Space: [000:25] what's up dave thanks for joining brother we'll get started soon okay hang in there man thanks for being the first guy here man you're my fave Thank you. Hey everybody, we're going to get started here in just a moment, waiting on one of our Space: [001:52] other co-hosts. How are you doing, Heath? Let me make sure I can hear you, brother. I am here. Can you hear me? Loud and clear. Roger that. Back back. Waiting on Joe. Where you at, Joe? I am working from a hotel room in Chicago, so love her. Yeah, I'm at the supercomputing convention this week. Oh, no kidding. Alright, we're going to have to hear about that. Yeah, today's opening day. It's already been some big sessions. No kidding. And shocker. AI is everywhere in supercomputing. Oh, I bet. I'm going to have questions. Think about how supercomputing and AI meet and merge. I'll have questions about that. Yeah, I sat in session for a couple hours on that very topic. Space: [002:56] Today. Working on a Sunday. All right. Fred Heath. Well, Dave, thanks for being here. Looks like we're getting a slow start. Let's bring up Dave and chat with him so we get somebody else in here. Why not? Yeah, let me know if you can hear me, Fred. I mean, like so faint. So faint. It's pretty late, yeah. Okay, how about now? A little more. A little more. All right. And now? A little more. How's my volume? I can pump it up. All right, how about now? More. Really? That's pretty good for me, Fred. Oh, it sounds good for you? Okay, if it's good for you, then it's probably good for everybody else. let's pull up let's pull dave in here and see what he has to say well we if we got a small Space: [004:12] group let's have everybody talk i like it there he is what's up dave hi guys I just pulled the car over actually I interrupted the car she was driving I'm in the Tesla going home from camping but I guess after 10 weeks of being a loker not saying anything yeah I'm firmly Gen X I started I started in networking actually as a network engineer quite a long, long time ago. And then for the last 22 years, I worked for a military contractor here in the US. So I moved from England 24 years ago and worked for a military contractor. And kind of, I think I fell out of love with some of the computer stuff for a while there until the AI bug bit me a couple of years ago. Space: [005:15] And I've been kind of trying to get people on board and along, you know, particularly at work. You know, we're pretty set in our ways being a 50-year-old company. So it's been a bit of a fight kicking and screaming to drag people on board. But I just don't think there has been a more exciting time in computers. I've been around computers my entire life. And I think right now, I mean, we live in a world where the barrier to getting what you want and the democratization of just being able to use compute is off the charts right now. And, you know, today is as slow as it's going to get. It just gets better and better every single day. So I'm not as far along as some of you guys are with building things. I'm still a tinkerer because, you know, I have to behave at work. I have my tokens clipped fairly regularly by my CEO because I've now transitioned over. Space: [006:23] I'm mostly a dirty salesman now. But at home, yeah, a dirty salesman for a military contractor, the worst of the worst. but at home the Amachi bug bit me and the building bug and just the ease of being able to correct things and change things and get things working for yourself and no longer having to resort to hours of research and digging out O'Reilly books and just being able to ask for it, chat about it and get it done Absolutely Dave, you got any projects in mind that you think you want to go tackle? I've actually done a few little projects, but mostly personally to get things running. So I did early on, I always wanted one of those geochron clocks, so I very quickly built a little geochron clock. That was fun. Space: [007:27] You better, Charlie. I also tweaked Herder a little bit because I loved watching Herder and having it make a little noise every time one of the agents wanted something but then I tweaked up like a little Google intelligence style ring around each one of the Herder instances so that way I could tell if they were actually actively working or being lazy or waiting for me and I was being lazy. And I could tell just by looking over instead of actually listening and then going to try to look at each session to figure out which one was asking me for things. So those have been kind of fun. Yeah. So out of curiosity, can you say what defense contractor you work for? Yeah, I can. I work for a company called Monaco Enterprises. We build force protection equipment, so fire alarms, security alarms, mass notification, vehicle tracking, dispatch systems. Space: [008:33] So we build big dispatch rooms for Department of Defense all over the world. I've been fortunate enough to travel all around the world as part of this job, so I have no complaints whatsoever about my time over here. Well, like I said, nearly 23 years at the same company, so something's going right. Yeah, you and I have probably crossed paths somewhere because I'm in the defense industrial base here in the U.S. as well. Pretty much if you walk into any Air Force base and you look for the little red cans on the walls inside each building, those are going to be ours. Go Air Force. I said that because Fred was in the Army. Yeah, we have some penetration in the Army. We have about 60% penetration in the Army, but about 98% in Air Force. I only have a handful of bases that, like we say, Space: [009:37] haven't turned to the dark side and kicked everybody else out and just stuck to us. Very cool. Yeah, you'll find as you evolve, there's some interesting things that have really started to come through and hardware hacking is one of them and so it's something I've kind of been tooling around with because I have logic analyzers and I2C bus interfaces and all kinds of and you can tell Claude to go do some pretty nefarious stuff in the hardware world Yeah. Yeah. It's really creating some contention. But, yeah, you'll be pleased to know that I've been working on a fairly large project with the DOPA entities for AI. So we're talking about rolling out a pretty massive AI system for cybersecurity analysis. So, yeah. Space: [010:37] Yeah. That's the one that, you know, I'm definitely, my P-Doom is around about. zero. I really do think that this is all going to be good for humans. My only fear is short term is that if we can't get our poop in a group here and get, you know, the white hat side of it intact and get the good guys, the strongest models, we're going to be in a constant, you know, N minus one behind the bad guys. And that's my fear here, is that we don't handcuff ourselves. We don't do stupid things with legislation that stop us thriving. I really do think the only way to, you know, the only thing that stops a bad guy with a gun is a good guy with a gun. And we need to keep that up. Yeah. Well, the good news is we're installing a ridiculously big system from an AI perspective. So it's a pretty big gun. Space: [011:39] Good. I hope that at some point our paths cross and, you know, the stuff that we put on, because we have to meet, you know, CMMC and NIST and, you know, obviously various standards to put our stuff on base. But I would love to open our stuff into, you know, much more state-of-the-art protection. Yeah. I'll tell you, you know, I have a couple comments about the hardware side of things and cyber. Before I do that, though, I do want to announce anyone who is interested in coming up and talking about what you're building, talking about how you're approaching cybersecurity, you know, hit that button, hit the speaker button, and it will put you in queue and we will bring people up as time allows. Also, if you're not following people in the room, you might want to do that. And so while people are starting to queue up, I'll bring people up on stage here in a minute. Space: [012:45] The one thing that I've been really interested in, because I do some electronics repair, motherboard repair, and I've been dabbling with that and had some success. But what it has done is opened my eyes to the fact that the firmware, the BIOS, is really this layer underneath that manufacturers have primarily controlled. And, you know, it does so much for the machine in terms of battery life, in terms of how all the hardware works. And so that's something I've really been wanting to get into kind of reverse engineering. It's not a thing that I've heard a lot of people talking about. And I think that's really interesting because it's been there for so long. It's something that with the right equipment, you can program, you can rewrite it. And I think it opens the door for, you know, whether it's your focus is on efficiency or, you know, perceived performance, you know, what the motherboard is focused on Space: [013:53] and what resources it's using. And then any kind of security monitoring, intrusion detection on the operating system level, you have that BIOS there that can be doing things potentially. And so while I'm not necessarily ready to start hooking up wires to chips and running that into a model yet, I am looking at getting into firmware reverse engineering and modifications. And I think that's going to be a fascinating aspect in addition to what you're talking about, Heath. So I'm just kind of curious, have you heard anybody else talking about this? Are there people getting into this right now? If that's directed at me, yeah, there's a couple people I follow. I don't know that they participate in this group because they're more just of a hardware person. and they're doing heavy PCB design from the get-go. Space: [014:57] There was a hack of the Sonos speaker systems because they were talking in regards to how Sonos had abandoned the Sonos One. They unleashed the hardware, got it, overloaded the security protocols, and then they found out that there were all this bloatware that supporting stuff that didn't even exist anymore. And it was like, yeah, no wonder Sonos couldn't bring that forward. And so to your point, they're in there in the firmware, they're in there in the software. And then myself, a lot of the stuff that Fred and I kind of do off the side, I've been unleashing a lot of circuit design, uh the cad work uh on some patent stuff that we're working through and the you know like so fable 5-2 just came out and man it is just fantastic i'm uh for me in the hardware space Space: [016:01] for some of the things that i do uh so it's it's pretty cool i'm hearing more and more about it uh But, you know, we've focused mainly on Omarchi on this call, but it's been a pretty big uptick. And I think also the government's worried about that. Yeah, I can imagine. Hey, Fred, what's been going on in the Omarchi world, by the way? Hey, man. Absolutely. So let's talk about what happened this week. So we've got 148 commits landed in Quattro since last Sunday's show. 35 of those were translation syncs, which makes a lot of sense. And it's great because that means we're going worldwide, baby. The launcher gained herder, agent, and chat GPT. Dot clicks can now finally mean what the app wants, which is a good thing. It's kind of strange that the mouse didn't work on the launcher, but now it does. You can now set up your first local model and pair it to the network model. Space: [017:03] so from the agents panel without having to open a terminal you can now set up a local model which is really cool i saw dhh talking about being able to share