the plan number two, let people know what what has been happening, even if I don't know whether it's a success or a failure yet because the story isn't written yet. So there could be a broader line of people that don't really want to follow the blow by blow on it, but we wouldn't mind dropping in to see a little piece of candy every so often. That's what T2Demos is for. Be happy to get as much pushing social media, of course, as we can, but at the moment there's only one or two videos there. That here's a teaser for another T2 demo, which I'm hoping to release next week although we'll see how it all works out. That should be sort of fun and nobody's seen it yet. In addition, also on YouTube, there's some new voices appearing as making suggestions and asking questions and had a certain amount of engagement and that was really good. Thank you folks a lot for raising issues and just letting me know you're thinking about this stuff because that really helps. Brian Canard, I'm not sure, I feel like he's a recent arrival, but I'm not positive. He found the research when looking for modular motherboard back planeless scalable connector design on Google. You Google that, you actually can find something back from 2008 that I've talked about it a little bit before that ended up in a wired online wired lab article. I don't remember exactly what hardware hackers create a modular motherboard. This is where I learned a little bit about quotations and accuracy in journalism. Nothing against Priya Ganapati. I don't know that I ever spoke to her directly. Him or her actually. I don't know her, I think. It says David Ackley, associate professor. We have CPU serial ports for connectivity on every two square inches. That's not actually what I said. I said on every two inches squared. Two inches squared. Not the same as two square inches. And even inside quotes, it doesn't come out right. There's the picture of them. The IXMs, the Illuminato X Makana, two inches squared. There are advantages to essential plant structure, but eventually we're running to great inefficiencies. I'm not exactly sure I use that word either. It's really hard. We have efficiency built into our brains so much as that is the nature of the good that even if you say you're doing something else, you're doing robustness, and they say, oh, will that lead to more efficiency? And you have to say, well, actually, no, at least less efficiency, but it leads to more robustness. Is that, oh, it sounds very efficient. So, and Brian Connard also linked to these folks who I hadn't heard about, that I googled for a moment, and it seems like they do FPGA stuff and this open FPGA project. And here's the thing, I've talked about the Illuminato X Makana, the IXM a few times before. But, right, when I looked long ago, FPGAs had unfortunate limits on dynamic reconfigurability. And, you know, so here's a little bit of show and tell. This is the Papillio Pro, and that's an FPGA right in the middle there, that big chip. This was designed and built by a guy named Jack Gasset up in Colorado who I actually worked with sort of on and off, and the idea would be, you know, he was selling these. And I thought, you know, what if we could take all these pins and move them around so that we could connect one of his Papillio Pros to other Papillio Pros. He designed, in consultation with me, the eight-fold mega-wing, like that, which is basically a connector board with, you know, there's a little bit extra stuff on it. And the idea is, you know, you take a Papillio Pro and it plugs in on a half stride, you know, slightly off one way and slightly off the other way. I don't know if I can actually do this, but when you get it right, you can plug them in. I'm not going to be able to get it right and we'll take the time, but like that. So you can tile the wiring into tile boards and you can tile the Papillio Pros behind them. We did one generation of this, never actually got it working. I did go down the road of learning enough about FPGAs to learn how to make it like a little CPU, a little tiny processing unit inside an FPGA, tried to start building the communications mechanism. So these things, I mean, if this thing had gotten working, it would be, you know, blazing fast as far as the communications compared to where the T2 tiles are so very slow. But engineering, I mean, hardware is just one straight up after another. And so when Brian Kanard, it turns out, you know, ICE40LP, Google that, it turns out, you know, that's an FPGA. And I really don't have him follow exactly what he's doing. He, you know, could help at least get some links or something that I'd actually look at. Maybe Brian and I could talk at some point. I've solved dynamic FPGA reconfiguration in an object-oriented way. I mean, if you squint, that's kind of looks like what the T2 tile is trying to do. Take a uniform, relatively uniform hardware platform, which is what a field programmable gate array is or does, and allow it to talk to others of its own kind, so that at one moment it can be acting like a drag, at one moment it can be acting like a sorter, and so forth, without the hardware actually changing, just by dynamic reconfiguration of the same existing hardware. So, Brian, in any case, thanks for stopping by and making contacts. All right, and this video, that is not my house. I think for the first time in the history of the universe, a T2 tile is booted up not by me, or not by us here in. Strider, who I've known since UNM days, has been working on software. In particular, he's forked a version of the underlying engine stuff to update, to upgrade the graphics and the interface stuff. There's a particular library called SDL that I use, the old one. That he's now upgraded MFMS, the simulator, to, and, you know, the next logical step to be able to merge that stuff, because there is so much commonality between the code base for the simulator and the code for MFM T2 that runs on the tiles, is that he needed a tile to actually try out, if he could port the code, the SDL2, it's called code to the tile. So, I shipped him a whole little care package with a tile and a power supply and a serial cable, and he is now beginning that road. He's put up a couple of videos, he's talking to it from a Mac, and he's got in the T2 splat code, other stuff going on, and in particular, he's got a journal file, I encourage folks to keep an eye on that, if you're more interested to see what's going on. That's very exciting, it's very terrifying. You know, these things are rare. But, Strider says, I found the tile to be much more robust than I had imagined. All right, so that's that. Finally, it's now been a year since we incorporated the living, since we formed the Living Computation Foundation, our nonprofit to try to back all this work, and to build outreach for the ideas more generally, as well as the specific computational ideas. Spring Incorporated, Spring of 2020, right now it's Spring of 2021, it's time to do our first tax return for the foundation. We've raised a little bit of money, we've spent a little bit of money, and we have our LCF nerds, we're now up to the 240 series of LCF nerd numbers, and we have monthly contributors, thank you folks, so much. Exactly how we're going to position everything, and I use the LCF, I use LCF, Living Computation Foundation, trying not to say the the, I use Living Computation Foundation as my main affiliation now, so step