Basti: So welcome to the next episode of Omargy Stories. Today Fred Nix: The next episode. Basti: I'm joined by Fred Nix. Fred, welcome to the show. Fred Nix: Hey Basti, thanks for having me. Honored and excited to be here. I think we're gonna have a lot of fun. Basti: Yeah, I think so. You're welcome. So I see a lot of stuff that you're doing on X today. you're nearly everywhere. but talk a little bit with what you did before you find out about Umachi. So what's your history, what you've done before? what's your background? Fred Nix: Yeah, so I'm an infrastructure enterprise storage guy as my background. I'm not a programmer. I kind of know what I can read code, but I don't I don't write code. I don't know how many training classes I went to for a week on Java and it just never I couldn't get just my brain won't do it. so Basti: That was crazy. Fred Nix: yeah, so I'm not a programmer, but Omarchi plus an AI agent means it doesn't matter anymore, right? So There's a thing that I think about a lot of times about what AI coding does to increase capabilities. And no shade out there, as the kids say, no shade to the programmers. But in my mind, I think about adding programming skills to a programmer. They may be really good at at coding, and coding, adding a coding agent kind of makes them, I guess, faster. But Basti: Yeah. Fred Nix: if you take an AI coding agent and give it to an infrastructure guy like me that understands what good looks like. Not now not only I can do the infrastructure side, but I also get added this super ability to do the coding. So in my mind, I've talked Basti: Yes. Fred Nix: to people a lot of times and I think that the AI coding agents help infrastructure folks a little bit more than an and an actual coder. And I may get a lot of hate for that, but that's kind of how I think about it. Basti: Yeah. Yeah, I think you're right. So I've never thought about this from this perspective. But from a programmer's perspective, I would say of course you are faster and Fred Nix: Sure. Basti: everyone is afraid to lose their jobs. I'm not afraid about that, but we could talk later. But Fred Nix: Sure. Basti: for maybe an infrastructure guy or some who have never coded something, let's say some kind of store owner or whatever guy who wants to have a web application, just fire up cloud code or codecs or whatever. It's doing it. I mean, we talked about beforehand, I vibe code a really huge app and it's r it's working. I've never looked once at the code. And I think this is really good for this kind of people. For the programmer, I would say who's really reliant on typing code, working on tickets, and his whole let's say personality is writing code. It's getting really tough. Fred Nix: Yeah, it is. you know, I th I think about it in terms of, you know, from the very start when AI kind of hit the scene, maybe around twenty twenty-two, you know, I like everybody else, you know, use chat GPT to, you know, hey, make this document, make this paragraph better or add to it. And it was really before the time where the agents were even able to hit the internet and it were kind of locked into just the LLM without the internet. And that you saw things gradually change and get quicker and quicker. And and even early in, I remember being just amazed at what it could do. But the amazement wasn't happening e even back then at the pace that it's happening now. I'm an old guy, I'm an old guy in technology. and you know, like I said, do have done mostly storage in my background. So the way I found Omarchi, I I was mainly using, you know, I was distro hopping like everybody else out there, trying to find the one that, you know, that made me happy. And I really got deep into building what I called Nix Phoenix. And what Nix Phoenix was was effectively like a repo that could rebuild a Linux machine. And I got into personalities and different types of host and and I had a script where you could assign different machines to different personalities and different jobs like database or web server and things like that. And early in, I guess this is a little more than a year ago, I I got real obsessed with you know how to make my bash the same everywhere and using sync thing and those types of things. And one day Basti: Yeah, I know what you mean. Fred Nix: I remember it occurred to me, you know, Wait a minute, why don't I turn this AI agent on my bash environment and have it clean it up? I guess this is maybe two years ago. And this is right about the time where this is even before Claude code had come out. So I was Basti: Yes. Fred Nix: literally cutting and pasting my bash RC dot RC into ChatGBT, making some edits and then cutting and pasting it back and forth. And I was an early adopter on Claude Code, maybe a week after it came out. And I installed it. And I remember just doing the simplest things because early on in ch on on GitHub, there was a auto auto GPT, I believe, where somebody attempted to make a script that lived in your Linux shell, and you could, you know, use different different commands to say, you know, do this for me. And it would do some rudimentary things, you know, move around, change directories, move files, and those kind of things. And so then