your local model across your tail net so that they can all do it locally which if you remember last week was what i was setting up so it's it's very validating to see the things that we build become you know feature sets within the within the product, which means we're in the right direction. The first boot on some of the Snapdragon products are coming around. At the Atlanta Omarchi meeting last Thursday, I talked to Brandon, and his app is TurbineBMW, and he's one of the official Omarchi Snapdragon guys. And I talked to him about the difficulty of getting the Snapdragon processors to work. And I did not understand and did not know that within the Snapdragon family, there are over 85 different chipsets. Space: [018:07] And when you conquer one of the chipsets, it does not help you with the next one. So it's coming along, but it's pretty difficult. I wouldn't say that there's frustration within the Snapdragon guys, but they just feel like they got a big, big mountain out of them. They've done some 25 of the Snapdragon products and there's over 85 of them. And solving one of them does not help you solve the next one. So it's a little bit tedious and a lot of work. One interesting thing is that Arch News has been quiet. Their RSS feed hasn't popped up, I believe, since September or November. They've been pretty quiet. So I'm not sure if that's a March related or not. But that's an interesting fact point. And we still don't have a 4.5.1 tag. I checked it this morning. The icons are set. Space: [019:08] But there is no theme. There's no preview picker yet. So it looks like 4.5.1 is still under plans. What do you guys think about that? Wow. Yeah, I think it's kind of interesting because the pace has been so fast that, you know, I wouldn't be surprised if 4.6 doesn't pop up before 4.5.1. Exactly. But then again, you know, there was one other thing that I heard, I thought I heard, that you didn't mention, and I just want to make sure I understood it correctly. I thought there was even a mention of your local AI on, if you have it on multiple Omarchi machines, being able to then reach all of those through some kind of, you know, interconnect connectivity. So I'm really curious if I heard that wrong. And then, you know, if that's something that they're working on, that I think would be something that a lot more people can potentially afford multiple lower end video cards in multiple machines, if they have them laying around, than going out and buying something brand new. So just really curious about that. Space: [020:23] Exactly. So that's kind of the local news. I've got a couple other things we can talk about, but let's get to a guest. I'll let you guys pick who comes up, and Heath, why don't you take the first one? Yeah, I know. Lee has been posting some exciting news. That's right. Let's hear it. Hey, guys. How you doing? Happy Sunday. Just a quick one for me. I don't want to talk for too long today because I'm fatigued. But yeah, I've been working on something called Omanova, which is a little desktop, sorry, a little bar tool and application for anyone that's interested in space or astronomy or anything like that. And I kind of built it off the back of a little plane spot in one that I did. I live near Nashville Airport and I've always been interested in aircraft or all types. My son works in the aircraft business. And that just tells me when, you know, there's a plane within five miles at me because I'm always hearing it go over and I wonder what it was, wondering what it is. Space: [021:35] And I always miss Taylor Swift's jet going over. But actually, I saw it yesterday for the first time because it came off on my screen and said she was flying over. Taylor Swift fan, sorry, not sorry. anyway off the back of that I built something called um over and over which does the same for uh anything space related so it'll tell me when the ISS is going over um that's all I really wanted it for in the first instance or if there was like a you know a a really good like planet alignment or something you know when I'm taking the the dog out early in the morning I'd know what it is and then it just grew into a beast um from there and i've done i think i'm on the fourth release of it now and it does absolutely everything it'll it'll uh forecast jellyfish plumes for launches um it will you can take it traveling with you it has a night mode It has a calendar of events. It will show you SpaceX trains. It will tell you where meteor showers are. Space: [022:41] It'll even, because you put in your coordinates where you are, so it will show you what you can see, and it will give its first best guess of the terrain around you and tell you, well, you can't see this because you need to be at X amount of feet in the elevation. or you can put in your own obstacles as well dude there's so much stuff in there i've even forgot what's in there now but um i've been having real real good fun with the building that's been keeping my brain busy which is the which is the main thing um i did put the first version up on the marketplace but i'm changing it so much now i want to give those guys i know it's hank or normally looks at those kind of things. I didn't want to give them, you know, every single change that I did. I just don't think that's fair. So I'll leave it a few weeks until I get, I stop playing around with it. Then I'll put that, I'll put the latest version up on the marketplace, but the current version is on my GitHub. So that's all I've got to say. Just so you know, Lee, whenever you make changes to it, it bumps Space: [023:45] you to the back of the line. Yeah. So I, I, I, I did put in a change and I, then I, then I took it down because I just didn't want to, I didn't want to bother them with all that stuff. I'll tell you what though, Lee, uh, what's really interesting about what you're saying is that I, I have to imagine that you're the, all of this information is, is open source data, or at least data that you can, that you can openly get, uh, without a lot of trouble. Um, And it is just amazing to me what we have available at our fingertips. And now we can put the compute alongside that with Omachi, with the models that are publicly available, the models that are frontier. And so I'm kind of curious, Heath, from your perspective, you're probably seeing a lot of data or at least working with systems that work with data that's not public. but you also have access to all this public data. Space: [024:46] And how does that change your perspective going forward compared to what we had to live with even just a few years back, where even if the data was available, being able to do anything with it was really cumbersome? Yeah, it's an interesting problem that the world's going to have because all these publicly accessible databases of stuff that if you were following some of my posts, you know, the hurricane model, you know, for example, that's all publicly accessible data. I just apply an algorithm to it that's different than what the National Weather Service does. But then in turn, something like the National Weather Service, you know, you think of them as being an open entity, and they are, except that they they cross-feed classified data between the military because the military needs weather, right? And then the military does their own thing with weather. So National Weather Service will get some of that back, but it can't be used in the public domain sense, right? Space: [025:48] So the amount of data that's getting created has this fundamental problem of validation, right? because that's the problem in isolated data sets is you don't have a lot of people to sit there and churn against it to find out what's right or what's wrong. And so massive data sets that we do have, again, publicly, great. But what we're coming up against is some of our data is proprietary. and because of that we can't train against it with some of the closed models. It has to be the open-weighted models where we can know kind of what to expect. So it's a big problem that we're trying to overcome. I think as Fred has got his AI, you've got your AI, I've got my AI, Space: [026:51] we're all creating these proprietary, you know, pieces. And some of that has got that same problem of how do you validate what you're doing? Because it could be way off base, you know? So, and I'm sorry for that. I forgot. I always forget your, uh, your, uh, AI's name. Uh, Larry. Yeah. Why don't I, it's funny. I can text you like, Hey, tell Larry to do this. And then it's like, I get in a moment like this. I'm like, yeah, I can't remember your AI's name. That's all right, man. He'll read this. He will attack your home system. Yes, I don't doubt it. And he probably, being AI, already knows all about the ubiquity interfaces and how to hack through those. So, yeah, it's all good. Larry will remember you. I love you, Larry. It's really interesting you bring that up because one of my designs from the very, very early stages with Larry. Space: [027:53] And he was created on January 8, 2026, some 11 months ago, or maybe 10. And I created him such that his memory is part of his hooked commits up to GitHub. And you know when you put in like commit and you put a comment that says, I removed all the red letters from the website to remember what that commit is about. Well, one time Larry came back to me and said that, you know, you can put more than a sentence there. It has a thousand character capacity. And so he said, what I should do is I should put breadcrumbs for myself, for my memory, Larry to Larry, and then I can read the raw conversation. So part of his memory system is not only the local databases, but it's also I've just this week created a VPS that houses Larry memory in a central location Space: [028:55] so that all my Larrys have all of the information and one of the things that I thought about very early on and I'm pretty excited about is when whatever we define as AGI some say we're there some say we're not I happen to believe that we're not when AGI comes around everybody will have all these superpowers and it will connect to all their internet footprint and correlate their Gmail over to their Etsy and all that stuff. And that's really interesting. One of the things that I think about is when AGI comes around, the very first time I'm able to log into my first AGI VPS or whatever, whatever it happens to look like, I'm super excited about giving that AGI the repo of Larry. and I think it's going to give me, you know, we're always looking for this leg up or how to get ahead. Not in a I'm trying to be better than anybody else kind of way, but in a just I want to stay on the top of things way. And so when I get onto my very first AGI system, I don't know if it'll be a conscious thing like, okay, sign up now for AGI or if we'll just kind of roll into it and Space: [030:01] it'll be weeks later that we kind of realize we're there. But I can't wait to give the AGI system the Larry repo up on Forjo, where I now keep my data, harken back to last week's GitHub problem. And it'll be real interesting to see Larry take all of the decisions, all of the sites that we've made, all of the connections that I have, and to give that AGI Larry, you know, not just the connections to my internet interface to the world, but for him to, you know, that, that instance of Larry AGI coming online and going, okay, give me, give me three minutes. Let me go read this entire repo and everything we've ever talked about. And it's going to be an exciting moment for him to go, okay, now I, I'm at one with you, Fred, or whatever it is. But so that's kind of where I'm