 Hey folks, it's T Tuesday. This is, we're back. This week my theme is reaching. The theme we had a couple of weeks ago, embracing failure. Not that popular, I wonder why. We'll see how this one does. Reaching out in a number of ways. So now first off, it's outreach. Whoops, and let's get over here. Okay. I feel like I'm doing a lot of outreach all of a sudden, and beginning to get a little bit more traction on Twitter, and so forth. So, right after the last T Tuesday update, I joined another Santa Fe Institute collective, workshop on collective intelligence. I started on Wednesday, September 1st, the day of the T Tuesday update, but it actually had started on the day before, on T Tuesday. And, you know, as usual with these SFI workshops, you know, there were all kinds of interesting talks. And for me, one of the most interesting things was to get folks who have been studying more or less computational takes on human societies, stuff that's actually involving people. Because certainly for the bigger picture of living computation, maybe a little bit beyond just the T2 tile project itself, that is all about understanding larger scale systems, actual living systems, as well as manufactured systems, as well as mixtures of living and manufactured systems, mixtures of natural and manufactured systems, not sure how to say it, and how they behave, and how they can be organized, and how they fail, and what we should be watching out for. Henry Farrell was talking about, you know, open access versus, you know, communication strategies that help flooding. I mean, there were lots of interesting talks. Jacob Foster was sort of a sociologist take, which was really interesting to me because it's something that I don't have as much contact with. That was very interesting. Anna Dornhouse, who I've seen a number of talks that she's done. She does all kinds of great stuff. Here she was talking about ants, and how they vary in their communication strategies, and what it might suggest, you know, and you know, look at this, ant hill in the middle, and then this whole apparatus so that the ants have to decide which path they want to take. And no matter which one they take, they're going to be going, uh, wondering about the purpose of life as they go back and forth. But the result of this is that, you know, there's this relationship between some species of ants are much more oriented towards exploiting it the best that they know right now and dominating, like solenopsis xylone, which is a nasty stinging ant, you know, starship troopers, or at the other end, much more oriented towards reinflexibility. There is Dory Mermek's Insanus, like the T2 Tile Project community. Lots of interesting stuff came out of that. In addition, this had been working, brewing under the hood in the background for a couple of months now, and it actually happened. Mark Hurst is a fellow that runs the tectonic broadcast that is broadcast on WFMU out of New Jersey every week. And he interviewed me a week or so ago for the open first episode of his fifth season of the tectonic podcast, which was yesterday. And that was great fun. And so this is, you know, this is his take on the T2 Tile Project, right? Dave Ackley has a modest goal of rearchitecting all computers, you know, ought to have some ambition. On the next tectonic, Mark and Dave discuss how the dangerously outdated ideas creating today's computers, the dangerously outdated ideas, and how Dave's teach to talk about building new decentralized to save the world from big tech. And, you know, one of my traditional problems is that I am terrified from a scientific point of view of over claiming. And so I tend to, you know, I don't make clickbaity titles, although I'm working on it. And I try not to get too far ahead of what the work actually shows. But it's really refreshing and helpful to see other people's takes coming back at me on this stuff. And, you know, there was a chat, you know, he found the original brick wall construction illustration that he used on the podcast page. And there was all this kind of chat, interesting stuff going on. I joined in, tried to chip in thoughts from here and there and so forth. It was really great. So, Mark, thank you for that. And, you know, Jeff and Michael, two guys who've been sort of working behind the scenes, helping out with the project for years in at least for Jeff, got this set up. And that was just super great. So that was the tectonic podcast. And that was yesterday, but that's not even the last thing. Then just last night, Ivan Reese, who is the current steward of the future of coding community, which is a podcast and all kinds of related material about trying to take a bigger view of programming and coding, which, you know, very compatible with again, the larger the hyperspace academy stuff that we are coders, philosophical perspective that I'm trying to pitch in the bigger picture where the T2 tile is going to fit in as an implementation example, that, you