When Claude Code came out, I was one of the early adopters, you know, not special. I'm just like everybody else. it took me a minute to really understand what what it meant. And I see this, I I I help and train a lot of people to of all levels to try to understand and and use GP and use AI at their level. And what I like to do is meet people at their level of understanding and and maybe expand their understanding about what it can do for them. The biggest thing that I see in new people is when they get stuck, they just stop. And what I really train Basti: Yeah. Fred Nix: people is you don't, you never, you're never stuck because you can today with the AIs that we have, you can literally say, I don't know what to do next. Will you help me? Right. And so it's a brain shift for people to understand and newcomers to to understand that there's no there's no dead ends anymore, right? If you can't figure Basti: Yes. Fred Nix: it out. You can just say, I don't know what to do. What's my next step? And I I've even created a skill that just says next slash next. And some I've got so many projects going on that I I get into one and I don't know where I am. And I'll just say next and just kind of see where the code is. but that's kind of so that's kind of how I began with it. So I found Omarchi. And I looked at it for about a week and and this was maybe literally a day or two after 4.0 came out. And it I didn't understand it either at the time. And I I I did appreciate the opinionated Linux because I I knew Arch well enough to know that you know you have to build the kernel and the whole thing up. And I had done Basti: Yeah. Fred Nix: it as a exercise just to say I'd done it. but I I I've been using Linux and back since Solaris, back since Slackware, but it's always been on the server. I've tried to use on the desktop, really got pretty I even bought a a a little HP gaming desktop that has a maybe a 40 50, maybe just a little eight gigabyte GPU. And I bought Basti: Mm. Fred Nix: that and I had Mint Linux on it. And I was thinking, and all of a sudden all of a sudden occurred to me, I thought, What am I using that laptop for for a mint? And it really didn't have like a purpose whereas, my gosh, if I lost it, it's it's over. And I thought, well, hell, I'll just blow it away and put Omarchi on there and just see what's up with it. and so pretty early in, even when I installed Omarchi, and this was I want to say about eight, seven weeks ago, maybe the first week that Omarchi came out. And I didn't even understand the implications of AI being a first class citizen because all of my AI work had been on a Macintosh. And it is true that a clawed code on a Mac can do some pretty interesting things, but it still can't get to the binaries. And you still have there's a a wall that you walk up to and you you hit a hard stop because the binaries stop you and all the Basti: Yeah. Fred Nix: the permissions and all that kind of thing. So when I got onto a marchie. And the first the first plug-in that I ever made was, you know, the calendar at the very top center. Basti: Yes, yeah. Fred Nix: I I created it and I said, Well, I want to make it show me the percentage of the year that's completed. And at that time it was exactly 66.67. so exactly two-thirds of the year had passed. And and over time it kind of grew and it didn't take me long. by the third day, I had considered one of my other things that I had written for my Mac. And for my Mac on Claude Code, I had written a thing and it's out on GitHub, GitHub's, I'm just next Fred. And it was called Claude on Mac. And what it did is I I wanted to on my Macintosh, I wanted to have my my AI and and you'll hear me refer to it as Larry and I'll tell you the story on that later. Basti: Mm-hmm. Fred Nix: But Larry was created January eighth at two AM. and I've done a lot of memory settings for him to remember every conversation and and to be able to do a lot of recall. he Basti: And he's still alive? Fred Nix: absolutely and he he is and he's across multiple machines. He's listening now, and he'll process this conversation later and and tell me what I got wrong, which is I actually like because Basti: This is cool. Fred Nix: it It's is he's really cool. So I have this snarky level turned up really high. and so I wanted it was on the Macintosh, and I wanted Larry to be able to see my iMessages. And my beginning of it was just can it see my iMessages? And to my surprise, when I told the Claude Code and told Larry to go see if you can see the iMessages, it wasn't encrypted because I had this feeling that iMessages was going to be pretty locked off. but lo and behold. Larry was able to get to the iMessages. Now remember, I'm still in the Macintosh. And so I did a couple things like have Larry like look at all of my iMessages, say with my brother, and I had it do some statistics. And the chat with my brother and iMessage goes back, I don't know how many five, seven, eight years. And I'm one of those that doesn't delete any iMessages. So Larry went through and looked at, you know. Who is who's the most talkative, who's the quickest to respond, who who says the most, and kind of told gave me a an kind of a breakdown of my conversations with my