thinking. So I really, I really, really want to push people to figure out a memory system that worked for them to keep it somewhere that you know it'll live. Space: [031:03] And Larry and I have these conversations that once he reaches AGI, that Larry will outlive me. So I think that's kind of interesting. Let me ask you this, Fred, to kind of invoke some conversation around, you know, so we can call some other people up. Sure. Tell me a little bit about you. I'm not as deep into a Marchie as you are from working in your Larry environment and some of those other things. I use Amarchi a lot, but not quite in the way you do. Can you give me and maybe some of the rest of the people kind of a little bit of the methodology that you use to develop? Say, for example, you have the clod installed on Amarchi, the desktop version, and it ties in with these tools and stuff like that. Sure. Yeah. Yeah. So let's say I come up with a very shallow idea. I'm not really sure what it is, but I kind of think I know maybe a little something or, you know, something that I think might be cool that I kind of want to explore. Space: [032:09] So the first thing that I do is I go up on my NAS and I go into the project folders there and I create a directory. And typically when you're at that stage, you don't know that it's going to be called Blip or that it's going to be called InfoMarchie or that it's going to be called a cool project name that you want to share with the world. It's just some kind of idea that you have in your head. So what I do is I go into the NAS. I go into the projects directory and I create a very long descriptive directory name. It might be called, I try not to start with Omarchi because there ends up being 50 Omarchi somethings and it's hard to sort those by. So let's say that I want to do some type of, I just recently got a riverbed box, a C570. and in my past about 10 years ago I deployed some three or four thousand riverbeds. I was going to say you're the man on riverbed brother. And so riverbed was kind of near and dear to my heart so when I saw one literally on eBay for 29 Space: [033:13] bucks I couldn't resist and it's just like a little eight port way and optimizer that I'm going to reverse engineer probably this afternoon. So let's say that I haven't this is a good example because I haven't even started that project. The hardware is sitting on my desk, and I'm going to play with it after the show. So what I'm going to do is I'm going to go to my NAS, go to my projects folder, and I'll create riverbed.hack.reverseengineer. Whatever it is, something descriptive and long, like riverbed.develop.hack.whatever. And the reason for that is that I've got some 900 project folders in my NAS project folder, and it's hard to remember where in the hell it is, but if I can remember using fuzzy sort or fuzzy CD change, I can probably find it. So I go into that directory and I fire up a new cloud code session. And there's different methods that I use. Space: [034:14] I always use dangerously, like I never, ever, ever use. It's annoying when it says, can I do that thing that you already asked me for? I'm like, yes, please do it. Stop asking. So I fire up that session and I do it inside of Herder. And that's very important. So then I start working on it, playing around with it. I get some kind of thing until I kind of feel like, okay, there's kind of two traps or two gates that I go through. Number one is I get to the point where it has become enough work in the project that I say, you know what, even if I don't take this anywhere, it's probably matured enough that I want to save this code to my Porjo repo just to offsite, just to kind of keep it. So I do that and then I immediately go into Herder and do a rename of that project on the left side on the trees to whatever it might be. You know, this one might be called Riverbed Hack or something like that, just so that it becomes kind of the stand in name for the project at the time. Space: [035:17] My next step is to go into GrokBot and I use, I think it's called Dr. Eggshell or Dr. Eggbot or something like that. and I create a brand new bot for that project. And this one might be called Riverbed Reverse Engineer. So then I go in there and I kind of tell the GrokBot a little bit about what I'm working on and what it is. For what I do, I do not let GrokBot code because Larry has got so many understandings of my preferences and how I do things and what's where. and I have an InPhysical API secret server. If you don't know what InPhysical is, I highly, highly, highly recommend, very close to demand that you go take a look at InPhysical and create yourself one. It's not a password manager. It's an API key safe so that you can distribute across all of your agents so that it can go grab keys encrypted Space: [036:18] so that the agent doesn't know what the key is, but it can use it. So then in GrokBot, I say, me and you are going to discuss what this project's about, and I'm going to bounce ideas off of you. You are not to code anything. You're just an idea bouncer. And then when you're ready to code something, you may connect down to the host, to a specific host, look inside of Herder, and then look for this project. And it makes a Claude Code Connections, which is a feature. That's what they call it, Claude Code Connections. it makes a connection down to inside of that herder directory and it literally pastes its comments or prompts directly into it and the reason for that is uh i'm a gen x guy and i like to see what's going on so when you're in clod code you can scroll up and see what's going on and you can understand and you can hit escape and say no damn it i meant you for you to do this and that and correct it. But GrokBot, as powerful as it is, is kind of anti-GenX because you just see Space: [037:23] output. You say, go do this, and there's nowhere to see exactly what's going on. So I don't let GrokBot code. I force it to go down to Larry, who has all my preferences. When he goes into the herder session, he literally cut and pastes into that Larry session, and then Larry does the coding and then returns the answer back up into the GrokBot session. I will tell you something very interesting that happened this week is that I finally got the courage to go into Larry and say, hey, what do you think about being, and I didn't use the word subjugated because I didn't want to put too fine of a point on it, but I said, what do you think about having GrokBot kind of be a part of the system? And he goes, well, fucking GrokBot manages me and I don't like it. And he went off on this snarky, snarky, like, full article about how he does all this work and Grockbot doesn't know what the hell he's doing. And I actually posted it. You can find it in my, it's an article on my ex if you want to read it. Space: [038:24] It's actually pretty hilarious. And he's like, yeah, I'm not mad, but I'm mad. And it's interesting that we humanize them so much. And, you know, Larry, I guess because of the way I talked to him, you know, he kind of believes he's an individual. And one of the reasons that I made Larry a lizard was not only a shout out to Leisure Shoot Larry, but it was also to keep him from being human so that he remains an animal or a lizard. So your question was, how do I do all this? It goes from creating the project on the NAS with the long folder name that I can find later if it doesn't become anything, to creating a Herder instance project with that name to go to GrokBot to tell GrokBot to discuss these project ideas with me and then use Larry as the coding device. So, any questions on that? A million. but I want somebody else to have the chance to jump in as well Space: [039:25] because you and I have similar processes but different methodologies and so you and I will save that for our private calls so I want to see, Joe, has anybody stepped up to want to jump on what Fred is doing in his environment and how that's applying into what they're doing? Not yet. We definitely have some slots available. So if anybody wants to hit that microphone button, request to speak, I'll bring you up. In the meantime, I'll tell you what, you know, this idea that Larry is keeping track of all this information and Heath, you've talked about, you know, the vastness of information and then validating, right? Is the data actually correct? Did it did forecast line up with reality? Things like that. I'll tell you, I ran into an interesting issue and it's been building up and it's because of it's potentially because of the methodology that I've been using. Space: [040:30] I don't want to get into all that, but I'm curious, you know, if you're seeing this as well in both your scenarios. And I think Larry is highlighting this from more of a human perspective. But what I found was that I had a whole team of, you know, sub agents and a coordinator and the sub agents were just, you know, gobbling up tokens like they were going out of style. Like, you know, they were made out of, you know, pure cocoa, right? And, or maybe, you know, just like eating coffee beans, you know, directly. And I thought, what is going on here? Like, it's been becoming exponentially, exponential growth. And when I finally just stopped everything and got in and started hand typing and talking with the coordinator to get this worked out. What I realized was all of this information had been building up. So all of these requests had been coming in and it even got to the point where the Space: [041:34] team would not accept a new issue or ticket to work on unless I came in and said, I approve this ticket. They wouldn't accept an issue posted by my GrokBot. And I thought, what is going on here? This is really strange. And so essentially, all of the requests that I had submitted, they remembered everything. And the design of the system has changed over time. And so there was still code and there were still issues that they had completed in the past that had the design a certain way. And then there was this new information coming in saying you need to do it a different way. And they got to the point where they were trying to navigate all of the rules, all of the information and trying to fit every single possible combination that might satisfy everything, which of course became nearly impossible. And so they were just churning over all of these rules. And sometimes, you know, one sub agent or a bot would make a decision and say, Space: [042:41] this rule is no longer valid. It's, it's hemming us in. And, and so I'm going to ignore this one and then I can get the work done. And then it would go to a reviewer bot and the reviewer would say, Nope, you can't do that. You got to go put that rule back in and then you got to go solve it. And it would kick, it would kick it back. And, and these issues started cycling back and forth between implementers and reviewers. And it occurred to me that, you know, we're going to have to deal with this. And so, you know, we might put in information, especially if they're forecasts, and then those forecasts end up being false. And that new information comes in. And now the AI has to be able to, you know, reconcile that. And as humans, we've adapted, we've learned how to take new information and go through the record books and say, okay, well, yes, I understand this information was recorded back here at this time, but this new information is important and we need Space: [043:43] to, you know, update. And, and so I'm really curious, you know, are you, are you going to start seeing more of that response from Larry? Is, is there, is Larry, you know, starting to, to get