know, part of the whole weird coders philosophy is that the way you can, you make, you put substance under words, you know, you can say these things like representation and ideas and truth and who the hell knows what they mean. Well, the point is, is we are now in a position that we can build machines, you know, using regular computers or new kinds of computers, whatever, that exemplifies, well, maybe what representation means is this, maybe what an idea means is this, and we can show it. And, you know, it may have more or less to do with how the brain and the mind works, but it's an actual concrete thing that we could look at, we can poke and we can see how it behaves independent of our conceptions of it. And that's what makes it valuable. And Ivan Reese, you know, reached out to invite me to come talk on the Future of Coding podcast, which I would be happy to do. We talked a little bit in Twitter and it's probably going to be in December. We'll see how that goes on news as it develops. That was yesterday, Future of Coding in December. And last week's update in televerses, Tesla, it's done relatively well by the standards of the T2Tile project. I mean, we are seriously nerds here. I mean, one of the things that certainly was helped driven home by work going doing the tectonic podcast is, you know, regular people, computer stuff, computer architecture, computer programming, that stuff is very, very far away from normal people, the modal person. And so it was important to figure that out and see where we were going with it. 186 views. It's number three. See, now one of the things that's a little bit bugs me about the T2, sorry, the YouTube studio setup is that I'm actually now competing with myself because I'm putting up the uploaded version that's all trimmed and, you know, easier to find. But the live stream is still up there too. So there's actually like 66 views on the live stream, plus 186, you know, it's important for me to stress this. So in fact, Intel versus Tesla would be the number one video in recent days for the T2Tile project. If we merge together the sets of views for the live stream and the, you know, okay, whatever. But there was an interesting thing. I just sort of ran across this. There was a little pop up hover thing that I hadn't seen before. This video is reaching a wider audience on YouTube leading to a lower click-through rate, wider less. So boiling down to don't freak out about your 4.2%. Now 4.2% didn't look that bad as far as the T2Tile project is going. But hey, I'll take it. You know, there have been a couple of views. The SystemD rant that I did, the hands-on with the Beaglebone AI, and now Intel versus Tesla are likely to be the, you know, the evergreen Tuesday updates. You know, would I rather it be the SystemRant and so forth? Yeah, sure. But you know, hey, we'll take it, whatever. I mean, the bigger picture for me is it's about reaching out into the world as well as looking into the project. And we need to do both to get where the project needs to go. All right. So that's outreach. Reaching up. This is the stuff that I talked about just very briefly at the end. Last time I'm going to expand on it in some detail and hopefully do a demo or so. It's what we saw in the opening video. So last week, I briefly mentioned the idea of persistent transience. And the goal was to have something that's bigger than you could put into a single atom, a single site, a single cellular automata cell to start doing higher level programming, higher level coding in some reasonable fashion. It's not just good enough to do a one-shot stunt. We need to actually build a programming environment that, excuse me, that we can code in and actually get some larger scale stuff done in. So what I did is this was the picture from last week, had a good bigger objects and so forth. And you know, I called a persistent transient, but number one, that's not a great name. It's internal contradiction. Sorry, I didn't like it. So there was this idea of the activated complex, the notion from chemistry where you get all the pieces together. That's the activated complex. And then the reaction happens. But activated complex, I tried using that activated complex.ulam, it was way too big. I boiled it down to Plex. So that's the new name for persistent transient, the Plex. And a Plex, you know, from the Greek and Latin roots, you know, it means like weaving or braiding, simplex means like one fold complex means many fold. So we're braiding together different pieces from different atoms to make a more complex structure. That's the process of Plexing. It seems okay. It's nice and tiny. So now it's transient Plex. Once again, it's got this gather gather to pull together all the bits and put them in this bit storage that used to hold them temporarily, do the event and then scatter them back out to all the changes get recorded. Now there's a new class called Plexer, which is a base class for any atom that actually wants