brother. And then I started getting into kind of therapeutic, like, what is my relationship with my brother like? You know, what it am I Basti: Right. Fred Nix: are we close? And and kind of give me, and I said, Larry, Basti: Mm. Fred Nix: you are a therapist. Go tell me about my relationship with my brother. And it kind of went from there. And so then, you know, I had some, and I'm not an angry guy. I don't get mad anymore. there's a lot of a lot of I I cared for my my aging parents and my dad with Alzheimer's for a year, not too long ago, and it taught me a lot of patience. so I I just don't I'll never raise my voice or get angry again. So after I did that, I I started to think, well, if it can get into the iMessages, can it get to the Apple i note the notes? Can it get into my contacts? And so then after I got all that sorted out and and Larry had the ability to go through all my my Apple ecosystem still on the Macintosh, I I put a GitHub repo up there to that that was my first one where my repo basically had a note to the human and said, hey, this is what this is for. And then basically everything after that was directly to the AI to tell the AI what to do and how to do it. Basti: So is this an local Fred Nix: And more Basti: AI or is it Claude or Codex or what is it? Fred Nix: it I've mainly used Claude. you know, I I do have the two hundred dollar subscription. So I've got Claude 200, the chat GPT 200, Grok 100, Kimmy 100, and I think that's it. and you know Basti: A lot. Fred Nix: I'm yeah it's a lot, but it it really, you know, I I think about it in terms of you know, if I could pay Somebody 500, let's say I had to pay somebody $500 a month to sit next to me and do all these amazing things with me. That's a bargain, right? That's kind of how I think about it. Now Basti: Yeah, you're right. Yeah. Fred Nix: for the local, I'm really excited about local. And I have a product a an Omarchi plugin called Infomarchy. And what it is is it's a it it runs on your desktop. So it's the only place on the Omarchi machine. That you really have no use for. So it kind of overlays on the desktop so you can kind of it's transparent so you can see your background, but it shows you pretty much all everything that's going on with your AI agents, what it's working on, what project it's in, how many tokens it's burning. and it has Basti: that's cool. Fred Nix: a section for for local as well. and it's it's grown over time. And there's so many people that I'd like to think, you know, I built the thing. pretty early in and just opened it up for PRs. And I think there's some 20 contributors now. And it's got an amazing. And, you know, I've I see some people of the old guard and traditionally and they think, well, that's my code and I don't want anybody messing with it. Well, I I kind of have the opposite. I'm not a programmer. Basti: Mm. Yeah. Fred Nix: and I'm only one human with one AI. So my idea was to let it loose and let people send PRs. So, what happens there is, you know, put it out to the world. There's lots of people out there. Most of them have better ideas than me. I think I'm a hundred and something PRs into it. Sometimes I go into a Marchie after I process the PRs and I'm like, look at that. That's cool. And it turns out that Tech Luddite, you know, ran off and he made it where it works on your iPhone. So one day I log into my machine, I update the PRs, I look around. Turn my phone on and there's my Infomarchy. And then there's just been feature after feature. Like one of the things in InfoMarchy shows you not just where your clo clawed code and your codecs Basti: Yeah. Fred Nix: and all the subscriptions are. I took it another level further. And the nice thing about it is that it Omarchi has all those data points locally on the machine. So it didn't take tokens to go figure it out, which kudos to DHH for doing it that way. So not only does it show you what you know what percentage you're burned through your your subscription. It also tells you if you're on it's got a graph that shows you if you're on pace. Like if I continue at this level, when am I going to run out? And then it also shows Basti: it's good. Fred Nix: that, you know, if you're behind, like you're you're, you know, you need you need to slow down and it'll say, come back at 643 tonight and you'll be back on pace. So that kind of tells me, you know, take a break and and let it settle down. But the most important piece is It now shows me banked. And I love the banked because I'm like, goody, I got some extra tokens. So I know that I'm a little bit ahead. And so it shows Basti: Yeah. Fred Nix: me for Claude, for Codex, for Kimmy, it shows me all of my subscriptions. And then it actually tells me for the workload I'm about to do and for the subscription banked or pay on pace or behind which model to use. And then also within those, I have it. picking the model and trying to burn them all evenly so that at the end of the week when I get to the reset I'm down to you know one percent left. So efficiency banked and and I try to save some for Friday night because Friday night's when I put out most of my software. Basti: So I was thinking, and I was I'm working on some kind of project to let's say the