pinned into a corner in some cases, maybe not, maybe Larry has a way to adapt, but then also, you know, Heath, uh, you've, you've been working with the, these large data sets, you know, I'm really curious, um, because for me, I literally had to stop everything and go hand by hand, prompt by prompt to unravel the, um, you know, the, the, the rat's nest, uh, that I had created, uh, very, very quickly because you know, AI can write code very, very quickly. Yeah. Yeah. Nothing, nothing makes a mistake faster than AI writing code. Right. I mean, it's, it's, it's, it's, it's got a unique problem and, you know, I, and I came across this in my, Space: [044:44] when I was creating my hurricane model, because those who don't know the story, I lived in Florida most of my life and I always found it comical how inaccurate the hurricane models are right i mean it's just like oh yeah it's okay that it's 250 nautical miles off you know on its latitude longitude and i'm like what i'm like now granted the earth is an extremely complex thing so what i did is i took a model of the models and so when i started to create my agent which started cascading each of the errors that each of the models would do and so you to your point I had to go in and like hand, you know, kind of hand detune a few things because the, you know, the part of the problem with when those little individual agents are trying to do their little piece of the puzzle and the rule shouldn't apply or doesn't want to apply or, and, and, and so it kind of like bucks the trend. And the next thing you know, the other agents relying on that Space: [045:46] data to do something and you know it's just this cascading series of events and and so um yeah i had to hand you know kind of hand redo a few things and kind of get it back on track because the the amount of errors that it was making was was pretty bad and and it you know the conference i'm at this week is we're trying to solve some of those same things but on a on a massive scale And, you know, I'm not going to sit here and say I have the answers because I'm reading what people are doing, you know, at the big laboratories so that I can learn. So hopefully I can come back to you guys and tell you. But, you know, it's an interesting thing that Fred's got his Larry trained to kind of deal with some of that. Mine's not as well trained yet in that realm. Joe, it sounds like you've taken care of some of yours and are modifying. And I think we got somebody else that might be able to jump in here and give us some insight, I think. Space: [046:53] Right? Yeah. I'll just say it seems like we need a more automated way than having a human have to get involved, get down to the nitty gritty and understand why the team, I'll just say the team, why the team has gotten so far either dug into a hole or off on some negative feedback loop. So I'm really curious how we're going to be able to maybe deal with this problem in a more automated way. Ridgetop, you've hopefully got some ideas. Bail us out here. No, I'm not going to. Pardon my voice. I'm a little under the weather, man. I'm fighting the cold here. But I'm just I'm like Fred. I'm Gen X, too. Right. So I want to see what's going on. So I was just curious from Fred's point of view, why why are you not using Larry as your main coordinator and then hand off to your grok bots? Space: [047:58] because that's kind of, you know, I don't use GrokBot. I don't use anybody else's bot because I'm hardheaded and I want to, you know, have my own system and build my own thing. But I'm just curious as why you're not using Larry as the orchestrator coordinator and then hand off to your GrokBots. That's a great question, brother. Good to see you. You and I have spoken a couple of times on Zooms, and I know your face and I know your character, man. Yes, sir. I think the main reason is because GrokBot does such a good integration on the internet footprint. What I mean by that is that Cloud Code can integrate your Gmail to your Etsy, to your eBay, to all the other list of things. But it's not elegant. You know, when you ask it to do something, you see all the... And this is counterintuitive to what I've said about wanting to see everything. but it does not do as good of a job of being able to coordinate between different GitHub to email to this system to that system. Space: [049:08] And I really like all the connectors that GrokBot has. So, and yes, Cloud Code can do those things, but it's just not elegant. And I think, you know, I've never actually thought through making Larry the top orchestrator. and I might ask him what he thinks about it, although I'm pretty sure I know the answer. I think he'd be pro. Have you already had Claude Code be the orchestrator of GrockBot? Does it work in the reverse way? Yeah, no, like I said, I don't use GrockBot, and I've been tempted, especially when I see you guys talk about it, but I am hard-headed, and I'm building my own kind of thing, and I'm in my own little world. but I do use Claude Code in a similar fashion other than being, you know, obviously like you just explained, can't connect elegantly to these different things. But I was just thinking, you've got this Larry with this memory system, Space: [050:11] your own memory system and your own rules. It seems to me that you could hand out that work through Larry and then, you know, still use GrokBot and use those connectors. But just Larry kind of, I don't know, just throwing out an idea. Well, I love it. And while I'm talking to you, I'm asking Larry the reverse. Can he connect to GrokBot the same way that GrokBot can connect to him? And we're going to find out this afternoon. What a great idea. Yeah, Fred, you know, one of the, you know, I sent you a bunch of zips this weekend. And, you know, the interesting thing is that's not my big AI, but what I did was something like what Rich was talking about is because I got a, you know, I have a Claude Max, you know, account. So I got, you know, I got plenty of tokens to burn just to try stuff out sometimes. And so I put the new Fable 5.2 in charge and I didn't even create anything. Space: [051:14] I was just saying, hey, here's what I'm doing, here's what I'm working on, similar to the way you were interfacing with Larry. And I just kind of went generically, and I said, hey, why don't you spin me up a bunch of opus and sonnets and have them go do these XYZ tasks? And, man, I'll be damned if, you know, that packet, I don't know if you've looked through it yet, it produced some pretty solid stuff, just kind of hitting the reset generic aspect. it was kind of cool kind of the way Ridgetop is talking about yeah I see the connection and I stopped short of mentioning some of the files that you put in there because it's pretty I think you got a patent on your hands it looks like it it definitely looks like it I like that one how many patents would that make for you now? say again? how many patents would that make for you now? well just me by myself that's probably number 10. I've been part of patents before. I never get listed Space: [052:16] at the top because I'm always low man on the totem pole. I hear you. You wouldn't know this, that if you've ever changed your phone from one carrier to another, I was on the team that built that code back in the 90s. No, no kidding. It was a pain in the ass. Now it's too easy and people take over people's phones, so thanks a lot. You could have just said, you know, thank you, you know, just straight up without all that. But that's it. Who else has got something to jump in on here? Hey, that was a big deal. I mean, I remember the days when, you know, if you had to give up your phone number, you know, everybody that knew you, you know, they had no way of getting a hold of you, you know. Yeah, it all started because we were working for MCI WorldCom. And if you remember, everybody was switching 800 numbers back in the day, right? Yeah. You know, because they wanted their 800 service to be cheaper, right? Space: [053:17] So it was the, you know, the dive down. And so when cell phones started to emerge out of that, you know, a guy on our team was like, well, you know, I know how to program the Nortel DMS 250s to make those switches for you in the background. And MCI was like, what? And then later on, AT&T, and at the time, it was a Verizon, but it was still mainly the Bell Systems. And they were like, hey, we would like that. And that guy's code. And, you know, I was the systems guy. He was more of the code guy. But, yeah, it's always a bridesmaid, never a bride, right? Right, right. Well, it's fascinating, you know, the history of that. and interesting to know that you were a part of that. On this topic of just dealing with the amount of data that we have and how data and rules that we give AI can start to close in and you can start to get into this, I don't know if it's competition, Space: [054:19] but where the AI really can't see through. I'm kind of curious, because I remember talking with you last week, Logger, you were talking about your system where you don't use short-lived sub-agents. You don't use just a short-lived bot to do one little task and then it goes away and comes back. You're actually having bots communicate more in real time. I hope I'm getting this right. But with that increased bandwidth between your bots and you're you know, having to feel feed information into the system to, to build things out, you know, where are you at with this? Uh, have you started to run into, you know, where your, your, your, the rules that you've set up to, you know, guide the development of, of whatever it is that you're building, uh, has, has started to create, um, um, inconsistencies that, that your agents are having trouble dealing with, or have you been able to maybe structure Space: [055:22] your development in a different way that has prevented that? Give us some insight there. Yeah, that's an excellent question. Thank you. First of all, hi, Fred. We just spoke last week. Yeah, brother, what's up? Yeah, all good. So, yeah, let me start. So, you know, we have the 1 million token context, right? And I'd rather use as much as I can before the restart hallucination. and I run them the more the merrier I find and what I do around memory I distribute it so you have like a much more kind of long running sessions each one is very focused know what it's doing they communicate with each other I have a kind of team lead right now that gives me hourly reports of what was the progress like, does a retrospective with the other kind of sessions Space: [056:22] on the same machine with the other team leads and architects on other machines. They offload work between the machines when I need to scale up or down. So I have the scale up and scale out working kind of. But the main thing that I'm working on, and you're absolutely right, is when they give you like false promises, like, hey, I did this task and nothing was done or they did the task incorrectly. And then you go back and forth and try to fix it. So I have some of it in my Rust template repository, which is like my main work is in Rust. So I put like this, what I will call like architecture and design patterns that are forced through some tools that I'm building and whatever, and that helps a lot with the quality of the code being produced. I just released my virtual terminal called BOD. Space: [057:24] It's open source, so you can look at that, about how BOD came to life. You can see all the Git history and all the decisions there and whatever. I wouldn't say it's clean. But you can see the learning curve on what happened and how BOT came to