to participate in this Plexing process. And the main event that the main method that it needs is get Plex. So we need to be able to know given an atom that the large scale virtual object that we're supposed to be made out of it, we need to know what class that's supposed to be. And that's what the get Plex method is designed to return. Now there's also a second method print Plex that I'll hopefully demonstrate for you that's got some new magic that I'm very happy about and it took several days to get to. So here it is L2 plate. Again, this was all off of the level two plates plates made out of level one plates, which we're doing as a process of sort of maintaining, you know, tweaking the balance between robustness and structure at the level one plate, we're saying that's going to be relatively fragile and tightly coupled and relatively small. Therefore, the next scale up the level two plate, they're going to be more decoupled. They're going to be able to survive without having everything all there at once, because they're going to be big enough that they're going to be spreading across tiles and tiles are going to be messing up and we're going to need to have a more flexible coupling between the individual pieces of the L2 plate. And there it is. Now, in addition to having L2 plate base as a superclass, it also inherits from payload T and Plexer. And here it is. So, you know, it gets very twisted, it gets kind of involved. But there are several stages, right? So we have the actual atoms, they inherit from Plexer, and they then report the type of atom, the type of virtual atom that they want to participate in. In this case, it's saying, I want to participate in an L2 seek plate, I'm sorry, L2 seek Plex atom, because the goal of this first test was to get level two plates being able to have a sequencer, a higher level control, without tying up additional spaces inside the L2 plate border. And from there, here it is. Now we get to the L2 seek Plex itself, which inherits from Plex because that's the, so this is the process that actually has the gather, gather and scatter methods, and in case it looks, you know, in the middle of a side of a plate. But that's not even it, right? The L2 seek Plex is the thing that gathers the bits together and then goes Shazam and makes it into the virtual object, the transition state. That is L2 plate sequencer. So we've got Plex, we've got Plexer. Now in this particular example, we've got L2 seek plate and then L2 plate sequencer. Wow. It's getting a little easier in my head. And so once you get to with the actual L2 plate sequencer, it has an update function, which is a lot like an atomic individual behavior function, but it's this whole upper level. So the L2 plate sequencer looks for packet coming and going to other from its neighbors, it scans out, it does sensing out into the world around the L2 plate, and it checks to see if it needs to seed neighboring L2 plates. And it's beginning to work. That's what we saw in the opening video. Let's see if I can get this to work here. Okay, I didn't quite get this as automated as I want to. All right, how about that? Do we get that? Okay. So what we've got here is a seed my element, which is going to pop up and make a little three by three bunch of my elements, which is great. And now we're going to take the seed L2 plate. Again, this is just what we saw. There it is. Look at that. Okay. So this is all what we saw in the opening video, but now there's the one more step. Let's try this. Get an Adam view window. So here, you know, now we're just looking at some random L2 plate on the border of the thing. And, you know, this has got all its particular details. And in particular, I think, can we now see this? Let's go over even further. There you go. Maybe we can see it. Can maybe see the bits changing down here, right there, except they're changing very slowly, because there's kind of a lot going on on the workstation. And this was, you know, this is what you get. Where is the L2 plate sequencer? You can't see it. There's only one piece of it in this particular Adam. But now, voila, there's a new tiny little button up here. And here's the L2 plate sequencer. Here's the distance it's got. Here's actually some little debugging output that's coming out in the middle. This is what going for the full structure of L2 plate sequencer and all that stuff actually got us. We have to write our own code to do this. But there's default code, which in fact is what this is using, that allows us to see the Adams, the result of pulling them together in the gather like that. And this, of course, has already, you know, I'd say close to already paid off its costs in the... Do we have two of me on the screen? Oh, dear. Well, okay, we've finished the demo anyway. So let's go back to... Or let's try to go back to... All right, there we go. It still looks like two of me because... Yes, folks, it is live. Live. Okay. So that's the demo I wanted to