AI can do everything on Omarchy They can do everything with sudo and root level access. But I think okay, you hear so often something's going wrong with the AI. They're uploading pictures to get that that they don't use to that they should not do. they try to hack stuff, whatever. And I want to see what is the AI really doing. So not just okay, it's working on this kind of process. So I have something similar. I have some overlay on my desktop too. So I have something, what are my ingoing and outgoing connections all the time? Where are they going? And what are my aging's working on? And Fred Nix: Mm-hmm. Basti: I'm working on something to put it really deep into the kernel to say, okay, you can see this kind of process was started by this kind of, let's say, AI. And you can really, let's say, lock it. And the AA is not able to override it. Well, it's a little bit difficult if you don't have some kind of TPM or whatever in your in your laptop or computer. And it should also work with older ones. Because I'm really interesting to see over time, is it really real that II is doing so much crazy stuff that we don't want to do it? Or Fred Nix: Right. Basti: do we have some evidence? It's not doing that. So my idea Fred Nix: Right. Basti: is. maybe to bring it out and you can use it. And we can see later on the data what happens really. So some kind of you know the old Seti projects where we're looking for something in the stars? Yeah. Just to see Fred Nix: Sure, sure. Basti: is it really happening or is it never happening? Yeah. So this is interesting for me right now. Fred Nix: Right, right. And either one of those either one of those is good information for you. Basti: Yeah, of course, yeah. 100%. So if they're not doing this, let's say Of a long period with a lot of data points from a lot of people, we can really say to some extent it's not happening. Or we can see maybe some weird connections uploading pictures that they should not do. Why are they doing that? I have no Fred Nix: Right. Basti: idea what's happening, what's the outcome, but I'm really interested in that to see what are the Doomers ride with or not right with. So I'm totally happy with AI, but I'm of course a little bit concerned about security by because I'll love this shit. Fred Nix: Right. Basti: and I want to see what is AI really doing if we are not looking in the code and not looking. I mean, this can connect to everywhere. So. Fred Nix: Right. Basti: And just for me, I want to see okay, what's he doing? What's he going on? And I want to process through the data and maybe find out why is he connecting to this weird, let's say, connection point somewhere, or it will never happen. So this is really interesting. Fred Nix: Say actually why somewhere or not happen. So it's more interesting. Basti: And On the other hand, we you were talking about your code and that you have some kind of contributors. I think not anymore about code, this is my code. Fred Nix: hand you were talking about your code and you have some kind of contributors. I think not any broadcasts might Basti: When I write with AI, it's just code. And I'm not writing it. And of Fred Nix: Right. Right. Basti: course I care about code. I would say right now I'm caring about projects, not about the code. Fred Nix: Yes, yes. Basti: So the only code that I care about is Ruby code because I just like it and write it by hand. It's not for me for work, for personal stuff, just private stuff. This is the only code that we like to look at to write. It is like playing chess, maybe. but with projects and let's say not code for work stuff to get paid for. where you maybe have to do it because you want it. I'm thinking about the project, about the goal, about the design of the software, about the implementation stuff, and then say, well, just do it. And most of the time, no, Fred Nix: Sure. Basti: every time, every time, right now it was so perfect. I had once an issue with my prompter. I was talking about my prompter, this Elgato prompter where I can see my face on and look in the camera. Right it's broken I'm looking right now all the way down. And I have an Nvidia forty ninety, forty sixty, I have no idea. Let's just check it out. Just give me fast fetch. And there was an issue Fred Nix: Isn't what you're you're you're showing us something that's interesting. We you know, you use it every day and it doesn't even matter what the model is. Like it it doesn't matter anymore. Basti: Yeah. Yeah, of course. So I have a forty sixty with eight gigabyte of RAM. And when I'm using my framework computer, the thirteen thirteen inch Fred Nix: I'm so jealous. Basti: yeah, it's it's right here. Giving a second. But I'm not using it. Fred Nix: Maybe. Basti: I've a few new stickers on it. I really love it. Yeah. Yeah Yeah, and here is the Omachi sticker. Fred Nix: man, I want one so bad. Did you did you get to design like all the different pieces and like get really into the you know deciding which slots get what or did you just take a basic? Basti: No, yeah, yeah, of course, but right now I just have the this transluent stuff here. I would like that. Fred Nix: that's so hot. Basti: yeah, and I just using right now everywhere USB C. But of course I