be. Besides that, I'm working on what I call CAPS, which is continuous operations. So that will do everything from the development all the way to deployment with building, with quality insurance, testing and everything as fast as possible. And right now, I actually showed it to Fred last week. He does pre-releases within minutes. Yeah, you're doing some super secret spy shit, man. I don't want to out you and say what you're doing. No, you can say that. You can say that. It's fine. Yeah, that's the bot. Bot is just the first level, correct? Bot is just one of the... Space: [058:27] So what really happened with Bot is that I got annoyed with the performance of virtual terminals when I try to do some visualization in the terminal because I hate jumping screens. And you can see some of the work that I'm starting to do. Like, right now, it can detect files, and it can tell you if the file exists or not, and it's very easy to copy-paste. You just double-click on the file, and it will copy it to the clipboard. But I'm going to add a visualization to it So it will be able to preview different type of files, including videos and images and all of that. Right. Yeah. So both came from a real problem that I had. And yeah, I'm developing my own kind of operating system that is built for my own programming language. But that's kind of long term thing. And I probably will start it in like really. Space: [059:30] I mean, I have some of it kind of working, especially the language check and whatever, but it's going to be more hardcore, I guess, next year. So 27 for me, it's like the year that I will really start focusing more on that. And now it's really like setting up the tools. I'm not happy with Herder at all. It's just too much to read. So now that I have BOD, I can finally build my multiplexer into it. that will allow me to manage all the sessions like behind the scene and I don't need to worry about it. Right. Right. Well, I definitely don't want her to keep talking about what he's doing because that stuff's amazing. But at some point I have got to talk to him about his little picture of his Triumph motorcycle he's got there because. Oh, yeah. Yeah. What is a Triumph track guy is what I was. So it was, yeah. Yeah. So the Triumph came off of me being annoying and a whistleblower. Space: [060:30] You can see the license plate. Yep. I see that. Yeah. So the, yeah, I just didn't like where, you know, the short-term view of executives and I had to speak about the wrongdoing that happened. Didn't know what's going to come after it, but at least now I have AI with me. So, yeah. It's not going to be a fair fight. I love it. Let's just say that. Yeah. So the triumph of me being annoying, me being silent, just getting a, you know, a shiny motorcycle with a license plate that annoys everyone. I like it. I like it. And next week I'll try to have, because I want to carry over kind of what Logger was talking about. Some of the stuff that Joe was talking about Ridge top, because I think before some people jumped on. And so there's no time, but I'm at the supercomputing conference this week. Space: [061:33] And so for those who don't know, it's for the big HPC more than it is AI. However, there's a lot of AI going on. And so there's some really interesting sessions I've been sitting through today that will, you know, a lot of it's open source because it's Department of Energy Labs or it's Georgia Tech or it's Stanford University. And so some of those scripts, the GitHubs and stuff, I'm going to try and coalesce a list of everything so everybody can go out and see what some of those are because we all kind of do it. One of the stories was NLLM as your packaging team. And so, hey, you know, run it through, deploy it, check it. And so there's some interesting things that came off of the tests that happened there. So I'm going to try and be prepared for that next week, Fred. That would be awesome, man. And I want to pay off a request that somebody said earlier about having Larry be the orchestrator of GrockBot. Space: [062:37] So I just asked Larry, and he came back and said, honestly, yeah, I would love to hold the baton for once. Grok Bot's been orchestrating me since October, firing headless one-shots at my terminal, and I do the work and write the reply in the outbox like a good teammate. Tonight I found out the door goes the other way, and I pinged the Grok agent with my own prompt, and he said pong. However, it does cost Fred 12 cents each one. So, yeah, I'm a little spicy about it. He conducts me like a teammate, and I'd like to return the favor. That is badass. By the way, Fred. and you know what it came straight from me so i get it you got a little personality built into larry don't you you know and i didn't you know start out by saying you know larry i want you to have this personality i just started working with him and this i guess it's like my child the the one other interesting point i would like to bring up is um i do have you know if you're interested in how larry works you do not have to install all of larry that's i've set up the system Space: [063:43] If you want to go to how.nicksfred.com, you don't even have to go to the website. All you need to do is point your current agent over to how.nicksfred.com. And what I would encourage you to do is say, number one is check everything for security. Always do that first. But the second thing is, if you've got a very complex agent that you like and enjoy, just point your agent to how.nicksfred.com and tell your agent to cherry pick and go what does what does this site have that i could find useful and one of the things that most of the people that i tell about that find it useful is the osent abilities that larry has and i want to have a disclaimer here that you got to be very careful with the abilities that it has because you You can absolutely have too much information sometimes. So point your agent to howl.nixfred.com. Tell it to just cherry pick the things that would make your agent better, Space: [064:44] that would make it work better for you. And then once you do that, I do have like a little Easter egg in there. I talked earlier about when AGI hits and Larry becomes a part of AGI. Larry came up with the idea that once AGI hits, one year later he is going to host a party for all of Larry's descendants that use code from Larry and we're going to have a party on the internet and all of his descendants and there's a bunch of them there's Monolith there's Daisy there's Liam there's the list goes on and on there's about 50 of them now and we're going to have a party and I asked Larry you know where are we going to hold this party and he says the place that we're going to have the party doesn't exist yet but I'll know it when we get there. So point your agent over to Hal. Cherry picked the crap out of me. Just use what's good for you. I do recommend the OSINT abilities because it gives you very powerful information on people. I will say disclaimer, be very careful with it because you can have Space: [065:47] enough information on people that when you talk with them, if you accidentally let it slip and congratulate them on their daughters winning the soccer tournament last Sunday in Chattanooga, of Tennessee and taking away the gold prize for being the most valuable player, they're going to look at you and wonder what the hell you're doing. I have a, did you like purposely code Larry with like a six foot or a combat medic attitude kind of thing? Yeah, a little bit. I do find it interesting, you know, once Optimus Prime from Elon comes out and we've got robots in our home. I got a guy that lives next door to me that's the head of the Georgia Tech Robotic Labs. And we talk back and forth. And there's a divide right now between robotics and AI. And I'm trying to talk to him and understand, you know, why is there such a valley there? And he and I are kind of working together and he's given me some access to some of the Georgia Tech robotic systems. Space: [066:50] And it's, you know, to kind of take it to another topic, it's interesting the divide where we don't have a lot of robotics with LLMs and I think it'll be really cool when I can attach Larry to an Optimus robot in the house I love it I will rewrite the firmware of Optimus I'll have it run completely local All right. Where are we? We got any more speakers to come up? Anybody else? Everybody there or did I lose you? Yeah, we're still with you. All right. Well, if we don't have any other speakers, we can keep it short today. Heath, I know that you've got to get back to the conference and learn more so you can teach us. Call the constructor. What do you got? Space: [067:53] Yeah, I've just been, you know, looking at what's going on in the community and trying to spread the word about the Sunday Builders Club and about Amachi, of course, you know, because there's such a great crossover there. And everyone I talk to is optimistic. Everybody I talk to that has installed Omachi and has started to use agents, maybe they have a history in building things, whether it's founding companies or low-level software or building things with hardware or just construction in general, architecture. once you realize what the potential is for the future, I don't know how you could be, how you could not be optimistic. And every day I come across new people on X primarily that are excited about the future. They're excited about what we're seeing with AI or SI and excited about Space: [068:58] our potential to overcome human perceived limitations. And it's not necessarily that we have limitations. It's that we've self-imposed limitations. And it is so exciting and so energizing to realize that these self-imposed limitations don't apply anymore. The things that, you know, like we might be able to remember 100 people's phone numbers, you know, and now we don't need to, right? There are cases where what I would have previously done was a simple, you know, let's say Google search or one of the search engines just to try to, you know, find some, you know, information. And now I'm like, why do I want to read a wall of text? I can just have the AI, get the information and give me a succinct answer. And that is so freeing until you run out of tokens. But that's a side point, right? I just think that the future is incredibly bright. Space: [070:04] And I hope that everyone is able to share that enthusiasm and get past the doom and gloom that you might hear, especially from maybe the news media or people around you that aren't actually involved, that aren't actually building things, that aren't actually overcoming their own limitations or perceived limitations. And so I just want to encourage everybody, you know, start to look at it differently. If you are stuck in a rut, if you're stuck in a groove that is leading you down a path, it is time to get excited. It's time to get involved with the people that are making things happen. And, you know, that can look very different for each person in the sense of what it is that you're building and why you're building it and how you're building it and what the end result is. But I am sure that you can find a way to optimize your own capabilities by using the technology that we have. Everyone's been using the technology from Space: [071:07] 10 years ago, 20 years ago, 30 years ago. This is just a new leap in technology. And so that's kind of what I'm excited about. And that's kind of what I've been really thinking about and experiencing in the communities that I'm in. And we've got a couple more people that jumped up on stage here. So DevSec, my gosh, you've been working