 Hey folks, it's T Tuesday. It's been two months since my last update. I feel like I missed miss you folks. I Think two months might be a little bit too long. I Feel I feel rusty. We'll see how it goes So the goals for today serial port concentrator demo. I failed at that. I have a bunch of pieces I'll show you what I've got And also von Joy Manon's collected works was supposed to be published. I failed at that as well Once again, I have pieces And I was supposed to have a huge fun and I kind of did that I mean it was it was family stuff and a lot of cooking and and just general fun So No foul. All right, so let's talk about this stuff So we'll talk about the collected works first. So This is my Place filler cover at the moment von Joy Manon collected works Since the von Joy Manon is a pseudonym. I figured the thing to do was to make it as absolutely as big as possible Here is the I don't know if it's possible to read this. This is the the contents as of now. It's far from done We've got the And just to cut to the chase here is search quiet wake. That's the science fiction story that I Wrote a couple of nano-rimos ago or last nano-rimo something like that That I actually sent out to a couple of science fiction magazines and let them turn it down That's gonna be essentially the only Substantial new thing in this whole book The path the best effort is the paper from 2013 and nine written by von Joy Manon and The living computation theory of everything is the transcript of the video over on the day back the channel And this is in fact move where a ton of time went to I was telling myself, you know I'll just take the transcript, you know downloaded from YouTube and just put in periods get rid of the Us and it'll be fine. It'll be ready to go But as soon as I started going through it, it's like I couldn't do that It's like, you know, okay, you know, there was a jump here. I Interrupted myself before finishing the sentence, you know, whatever it was So I actually needed to go back in and kind of if it was gonna be readable as an essay and that was the idea I Had to I had to start working on it. In fact, I'm still working on it So and I don't exactly know when I'm gonna get back to this stuff I want to keep pushing it as best I can but flipping and back and forth between hacking head and English head is a bit disruptive So we'll see how it goes the main goal right now the main goal for December for the next to t Tuesday update is Get hacking absolutely get hacking. So in that context the serial concentrator design So just to remind us We are now in the t2 brain challenge where 60 days into the t2 brain challenge and really don't have nothing to demonstrate But the big idea was instead of having the t2 matrix be this isolated universe that ran chemistry and biology all by itself We were gonna admit that we're just gonna pick a chunk of it a finitely scalable chunk of the matrix and connect it somehow to some external World or simulation of a world and that was that that is the plan and The idea of how to connect with each t2 tile has a serial port So the idea was to somehow concentrate all the serial ports down and feed them into a laptop Run a simulation on the laptop that represents the world and go from there And last time where we ended was the idea that we could actually use beagle bone greens That's the same system that is inside the t2 tiles already and each of those at least in principle could hold Could could handle up to six you arts And so this was the idea and after the last update in October I Started heading for this and Didn't go so well You know so this is 19 tiles, that's one surrounded by six surrounded by 12 That is a lotus of the basic unit of tiles that we've been assembling a single lotus has a single power supply And that is what's gonna take At six you arts per it was gonna take four Not all of four, but three plus Beagle bone greens just to fan in those 19 Serial ports and in the process each of these lines is gonna be four wires a transmit line in the ground a receive line in a ground So we are talking 76 wires that have to go somewhere and all of that is Multiplied by four because we've got four lotuses at the moment on the grid and I still wanted to make it bigger So just to set the physical aspects of this This is the key master, but it looks like all the tiles and this right here is the debug header Which can has the pins for the serial port for the tile And then the thing to notice is number one it plugs in from the front That was very deliberate because the purpose of it was to be able to debug in the middle of the array But that does mean that the obvious idea of taking these You know zillion wires and tucking them around Into the back so that they won't block the The screens on the tiles and mess things up visually Is a little bit harder to do so that has to come out this way It could turn and then head down But in this picture if you can see it the the gap between the two yellow arrows That's the clearance between one tile in the next the blue there is the circuit board of a tile in the circuit board of the next one There's not a lot of room you could feed a flexible a flat cable flat ribbon cable through it You might be able to feed some wire, but it would be ugly and then you would have the same problem behind the grid It was a big mess a big mess and I was thinking about all sorts of things What do I want to do with this and then finally a couple of weeks into October the pin finally dropped for me What about doing daisy chain loops through multiple serial ports and the idea here was this whole thing runs Incredibly slow and that's because the t2 tiles are prototypes And I don't do hardware and I don't do anything, but you know there it is if they're running And the the base rate that we're capturing data is one frame every five seconds That's when the camera the Fuji takes another frame and so that's going to be the heartbeat of this simulation and so If we're saying all we need to do is to get sense data from the world out to all the t2 tiles That want to see it and get back motor commands that want to apply to the world for that one unit We have five seconds to do it And the the the serial port rate is you know a hundred thousand bits per second basically or divide by ten Packetize everything Send the packets around with a hop count and then we could chain through a loop So instead of having separate wires going to each t2 tile We have a wire going from one and one to the next and one to the next and then back around And this is the