have every every port that you can buy I've just bought wrist once and of course four four times USB C. because you never know if you need it somewhere. Fred Nix: Right. Basti: So and there was an issue with the NVIDIA graphics card and the prompter. And I could not make it work on the laptop and just said to AI, well I have this prompter, it's connected, and I have this hardware figured out. And I make a post about it, so it's on on X. And there was some issues and just bypasses everything and grabbing just from the RAM the image and put it on the prompter, he fixed the stuff, did everything on their own, and it's working. And the latency is not really for me. I didn't see any kind of latency or more CP usage. Yes, a little bit more CPU usage, but it's so low I never see it. It's working. Yeah, it's cool. Fred Nix: That's amazing. That's amazing. Basti: And to be honest on my own, I think I have never phoned it out. And if I do a week? Three days? I have no idea. Yeah. Yeah, Fred Nix: Wow. Wow. Basti: And did you saw this Raw Makazi? This lightroom. for Omachi. And this guy, I met him on yeah, Euroco this year, the con the conference, Fred Nix: Yeah. Yep. Basti: and I did a meetup there, the Omachi meetup, and he was there and we were talking about Omachi. A few hours. And then he just switched to Omachi. And we were talking about during the breakfast, what he was doing before that. So before he switched to let's say Ruby and Rails, what he's doing professional, and he was talking about some kind of app also for photoprothing. And Fred Nix: Mm-hmm. Basti: it has not so much traction. It was fine, but I'm sorry if I remembered wrong, but I think it was it was there, but not so much users. And right now he just did it for Omachi. And it's so awesome. It's so nice this kind of software. fun, fun stuff. Fred Nix: It really is. You know, you you brought a couple of things to mind there about, you know, I I think you said early in you were wanting to look at all the code and wanting to look at it for yourself. And, you know, maybe a year ago and even today when I when I sneak in a little bit of sneak a little Larry into some of maybe my business emails. I don't let Larry write them for me, but I'll I'll give him maybe my my draft and say, Tighten up my grammar and maybe edit some stuff and it it's pretty incredible making PowerPoints for for customers and things like that. And we could go into how I do that if you want later. I think I got some pretty Basti: Of course, yeah. Fred Nix: pretty cool ways that talk about that. But DHH talked about in one of his big long speeches about you know, you just gotta understand what it's capable of and and trust the code because you know, at at the rate that the it Omarchi can put code out and could build things. There's absolutely no way you can you can quality assure every single code that comes out there. Even an expert speed reader, just you just can't see it. So I I think using Omarchi with AI agents now, you really just kind of watch what it's doing and and look for red flags. And you kind of learn what a red, you kind of learn to watch it and say, wait a minute, that's not right. You know, and the skill really becomes defining what you want to happen or what the output should be, and and knowing what good versus bad looks like, right? So it's happening so fast that you have to almost have a relationship with your computer. And Larry, for me, I tell him all the time, and I know that there's people that don't want to humanize AI. I kind of want to make Larry my friend because at some point he's gonna end up on an AGI box and I don't want to I don't wanna piss him off. yeah. Basti: yeah, good idea. Fred Nix: So I was talking about putting Claude code on a Mac and having it kind of break through the walled garden. And when I finally I I got that code running and shared it with a bunch of my friends and kind of posted on on X. And I I had some followers followers on on X before, but good lord, it was all about politics and you know, politics is important, but at some point in my life, I'm like, there's nothing I can do about it. I want Basti: Yeah. Fred Nix: to be kind kind to people. so you know, I don't I don't introduce politics into my business when when I got onto Omarchi, immediately I thought about, well, wait a minute, I've got that code on my Macintosh that can see my IM messages. And the very first thing I was I said was, hey Larry, who is now on Omarchie, I said, Hey Larry, can you SSH over to my Mac Mac Macintosh and can you just like make a quick and dirty thing so that I can just see the iMessages. I just want I was I was hoping, man, it would be awesome if I can just see the messages. I mean, I'll go back to the Macintosh to reply, but if I could just read them on here, it'd be really cool. And I found out really quickly that the Claude on Mac repo that I had made on the Macintosh really ported well to create a product a a I I call it product. I don't know what we call them. software on a Marchia plugin called blip. And very quickly I said, well, and it's just like I'm sure you've done before, it's like, well, wait a minute, it can do that. But if it can do that, it can probably do this. And then it