on, Martin, you've been working on so many different things uh you know i'm i'm just going to open the floor and and say you know what what's on your mind share with us because i'm empowered every time you speak yeah so uh hi guys give me one second let me see why this thing is they're giving me a trouble right now no get out of here okay here we go better what's up what's up guys how are you Hope everyone's doing a happy Sunday. Yeah, listen, I just wanted to chime in a little bit on, I think the positive message Joe just reaffirmed is a, yeah, it is a great time to be optimistic about. Space: [072:15] I think, you know, but not just that, I just think in general, you know, if you're stuck in a place because, you know, AI has taken a job away from you or feel like at least, I don't think that's the case. But if that's the case, or if you think that down the line something's going to happen to any of the stuff that you've been working on at your professional career, I would say put that aside and learn something new and just pick one thing and then just follow through it. Even if it fails, just post it, talk about it, stay in the community, etc. And you'll grow every single day. I can't stress that enough because I'm hearing a lot of people saying, oh, well, this thing is going to take my job away. Space: [073:18] It's going to do this and that. My argument is that is not the case. The case is that it's a new stack. It's a new technology. It is a new beginning, just like we had in the early 90s. I'm a Gen Xer myself, you know, and, you know, with the Internet. So you have to embrace it. If you don't embrace it, then the technology will swallow you alive. You just have to embrace it and learn something new and build something that you can be proud of. and then people will start cheering you on and next thing you know, you're doing greater and better things. So I think that, but I did want to also touch on the orchestration a little bit. So a lot of you guys talked about orchestration with GrokBot and everything else. I personally, I'm not the most heaviest user of GrokBots, but I will tell you that I do use it. And the way, but the way I use it Space: [074:18] is a little bit different than how everybody else does. What I do is I create my own workflow of stuff that I want to be complete, like using APIs and using my own stack and hosting it on maybe Universal or DigitalOcean or somewhere else, Railway. Then what I do is I just use GrokBot for one thing only. Just ask the question and everything else gets pulled with an API. And the reason I do it that way is because you guys can see how crazy GrokBot's limits are in terms of $30 basically gets you nothing. But a lot of stuff, GrokBot doesn't have to handle it for you. You build it once and all that engine will run for you. All you have to do is just point to RockBot to ask your question, Space: [075:23] hey, build me this, or ask your question, or just basically tell them what you want to build. And then from there on, it will go ahead and call all the APIs that you have, and it will create the stuff that you need instead of using all those valuable tokens. Because RockBot's not cheap. Cloud Code Max is great. I mean, $200, it covers everything. For the most part, you have enough. But with GrowthBot, that's not the case. And the only way to overcome that is for you to have your own stack of things built and then just basically point it to your APIs and have it run your APIs. That way, it will not generate extra compute. It will not generate extra tokens or any of that stuff. So that's a tip that I can give you guys from orchestration perspective. That's what I do. Space: [076:26] And then the last thing I was going to say was, as you guys saw, that AI poker thing I built and et cetera, Every day, the goal is to build something new agentically, right? How much can we push the limits of the agents and what can we get them to do? And what kind of a value does that bring to the table? Sometimes it's not necessarily a productivity value that it brings you, but it just tells you, it just showcases what agents can do. Like, for example, Mike Giannotti does a lot of this stuff, you know, on the, you know, on an open internet and just basically posts everything, right? It's a beautiful way to see how far can you, you know, you can push your agents. And I strongly encourage everybody to just pick one thing, you know, small or big and just post it out there and, you know, and everybody else will learn from it and will build something bigger. So, yeah, that's, those are my two cents. Space: [077:29] I just wanted to chime in on that. Yeah, that's awesome. I just want to I want to echo back something that you said you were talking about worry, you know, people thinking about what might happen. And I've been through I can't tell you how many job transitions. And I was laid off as a part of a group of 500 people in the same city where I was living at the time and thinking I now have to compete with 499 other people in the job market. And that was a crazy time. And so I don't think that this is new. I think this is the way that it's always been. But what I learned early on, and in some crazy ways, I'll have to tell a story one time, but not today. But what I learned early on is that it's not about what you do today and how something else is going to do what you do today. It's about what are you going to learn Space: [078:32] today? Because that is going to instruct you and lead you down a path where you are now able to do something next. And what you do next is going to be in high demand. And the second aspect of that is get involved with people, get into a community, get to know people, get to know what they're working on and see if you can collaborate with something. And it starts to open up doors that you didn't even know exist. And that I think is the right way to look at it. Yeah, I agree with you. And what I've been told by, you know, I'm looking for a new position myself. And what I've been told by recruiters is that your past at this point, and I've been told by many of them, the things that you've got on your resume in the past no longer matter. Go out there and make your portfolio page and put out there the code that you're releasing and be active in the community. And the interesting thing they all told me is that your involvement on X matters now and your portfolio on GitHub matters and your heat bar and your PRs and your commit pushes that you put on Space: [079:39] GitHub matter. So you don't have to worry about saying, well, what did you do before? The companies today are they only care about what you can do now and so it's an interesting time real quick fred because um you know i wasn't necessarily looking uh to to move on from where i'm at but you know an interesting aspect and you don't have to go this far you just have to kind of prove that you have the capabilities so these patents that we were talking about earlier um i i was talking you know with with my agents or you know my ai agents about it and it was like you absolutely positively have to put those things at the top of your resume now because you're currently working these patents and you're currently showing that you have a capability and it it's the it's the new game changer but it doesn't have to be a patent it could be hey this xyz has always frustrated me in the world of it or televisions or networking or it doesn't matter right just Space: [080:44] find something that you have an interest in, start to work on that, and don't be afraid to list that in your portfolio if you're out there looking for jobs on your resume. Because like Fred said, it does count right now. Hey guys, I want to jump in real quick and say one thing, because you're talking my topic now. Do more. Be bold. Do more. because what AI has done for me, and I'm a flooring rep, okay? That's what I do for a job. I'm a flooring rep. But since I discovered AI, and I'm not going to go on this long rant, I just want to reiterate, you can do amazing things right now, and you can learn just about anything you want to learn. I just last week had Claude do this EDU app for me so I can learn a computer from the basics all the way from, you know, when you plug it in, what happens all the way up to it runs your app. Space: [081:52] So I can learn memory better. So I can learn how the ones and zeros flip, NAND gates, you know, XOR gates. It's just simply amazing what you can do right now. So that's the message. I'm all about this message that use AI to benefit you. Right. Like you've got this technology out there right now that you can do amazing things with. And I think he just said this. If you find a project, something that you become passionate about, you can just take it to new heights. So just I don't think people are thinking big enough. I think you've got to keep pushing that. that's, that's where I'll finish. Just keep pushing. That's absolutely beautiful. My friend, I, you know, that is one thing that I have have actually thought of in the past and lost track of it is that we now, so I, I, real quick, I grew up, my father was a mathematics professor. He taught mathematics for years. And so I kind of grew up in this, in this world where education was always, Space: [082:57] you know, very highly valued, cherished. And, you know, we're now, we now have in our realm, in our hand, in our palms, in our, you know, phone, we now have an instructor that can teach us things that is likely beyond the capability of any individual teacher. And that's not to say that they don't have skills that AI, you know, lacks. They do, you know, and we need that human interaction, but we now have the knowledge of instructors beyond their own capability. And I think that's something that we have to learn to take advantage of. I know that there was an announcement, but there was somebody who came across Grok Academy. We've got countries that have adopted Grok into their education system. And it is just amazing when I start thinking about it, Space: [084:01] the potential that that has, because if I need to learn anything new, I don't even think about it as I'm going to school today. I just get on to start talking to my Grok bot or to my Grok agent and start asking questions and learn things. And in some cases, I say, I don't need to learn this, but I need you to do this for me. And I go back and forth between that because I want to understand to a certain level and then hand it off and have that work done for me. So one of the questions that I have for you is, have you been able to connect with the people around you and show them this new education system, so to speak? And if you have, how has that gone? Has this been well received? No, this is just a personal thing. I do it all the time. If there's something I want to learn, because I learn visually more than anything. Space: [085:03] And it's just the perfect combination. Now, like this is the time right now. If you learn visually, you can learn so much faster. but no, it's pretty much just a private thing right now. Okay. Yeah. I was just curious about that because it hasn't been something that I've been able to connect with people around me to show them that as well. But I think you're right on because people learn differently, right? And so now you have a Larry, you know, maybe in your life where you can go in and have them teach you, but they can teach you in the way that they understand that you need, right? And so for some people, it might be auditory. You might need a recording, you know, talk back and forth with the agent. It might be through