picture of it. So We have a us a USB to you are it a USB to serial port connector I've got bunches of those for doing debugging already But now the idea is the transmit line would go to the receive line of the first t-tile t2 tile in the loop And then if it was if this packet that comes through there was meant for it It would handle it and send the response out the the Transmit line, but if it wasn't for it would just retransmit it to the next tile So the transmit of this to up becomes the receive of this one transmit receive Transmit receive all the way around until a loop So each tile can talk to the to the center the world Axis but only by going all the way around the loop and I did a little bit of arithmetic and if I didn't screw it up Then it says we could easily have 30 to 40 tiles in a single loop and Still get a packet out so that everybody in the loop could see fresh sensory data And if they wanted to they could return fresh motor commands And still have enough time to finish the basic heartbeat So two or three loops with just pairs of wires snaking all around would be enough So that was pretty good and I set down to try to make some prototypes just to scratch them together I got some board. I cut them up into little bits that could fit by the serial port lines I got some right angle headers So the idea is we'll put those on to the board and then push the header here down on to the pins And that's how we'll make the connection. I took an old serial old Ethernet cable and I cut it open to get the twisted pairs to send the signals along And the blue and white arrow blue and white wire was the most tightly wound and I didn't even actually know this Did you know this that you know in a typical Ethernet cable? There are four pairs of wires and the pairs are twisted to help resist Interference from electromagnetic stuff, but at least in this cable and I think in general each pair is twisted a different amount of twists So that they'll avoid resonating against each other So in fact the blue and the white is the most tightly wound in this particular cable I had no idea about that And so I built them. I mean, it's there's no electronics There's just you know, take the take the Receive wire and send it down to the appropriate pin and ground wire and take the transmit wire back and so forth These don't have a heart attack. Oh my terrible soldering and so I made a few of these So here's an example. This thing is now plugged into the debug port on a tile Here comes on the green and white here comes transmit from upstream on then on the brown and white There goes transmit to the downstream And there's an example where between both of them all hooked up. I Can't show you this working. I didn't get it working. I got this far and then switched to fiction Okay, but the other half of it is okay, so say we can do the serial concentrator What is it going to talk to what is the world that we're gonna live in and I didn't really know much about this And so I went to the brain trust. I went to the t2 tile discord and Well, I'll come back around to that So the the idea for all of this is to do the there's a famous a nerd famous book called vehicles by By by Valentino Breitberg. Yeah, this is a copy here That the idea is that it's a lovely little book the introduction let the problem of the mind Dissolve in your mind and he goes, you know, so he's a biological Cybernettist or something and he as he was working on studying the brains of animals I felt knots on time distinctions dissolve difficulties disappear What that I compared to my first naive approach to the problem of the mind and you know I feel the same way I feel like as I've been working on the t2 tiles if I'm working on Distributed robust local synchronization best effort all of this stuff And staying close to the physical spatial computing spatially mapped That all of these things that seem perplexing about how the mind is supposed to work They dissolve into the physical and that's what the Breitenberg vehicles book this book vehicles is trying to get at So it has this very abstract notion of a little car a little vehicle So the car is the rectangle that's got two wheels in the back, which are meant to have motors It's got two sensors up in the front that are meant to sense light And the idea is is you can connect the sensors to the motors in a variety of different ways And then get different kinds of behaviors So for example if in this one in a if we have a bright light on the left Sensor then that means the left wheel go faster And if it's a bright light on the right then the right wheel will go faster And the net result of this is that a car a vehicle on this design will automatically turn away from light because The brighter light causes it to turn away But if we cross the signals as here then Vehicle B will turn toward the light. So if it's brighter this way then we Drive the opposite wheel and turn towards it And the point of this is is that we can think about it in terms of Desires, you know and Breitenberg just goes for it, you know, this this one is afraid of light This one loves light, you know, he's just goes right ahead and uses mental terms emotional terms to describe this stuff Even though it's completely simple trivial I want to build these I want to build them starting with the simple things but then carrying on to more Complex and programmable ones. So I went to the discord. I said folks, where how can I get these Breitenberg vehicles? folks Tim Teague's suggested pie bullet, which I looked into AJP Kim suggested this thing called Mujoko, which I had never heard of And after going through it for several days, it seems like Mujoko is the what I'm going to try So in fact, here it is This is my little Breitenberg vehicle or at least it's part of a Breitenberg vehicle All of this is pieces. None of it is hooked up yet but 2024 So that is it The uh, uh, so hardware source serial port concentrators failed at that but may have got pieces Software protocol. I did I did implement packet stuff and so on, but I haven't designed actual packets yet So next time deploy parallel interface. Well, this is the official schedule But what I want is a single loop interface an actual thing that daisy chains through a bunch of tiles I don't care how many uh, actually running and a Breitenberg car with sensors and stuff running in the simulation They don't have to be connected to each other We'll see It's good to be back. I hope folks are okay and you know, happy holidays. It's gonna be 2024 before the next update I'll see you then