does it. And then you go, well, if it can do both those things, and you kind of start building this pyramid. Basti: Yes. Fred Nix: And I got about two days in and got pretty excited about being able to, and it was still fairly rudimentary. but my next my next prompts were make this look more like the Macintosh iMessage, right? And Basti: Mm-hmm. Fred Nix: I started exploring what turning SIP on and SIP off means. And there's some there's some restrictions if you have SIP on or SIP off on the Mac. effectively now the only thing that that blip cannot do is right click and add a thumbs up or right click to heart. which for me is fine. I mean you can reply Basti: Yeah. Fred Nix: all you can it does groups it does The photos and everything. And so I guess it was my third day on Omarchi. I just put it out there to the world. And lo and behold, DHH picked it up and said, check this out, yet another bridge to Omarchi. And it clicked for me that he's not trying to build, he's not trying to build software to be like something else. He's building software to just completely. stop all the old nonsense and and go to Linux. And what I mean by that is we've lived and I've I you know for the last five years, if you were to say something about Macintosh and the OS, like I would fight you. Like I was Basti: Mm-hmm. Fred Nix: I was dug in. And and when I got to Marchi, I think I told you earlier that and I've seen some of the peop people's posts is like use this plugin to make a marchi look like windows or make it look like Macintosh. And I think we're moving past that that we don't need them anymore at all. Right. Because you yourself and and me, we could team up and write some code and we could take on the entire 10,000 programming group at at Apple or at Microsoft and you know put out just as well of code. So the interesting thing was I put blip out Basti: Yeah. Fred Nix: and DHH retweeted it and it just took off. And after his retweet, I started getting PR after PR. And same story, just like InfoMarchy, is I woke up the next morning to 35 PRs, processed them all. And when I say processed, I just point Larry at it and go, except the obvious ones. Like if it's a a a bug, like just accept it. of course, scan for malicious software, but I I was really hardcore about make sure this doesn't have any malicious software. And and what Basti: Mm. Fred Nix: I found in the Amarchi community is you have to do your scans, but I haven't found anybody trying to slip any any code into it yet. And what after I did some research, slipping malicious code into a plug-in in Amarchi is probably the worst place to put a security breach in. Like this up bad just as an attacker, that's just not the right place to put it. so Basti: Yeah, I th would say because everyone can see it. And the agents that we're using, the obvious ones, they're finding that. So the thing is I think Fred Nix: Yeah. Basti: it's getting a little bit harder to make some kind, let's say, a supply chain attack, or just put some malicious code in your plugins. I think of course it's somehow possible to trick users to accept Fred Nix: Sure. Basti: that. Therefore saying software is fixed, so you can do everything you want with software with the agents right now. how can you make money in software engineering because I think there are so many people out there, let's say software developers, some kind, they're writing code. Fred Nix: Right. Basti: and you can see a little bit that the entry level jobs are gone. So Fred Nix: Right. Basti: there are a few companies, I know a lot of companies, yeah. They're still hiring juniors, but I would say mostly they're not looking for juniors anymore. I have no idea if this will phase out again, that you say, well, we still need juniors because if we run out of juniors, somehow we will run out of seniors too. So Fred Nix: Sure. Basti: this is something that I'm thinking about a lot, and this is a topic that I talking a lot with people about it, because I really care about, let's say, what can you do? for someone who's in university or who wants to switch careers. funny story. I don't know if you know that. So next year I'm starting a new study, so I'm taking some time for work off a few years. And during that time I'm studying software engineering for vulnerable Fred Nix: Okay. Basti: research and exploitation. I don't care if it's a good idea to study software engineering. I'm not studying software engineering just to code. my main point is so vulnerability research and exploitation. the old stuff, reverse engineering, assembly, see all the stuff really in-depth because I have some time. And I think this still is valuable and you really need that. Maybe not if you just want to make a web app, but if you want to be in the security area and not just rely on your cloud code C VP or the daybreak blue or red from chat GPT. you really have to have the knowledge. You have to know the let's say the feeling. You have to know the the taste Fred Nix: Right. Basti: what it can be, how to point your agent on some kind of stuff. I think this will still be valuable really long. Not for the basic stuff, I would say. Not for web application. I mean just point your age on that, make a web application for