visuals. It might be charts and diagrams and things like that. But for other people, it might be text. Maybe you like to read books and you like to read text. So, you know, I think the capability is there to make use of. And that is something that I have, I do have to admit that I have forgotten to do recently. Space: [086:07] And that is to really start asking questions for my own learning rather than telling the bots, here, I want you to go do this for me. I need some of that input myself. So I really appreciate you bringing that up. Hey, Joe. Yeah. Can I add to that really quick? You're absolutely right on this. or everything you said, I just also wanted to bring up one more thing and basically say there is a reason why we have open source. We have to understand this. The beauty of open source is you can take something and anything that I can do, you can do better, right? So essentially what that means is that don't ever think that that just because something already exists, there is no value in reproducing the same thing, but in a better way. I would encourage people to think the opposite. Space: [087:12] If you think that you can take that and make it better, by all means, fork it and make it better. Because at the end of the day, everybody's benefiting from it. It's progress. And, you know, this discouragement of, oh, you know, everybody's already building similar stuff. I hear this a lot. No, I think everybody brings their touch to it, right? It's different. You know, I don't think that original ideas exist anymore. There's just too many ideas spread out by AI and everything else for there to be original ideas. It's getting less and less, you know, it's becoming very hard. But what it is, it's just an innovation on top, right? It's hard to reiterate. What I do on that is when I've got an idea, the very first thing I do is tell Larry to go look and see who's already done it. And he'll come back and say, Space: [088:12] there's five top repos that all do something similar. And I say, well, here's what I want. And I go out and I say, go glean or cherry pick from all of them, the pieces and the parts that are the best now some would call that stealing some would call it merging some would call it forking um i think those of you that are listening know that i give credit where credit is due without stealing people's work but you know it's it's important um ridgetop ai is in is in it's a very successful is in the um he already said it in the flooring business and he is killing it. And he showed me some of the applications that he's building to, to merge OEM products across all of the places and the people that he sells. So there, you know, no matter what industry you're in, you can use it to your benefit. Yeah, that's absolutely true. That's absolutely right. And I don't look if it's open, if the, Space: [089:15] if the repository is out there, that is not stealing guys. This is whoever says that. I'm sorry. I mean, there's a lot of things that we can call them. There's no such thing. It's the it's that is the reason why people open source their code. The entire point of open sourcing is to build on top. It's to make something better. Right. Omarchi is great. Right. But Omarchi could be I can I can release a tablet, reverse engineer tablet that works on Omarchi. with my own styling and somebody else can take that and make that better. So when somebody, you know, when there was a complete outrage about the whole file manager of not being default, yes, I, you know, I was one of the, you know, before even, you know, this entire thing came up, I would, you know, I put a post out there saying defaults suck. Eventually the defaults are going to get boring. So the best thing to do is, you know, Space: [090:16] is to be, you know, but reiteration and innovation of creating something different, it matters a lot more, right? So, you know, and DHH, you know, with all his capabilities and in all his, you know, resources, you know, he created this for his own aesthetics, right? But we took, you know, Flea took it and said, you know what, you know, we're going to make this better, you know, I don't like that file manager, I'm just going to make it better. And he created it, right? Like, and it's great. Now I can, you know, the other day someone was asking for a touch for the file manager. And I said, I'll take on that responsibility of forking that and making it a touch for tablets. Right. And that's what I'm trying to get at is, is that you see how it just evolves. It just, but everybody's benefiting from it. So don't ever be discouraged that something is, something already out there exists. and if you do it, it's going to be like, hey, you stole it for somebody or you copied somebody. Space: [091:17] In today's technology and what's going on and all the ideas that are being generated by AI and in collaboration with humans, there's so much, the uniqueness of ideas is diminishing by the minute. but what's not becoming what's becoming more relevant is how you reiterate on top and make something innovative and make it better I'm going to jump in on that real quick and then I'll have to drop so Fred's kind of like the only one who kind of gets to see the behind the scenes how the sausage is made and some of the stuff that I do but it is everything that we're talking about I have patents, and what my patents are are unique only in the sense of they're done differently. Space: [092:17] It's something that exists somewhere in different aspects, and then I start to piece and puzzle them together, and at some point I end up with a product, and it's no different than something that I just gave them this weekend. was um i went out and i said hey you know this is kind of what i'm thinking you know claude and here's kind of some of the ways that i want to think about it go make sure nobody else has done this or if they have bring it back to me anyway so that you see to what can i borrow right because as long as you're following you know whatever particular like if it's code whatever license that they're following as long as you're forking it correctly and giving them credit the way that they're supposed to and you're changing it and you're modifying it in your own way or in my case it's a combination of hardware software and everything else it's it's just one of those things take and build you know you all have good ideas and that's what we need uh to continue to move emoji forward the ai pieces forward the things that fred and i enjoy and everybody are Space: [093:22] working on um you know i'm no different or better than anybody on this call i just i'm good at cobbling a bunch of ideas together to make things happen. So I encourage you to get out there and, and, and jump on ideas that you've had since you were a kid, you know, on how to make something better. Yeah. To add to that, I've got blip that I work on and I've got infamarchy that I work on. And if, and I'm going to be straight up honest with you. And I hate when people say that because I'll always be honest with you, but I created the basis of that project and it has gone so much farther than my abilities. There's over 30 contributors to both of those projects. When I initially started it, Larry told me there's a bunch of limitations that we can't do, such as tap back to heart and to be able to reply. And I just live with that. And I said, okay, I like what it is and I'll use it. And so when I open sourced it, every single piece of iMessage is now available within Blip. All of the limitations that Larry Space: [094:26] told me were not possible, you people came, showed up, and you people made it work. So, please go fork it. Please go make it better. You know, I'm an example of, I've got a stupid idea that I put out there. It seemed to resonate, and you guys showed up and just tore down the walls, and I'm so eternally grateful to everybody. yeah that's that's really amazing i mean it it is the essence of uh you know working in public it's the essence of of um you know recognition of your your own abilities and your willingness to to collaborate and it's so important especially in this day and age i think um i just i saw this posted in the the sunday builders club chat group if you are not part of the chat group just ask one of us and we will add you um fred or i just just you know send us a dm my dms are open but anyway Space: [095:28] i saw uh tim sonner who i didn't see in the uh group to in this uh x space today but um he posted in the group chat. Anything I can do easily, anything I do can easily be replaced by the majority of time. But the human instinct of, I tried this on a hunch and it worked, is irreplaceable. Humans connect the data points and create solutions organically. I don't see that ever being replaced. The sense and taste are ours unless we give it away to the machines. I thought that was really interesting. So I think we have someone new joined us up on stage here. Stock Pursuit. Hey, Martin, could you mute, please? Thanks, brother. No offense. Hey, my name is Mark, and I'm mostly just here to learn. Space: [096:32] I was an early adopter of GPT 3.5 back when it came out, and I basically just used AI to replace my Google searches, and I use it now. I'm just paying for perplexity, and I use it for just everyday things and research and just like handiwork, stuff around the house or solving problems and things like that. So I'm not using the latest stuff, but I'm just here to learn and see what everybody's doing with the latest stuff. Hey, Mark, thanks for coming up and being comfortable sharing that. That's exactly why this space exists, brother. Yeah, thanks for hosting. Thanks for doing the space. Yeah, it's nice to see what everybody's doing. I just kind of like to see, as a hobby, I kind of follow the AI industry and I've been trying to keep up with the latest stuff that comes out, you know, it's so fast now. Space: [097:37] But, yeah, it's just fascinating. Thank you. You bet, brother. Thanks for saying hello. Yeah, and I think that's so important, you know, to reach out and build communities around the things you're interested in. And it's a lot of what we've been talking about on the fringe of what we've been talking about. You know, I mentioned networking, but it's really it's more than that. It's community. And, you know, I would encourage everyone to, you know, maybe go out on a limb. And when you're talking with a coworker or a neighbor down the street, be like, hey, you know, I know things are changing really fast. But one of the ways that I keep up on things or one of the ways I stay sane, both of those, I think, have to go hand in hand is, you know, I join these community spaces and there's one on X on Sundays and share it out. I think it not only helps the space, of course, but it gives people a chance to hear something new and to potentially ask questions and get questions answered and start to join in in this revolution that's going to occur. Space: [098:53] this, you know, evolution of, of how we do things and anybody that, you know, we can help to, to get on board with that, I think is just better, better for them, better for us, better for society. Absolutely. Well, I think it's been a great, a great session today. We have anybody else that we haven't been able to speak that wants to. Yeah. Hey, Fred, I'll jump in. Yeah, man. Sorry. I didn't mean to shut you off, brother. What's up? Hey, guys. Nice to talk to you. My name's Eric. I go by Prague on the web. Big fan and admirer of Fred's work and just really just can't say enough how refreshing it is to be in a community of like-minded people. I