by step, all right. And so that's the news. Now, we put out these little videos, we put out these little art goofy things, like don't cross the streams, and so forth, and people can look at them and they can think they're interesting, or funny, or pretty, or not, or whatever it is, and that's all good. And I think T2 demos are really important, not just for outreach, in a sort of art and fun to look at, well, partly fun to look at sense, but also in the sense of training our eyes what these systems look like. They are not kchunk, kchunk, kchunk, kchunk, like the master synchronizer that makes everything perfect, there's always little ripples around the edges, and that's the way these systems work. And so having these demos all different ways that shows, you know, things spreading, things coming together, whatever it is, is all important, I think, for training our eyes. The bigger picture is that we're going to take this bottom up, distributed, robust first approach, and work our way back to, work our way up to a computational framework powerful enough to do useful work. And there we go, right. And so this kind of comment that comes up on YouTube, about when can I start to deal with my Excel dialogue using this sort of thing, or when is this going to implement a spreadsheet, or could I use this for video gaming, it's a standard kind of question comes up. It's all variations of what is the killer app for this thing. And my answer is, if there's a killer app, I wouldn't tell you, because this is early research, this is research and development, but that doesn't mean there isn't a goal, a plan, a path. So I don't even know if any of this is actually visible, but so here's the idea, we're building a whole new computational stack, from physics to people to society. And at each scale, things get bigger. So we've got physics is concerned with picoseconds to milliseconds on this cartoon story, chemistry is milliseconds to a second and so forth. And there's computing analogs for all of this, that from one bit to an event window is sort of the domain of physics. And from one to a thousand sites perhaps is the domain of chemistry. And right on up the stack. And so what we're trying to do is build our way up that stack. And here's how our progress that we've made. We've got actual tiles running that are doing bits. We've got languages for supporting, manipulating those bits for the chemistry, for the order of a thousand sites or so. We've got various demos and we're going to keep working on that. But we want to start moving on to the next step, the biology step, where we have something like a cell, something bigger than an atom, much bigger than an atom, that there are many copies of it that are related, similar, differentiated, identical, whatever it is that interact to perform some larger problem. Because that's going to be the key to reaching a first level of utility. We're not going to expect to find a, to make a marketable app using individual sites and atoms. It's got to be stuff that's doing something complex in a collaborative fashion. And that's what this is getting at. And that's where we are now. We're sort of at the tipping edge between sort of digital chemistry and digital biology on this point of view. So, you know, in physics, you know, MFM, MFMS, the simulator, MFM T2, the tile engine, those are physics level, ULAM, SPLAT, and SPOT from last year, the stage priority operation teams, those are dealing with coordinating, you know, groups of atoms, you know, perhaps hundreds or thousands, but not millions or billions, not directly. There's going to need to be more structure. There's going to need to be more organization in order to do that. And biology is the next stop. Multicellularity, reproduction, differentiation, and signaling. So here it is, running, taking a lot of time. I'm going to chop off part of this here. But so the goal, my goal, A goal, a proposed goal for 2021, is to actually get high enough up in the biological systems so that we can do simulated simple control system demos. Like we have a simulated robot, perhaps something as simple as a Breitenberg vehicle that sort of turns toward the light or perhaps even something more complicated, like a cart pole, a traditional problem for simulated robotics. But what we do is we take a simulation of it and we connect it to the real grid. We have new hardware that does input and output in some fashion around the periphery of the grid, perhaps injected into the center of it, who knows. And we let it flow through doing its biological, chemical, physical type bottom-up processing. And we take signals off it and we say, okay, as far as the simulation is concerned, that was 100 milliseconds or something, even if it takes 20 minutes in reality for us to do it. That's the goal that I would like to achieve. I'd like to achieve a lot more goals, but that implies many of them. All right, so yeah. And so getting to reproduction and differentiation cell structures is what I've been focusing on, what I was trying to focus on for the A-Life 2021 that I did not end up submitting to. I was working on stuff that went all the way back to 2016, which was a very simple-minded method of copying structures. This is an example of a structure being copied that I called the 2D printer back then. I was trying to refresh it now over the last couple of weeks. And it's exactly one of these horrible early technology failures, like rockets falling over left and right and crashing and the early attempts to fly with airplanes with six wings going three feet and then folding up. That's what was happening to me, trying to get this replication method working. So here I've got a scribbled little blue thing of stuff that's meant to be copyable by the 2D printer technology. And the idea is it passes swap lines through a special kind of swap line to copy line by line and then deposit them until it's made a copy. So there's a bunch of swap lines going through and there's some kind of horrible transporter accident and evil Kirk shows up on the transporter pad or in this case it's actually more just sort of a puddle of Kirk-related goo is what I got this time. Eventually I fixed enough of the bugs so I started getting more interesting failures like this one. I got an absolutely perfect copy but mirror image talking about evil Kirk. And so now that's where we're at. This is running on the Keymaster. We've got a little something make a little squiggle and let's see if we can reproduce it. So that's the next T2 demo. In the service of the larger goal of we need to be able to make cellular structures bigger things hundreds of cells that would be bunches of them which will rebuild themselves from scratch and if in case something goes wrong and reconnect and so on. That's where we are now. We shall see. Going forward I'm going to keep working on the new T2 printer stuff. The copy area the reproduction system because it's getting close and it's it really is kind of cool and interesting things go wrong. Still but also for the next update it's going to be the hyperspace academy. Second lecture is this whole other thread that will come together in the end if we imagine to enough work that we are coders. We ship code way of looking at the world. Way of looking at human interactions. Way of looking at computational interactions. We shall see. Thanks for being here. I hope you're doing all right. Hope to see you in two weeks.