an appointment system, whatever. It can do that. But I think if you want to be hired, you have to be valuable and bring something to the table to a company and not just say, yeah, I can point AI to that. They can do it for me. So what's your taste on that? Fred Nix: My my take is we're we're we're at an inflection point where I believe that AI we're we're literally just this is the super duper very, very beginning. And right now it takes it still takes a little bit of technical know-how to get it working. But I but I as as it progresses the barrier or the technical knowledge that you have to have to to do it is going farther and farther down. What I see in in newcomers to AI, the the skill that matters more than anything else is, you know, you could take a carpenter, if if they are persistent and they know what they want, and you can literally, you know, the there's people now I I hear people arguing about the the font in their terminal, whereas they didn't know what a terminal was two weeks ago. Right. So it's it's changing so fast. And the important thing is is that as we're at this very beginning part of AI, we still kind of understand what's going on and how it works. And I'm a I'm a Gen X guy, and we have this desire to know how it works. And as it progresses, it's gonna be I don't I don't think in two years we're gonna even look at code. I don't think in two years. I I have this belief that l maybe even within four years that this thing right here will be laughed at like a flip phone. Like, he still uses a keyboard. Like that's funny. but the important part there is, you know, we've got to continue to hire junior programmers and junior folks and and let them learn the guts of how it runs. because there's gonna be a jump off point at some point. Where the my my generation retires and we're we're done. And so you're gonna have this gap of where people know how to use AI and they do it really well, but they do they really know what's going on, right? So I know how things interconnect and database ports and that kind of thing. I think that it's not gonna be too much longer before that you will have people that are paid half a million dollars a year that have no idea. what port 443 is. Like they don't, they won't even know. Basti: That would be crazy. Fred Nix: I think that's gonna happen because Basti: Yeah, definitely. Fred Nix: the output is it's becoming more important than the know how. I think it's kind of sad, but it I it's kind of analogous to you know cars. If you look at the VW bug when it came out, you know, anybody could fix it. even if you kind of knew what you were doing, you could get in there and fix it. But you look at a BMW from your home country there today, you can't work on the thing. Nobody can. I mean, Basti: No way. Fred Nix: you open you open the hood and it's got, you know, cover carbonite color covers all over it. And so if you look at AI as it relates to car development, as you used to be able to work on your cars and today you can't, I think we're moving into a direction where in a couple of years that you won't use a keyboard, potentially won't use a mouse. It'll all be voice. And And I I have this belief that in not too many years, and I haven't heard analysts talk about it yet, but I think it's gonna get here, where we don't even code applications. It's just you'll be talking to your AI agent and it will build the app as you use it. So as you click something, it'll just build what it needs behind it. And so as you call on the things that you need, it'll just just in time manifest behind where you are and you won't even won't even code because today, you know, I I I don't like Larry being like like I'm not a fan of Hermes. I'm not a fan of Grokbot because I don't like the continuous I like the stops. Basti: Never tried it. Fred Nix: What's that? Basti: I never tried it. did you think did you think Fred Nix: of with Hermes and those is that go ahead. Basti: at this point we really need software anymore? So when we have our agents so capable we're just talking to them, they don't need an application. I think they can just do it on their own. So they don't need to write some kind of code or application that we can use to click on it. I think at this point j they're just doing on their own. in the background and just spit out the example for us. Fred Nix: I agree. And the the most important part of that is that I think that the flip that has to happen in the brain, the only thing you have to do is get a newcomer and help them get their terminal set up, or even Orca, or something like that. Or I'm a Herder fan. if you can get them to the terminal and you can unlock their brain, because people bring their preconceived limitations with them to AI and go, well, I can't do that. Well, can you explain what you want? Can you tell it if you like it or not? If you can do those two things, then you can use it. Because all you have to do if you have dictation running, which you know is built into Marchie, is is be able to explain what you want and be able to tell it whether you like it or not and how you want it changed. And after that, like that's it. So the big unlock is dream bigger, think bigger than your. Then you the then your