show. Now... No, no, no. Decent. I'm all messed up. Select a scene. All right. And now... Oh, yeah. Well, actually, this is another thing I wanted to show as well that I'll do real quick and then I'll finish. I actually have gotten a little bit back into the tiles. And there is one new feature that I've been dying to see, dying to have for years. I'm sorry. The focus is so terrible here. But the important point is, you know, here... Everything's messed up today anyway. So the white... Here's the key master. I'm going to plug in an Ethernet cable. And can you see it? Not yet. Uh-oh. There it is. I don't know if we can really see it. But right up here, there is the Ethernet address, which is something that I've been wishing for for so long because when you plug in one of these tiles, you know, who knows where it actually lands. So all I can say about that is I am gradually getting back into actually What a disaster. Why did that happen? Oh, of course, because I didn't go back over here. All right. So the insiders report. Let's call all of this last stuff the insider report. Supposed to look for a tile failed event. Didn't do that, but actually did get into it. I have started looking at the code. I got to the, you know, the first place where the stuff fails when there's a timeout. I think I'm going to, you know, go ahead and build a packet buffer so that we can keep the outgoing packets until we actually see the next packet coming, the next packet coming back in the handshake sequence for a given event and rebroadcast if necessary. Will that solve all the problems? No, it won't. There still will be bigger failures. They're going to have to be dealt with. But I've started a new branch on the GitHub repo for MFM called Don't Worry Be Crappy. And I'm really trying to convince myself to just let it all go. Let the errors come up to a visible to user level, and let's just start dealing with it. Let's just see what actually happens. All right. So that was the demo for that. All kinds of stuff happening in ULAM 5 land. I am going to save that for another day in the spirit of it would actually take some time to go through some of this stuff. And some of, you know, it's been three years since ULAM 4, something like that, at least. And so a lot of these like multiple inherent and shared base classes, that's really been around for years. And they've really been key to all of the stuff that's been starting to work, the plates, all of this stuff absolutely depends on that structure. I mean, could it be done another way? Absolutely. But doing it this way has worked out really nice. So as far as I know, the few folks in the world who are willing to actually play with ULAM code are all using the develop branch from GitHub. So the fact that we haven't officially released ULAM 5 is not the end of the world. But I want to get it out just to draw a line under it. And in particular, since I am making some breaking changes, I would kind of like to delay ULAM 5 even yet a little further and consider some other possible breaking changes so that maybe we could make the pain at least be finite in duration. All right, that's it. Reaching out, reaching up. We'll see how it goes. I know I'm supposed to be working on the tiles. I will continue to work on the tiles, at the very least designing how to make the next step as far as making it more robust. But I need to pursue Plex. That is feeling a lot of possibility. Right now, the level two plate is looking pretty robust, but it's not moving yet. So that would be the goal that I personally, secretly, if I'm allowed to do stuff that's still on the simulator rather than on the tiles, I want to see L2 plate moving by next time. We shall see. Hope everybody's doing okay. Thanks for stopping in, again, here live or checking in later. Get to see my incredible embarrassment as I flail around, pushing the buttons on OBS. I feel like this is part of an arc that I'm sort of going through about learning how to go live. In my mind, there's always been this huge distinction between writing and speaking. And over the years, as everything's gotten recorded and zoomed in all the stuff and since the pandemic especially, but even way before that, the distinction between what is the difference between speaking and writing, really. And I now sort of understand or believe that the difference between speaking and writing is that it's speaking if there are witnesses. It's speaking if there are folks there that could, at least in principle, interact live with you. That's what makes it speaking. And that's why conversations are sort of automatically kind of interesting in their way as long as they're legit, as long as they're not scripted behind the scenes. So I'm going to keep on practicing doing these things live. Hopefully I'll get sort of smoother going forward. But you know, hey, it's the failures that drive the plot. It's the failures that make it drama. Thanks for coming in. I hope to see you in two weeks. Take care.