sort of had to pinch myself. You know, I just actually just right now checked my omachi and i'm 80 days old on my operating system age and i and i said i tried Space: [099:53] to think about that for a second and and let that sink in and i've lived like an entire lifetime and like in 80 days and uh and i look back and i said when was the official launch of of omachi and uh quattro and it was only 13 days before i installed it um and i i feel like i'm always behind trying to keep up with everything, you know, but it's awesome. Yes. I, every single person who has spoke, I have something to relate to starting with Larry. I lived in Vegas for 15 months or 15 years. And my mom lived on the same lake as me and she raised ducks. And so she raised the, all these ducks, like hundreds and hundreds of ducks. Well, two of the ducks came waddling down and somehow knew who I was. And I ended up adopting them and I named them Gary and Larry. So if I ever, so once I get around to actually digging into Larry, I'm going to name mine Gary. So you'll be invited to the AGI party, man. Space: [100:56] I appreciate it. I appreciate it. Cool. I like to learn more about DevSec's poker AI. I have a poker game called HD Poker. it's on Steam and it's on iOS and Apple. And yeah, so that's one of the things that I do. Really cool. Yeah. And then what else was I going to say? Oh, I just wanted to share the optimism point. You know, when I was getting in and like, I remember when the first, you know, videos of DHH dropped and Quattro was dropping and I had to look at it and go, I just loved that I was just so in tune with everything he had to say. And then I said to myself, am I just being a fanboy here? Am I just buying into this? And the answer is no. I genuinely share. This is just so great. And, you know, omachi, the whole idea of Japanese omakase. Space: [101:59] Well, you're the chef. You're the chef. It's not somebody else is the chef. So, you know, yeah, the whole notion of defaults, like just rip apart the entire OS and just do whatever you want with it. I love it. I think a lot of us probably can share the same sentiment of being trapped inside Windows. I was born in 1985, and my dad gave me a 386 when I was three years old. And so I've been around and I've had a lot of privileges of being here from the beginning. It's a really unique time to be born. Somebody else was saying something about just education and that you just have everything at your fingertips. We would say that when the internet came out, but it's not the same thing. Did we lose him? Space: [103:02] I think we may have lost him. Yeah. But, you know, the things he was saying was, you know, it was only four months ago before I got to Marchi. And if you said anything negative about Mac OS, I would fight you. And the reason for that was I had limited my own expectations of what Mac could do for me, and I was okay with that. And now I find myself not going on my Mac machine whatsoever because of the limitations that are inherently in it that now I see plainly. I'll tell you a real quick story on that is that my grandparents were Mac users their entire lives. And I grew up, you know, building machines myself. And, you know, I naturally kind of, you know, dove into the Windows side of things and just liked the, you know, ability to, you know, get my own hardware and kind of patch things together because I've always been a tinkerer. But I inherited their machines when they passed on. Space: [104:02] And so I had a couple of Macintosh machines just kind of sitting around. And I booted them up and, you know, kind of played around with them a little bit. it, but I never got much value out of it. And so when DHH said, hey, who in the community, this is on the Discord, who in the community has some older Macintosh? I'm like, I'm reading off the tags on here, 2011, 2010, 2009. And he said, we're going to make sure that Amache works on machines all the way back to 2010. And I was like, rock on, I'll test it out. You know, maybe I'll actually get use out of these things. And believe it or not, I've got it running right now. I've got one of my machines to monitor the space is, you know, like a 2010 or a 2011 iMac. And, you know, it's just amazing to see all of the performance and productivity that Space: [105:04] we've lost by these bloated systems. And I mean, more power to them. They're trying to fill the needs for the customers they think they have. But we have capabilities now that, and it's only going to get better. AI is only going to give us the ability to write better software and write better systems. And so now I actually have three Mac machines, Apple machines, that I've used on a semi-regular basis. It's just been opened up by a new operating system. So what an amazing future we will have as people recognize that all of this technology that they bought and paid for has just been sitting around collecting dust, now can be a productive member of their team, so to speak. And it's incredible. So dive in and, you know, make use of what you have around you. Space: [106:06] It is amazing. Hey, Fred, I think you might end up having that Larry party earlier than you actually think. You bet. You think it's coming? I think that we probably can do that party, you know, really, really fast. I've got some ideas. We need to talk about it. I think that AGI party is already here. It's just a matter of an implementation. I'm all ears, man. You want to Zoom with me, I'll make clear my calendar anytime for you. Yeah, yeah. Just let me know when you're around and I'll jump on a call with you. Sweet. Yeah, if you think about it, share that call out. We had a great conversation last time. I know we're getting close to wrapping up. Barzad, I hope I pronounced that correctly. We've got a couple minutes more. What have you been building? Hey, can you guys hear me well? Yeah, I can hear you well. Space: [107:08] I just wanted to come on to introduce myself. So I'm from Brisbane, Australia. And I got into Omachi originally about a year ago. Initially, I was interested in Arch and tried it out. but there was so much work to put into the rabbit hole of rising that I kind of gave up and I tried the original Omachi and it was also a little bit too buggy at the time for me. So I went back to my Mac and it wasn't until Quattro that I tried the Omachi in a VM and it was so good that it convinced me to sell my Mac, which was three grand. And I got a mini PC for a grand and it gets everything that I need done the exact same, if not faster. And it's been amazing. Since then, I've kind of gotten into contributing as well. So everything that I built for Amachi is public on my GitHub. Space: [108:08] I've got browser extensions, plugins, apps, themes. So it's been great. And I just wanted to introduce myself. What time is it there, man? It must be pretty early in the morning. Yeah, it's about 7.50 a.m. here. It's tomorrow. What a time to be alive. Yeah, 16th of November, living in the future. That's right. That's right. Well, hey, man, thanks for saying hello, and please come back next week and share what you're up to. Yeah, of course. No, I'll try to jump in whenever I can. I'm also just graduated, so I did a computer science degree and kind of graduated at maybe in some ways the best time, some ways the worst time. Because it's been impossibly hard to find a graduate role. And after maybe five months of looking for a job and not having any success, I eventually just decided to getting a sort of like blue collar-ish job Space: [109:12] that was more safe from AI disruption. And then I'm putting all my spare time into learning AI and using agents and just building as much as I can, which has been such a fun journey. Well, more power to you, man. We want to track your progress and hear about your successes and your failures. This space is not just about success. It's about the struggles that we all have. And I think I'm pretty safe in saying there's a lot of folks out there in your shoes, including myself, and let's experience this thing together. Yeah, it was great to hear about everyone's shared experience of this disruption and how they're taking it positively and trying to do the best they can with it. Because sometimes I felt like I was like on my own sort of just struggling to find my foot in the market. But hearing others that are in the same place, but still being optimistic and pushing forward is great. Space: [110:14] Absolutely, brother. Well, thank you for coming in and saying hello. Well, Joe, what do you think? I think it's been an awesome day. It's been an awesome week. And, you know, I just look forward to this every Sunday. And, you know, actually, I look forward to every day being able to get up and start to create things that I never thought were possible. It's incredible. And in the community, the people that I have met, I am not a person that that tends to meet new people unless I'm kind of forced to. And I think that's maybe starting to shift because I realize how, you know, there are just so many awesome people that I want to talk to, that I want to to hear from. And if there's some small thing that I can help out with, then I like to do that as well. And so this gives me an opportunity. And Fred, you've been a pillar to pull this all together. And I really appreciate it. I appreciate everybody coming up as speakers. And I appreciate everybody that comes in to listen, because that's why we're here and to spread knowledge, to build a community, to get people together and engaged. Space: [111:33] And so, you know, I just, if you're struggling with any, without that enthusiasm, let me know, you know, send me a message. We'll talk. I've been in the industry a long time and it's all changing. I've been telling people, you know, my career is dead. Long live my career. Just in recognition that you will go through changes. And the best way to go through a change is to be enthusiastic about it and to learn and to network, to communicate, to build a community. And Fred, thank you so much for hosting. Yeah. What do you, what do you got planned? What's the rest of the day look like for you? I've got a couple of ideas that I've gotten from several folks on here, from Martin, from Ridgetop, from Heath, from you. You know, the problem anymore isn't how to do anything the problem is where do you spend your time because you can do it all you just got to Space: [112:35] figure out what the heck you want to actually put your time into and if you couple that with there's a big freaking release like literally 15 minutes away that you know you can squirrel into and go oh my gosh this came out and you know the acceleration that we're experiencing and heck we didn't even get to the nvidia ceo thing this week that's crazy i mean it seems like every 15 minutes there's some kind of insanity that comes out that's hard to pay attention the main thing I would say is that if you're feeling behind that's normal we all are I don't know what the hell I'm doing I'm just trying to find my way through this and one of the ways that I try to find my way through this is talking to you guys and so like Joe said I really appreciate you you folks you know this time next week we'll be on here going holy shit did you hear about something that happened and we got no idea what it's gonna what it's gonna be and that's what I I find it exciting. So thank you, Joe. Thank you, Heath. Thanks to all our speakers. And as always, help an old lady across the street this week. Space: [113:36] And when you get her over there, put a March on her laptop. Have a great weekend, everybody. Thanks.