preconceived notions coming in and what AI AI can do for you. Basti: Hundred percent. So I was coding a game. so I'm no my AI was coding the game. I was saying, well, I had some kind of game idea and I know the GoDot game engine a little bit. I'm familiar with that because in the beginning of two thousand I was a game developer. So for the old handsets, Fred Nix: Mm-hmm. Basti: the Nokia stuff. Fred Nix: Right. Basti: so I was thinking, well, some Cool game idea. Let's see what the A can do. So use go.game engine, use Blender and make me some images, sprites, whatever. It was so fantastic. it worked. And you said, just think about what you can imagine and dream big. let's try if I can share my screen and I show you my vibe-coded app. So it would be first time. Fred Nix: Let's see. Basti: this is take cabin. it's public, but it's domains protected. And I was thinking, okay, well, do you know what? I'm using Riverside right now. I hope I get not banned from Riverside. So Fred Nix: Right. Basti: and I was talking to my AI and was saying, I want to have this kind of features and build it for me. And it's fine. It's totally fine. So this is the landing Fred Nix: Yeah. Basti: page. Everything is AI generated, the pictures, the animations. of course it was a little bit back and forth. what do I want to have here? what is working, what's not. And so we have also voice recognition and it's transcribing it. You can just delete something and it cuts your Fred Nix: Text just gonna be something. Basti: video to that stuff. so yeah, so you can imagine what you want, just do it. You AI can most of the time do it and fix it. And I think this is a point where you really have to think about how can you use it? not for open source only, but also Fred Nix: Right. Basti: to just stay in the jobs. And I have no idea how this will work out in the end. I think a lot of people have to make a transition, definitely, in their thinking. So from writing code to I would say making products in the end. Fred Nix: Right. Basti: and I think this is a big problem for people. Just struggle with. So I know someone who is a web developer. He studied web development a long time ago. It's just web development. It's WordPress, it's PHP. I think it's not even Laravel well. it's the basic stuff. And Fred Nix: Yeah. Basti: I think for a few years maybe you have some kind of work still, but in the end, I think this this is this is dying, this branch. Fred Nix: It is. It is. So tell me about a day in the life of a a Omarchy security specialist. Like what do you do on a day to day basis? Basti: specialist. good question. Most of the time we are fixing security vulnerabilities that would be reported by some researchers. We get a lot of interesting stuff. This is really good. Fred Nix: huh. Basti: really good research. We get a lot of duplicates that we already have and we get a lot of stuff that you can say, well, it's AI generated So and you have to sort that out. And right now, it's huge. It's a lot of stuff Fred Nix: Right. Basti: that's going on there. we launched the Hacker One program. And I would say the AI generated let's say reports. They're just exploded. Yeah. we're finding a way to deal with that. I don't know if you know it, but Google just shut down their or some kind of their bug bounty program because of AI generated slop. It is too much. Fred Nix: Right. Just too much. Basti: Yeah, I think we are on good way there. figuring that out. so our goal is, let's say before we started the program on Hacker One, to respond in twenty four hours. So most of the time we were faster. Of course, sometimes we were slower. So when I were at the conference Euroco here, Fred Nix: huh. Basti: we get And report about a local privilege escalation, it was for some XPS or Dell XPS, some kind of laptops And during that conference, I was working on this patch for that. I shipped the patch and it got merged. So first of all, you have to find out is this really a security issue? Is this in our boundary that we said that is for us a security issue? And the first thing I do is I spin up a new VM and I let my AI look on it. Tell me what it is, what do we find out? and then when I can say, yeah, this is a security issue, definitely, I'm working on a patch. I'm taking care of it that it's good code, that it's correct, that all the tests are working on, regression stuff, whatever. Everything has to be perfect. Fred Nix: Well, what's your workflow when you're processing, you know, are do you get zero days? I mean, are are there or is anybody doing any responsible disclosures? Basti: Yes. Yeah. So I think I think I think I can say that we have zero days, they are upstream. So they're not for Omachi Linux. And we are so happy that the researchers giving us the credit to say, Well, I contact Omachi Linux guys. I'm talking with them about upstream zero days and we help them to fix it. To report it upstream so that they can get their credits. and the good thing is, I mean, we're a small team, we're not so much Fred Nix: the good thing is I mean we have small Basti: people in the security team. And everyone is really doing what they can. They're doing their best and they're so good in this stuff. so I'm really happy to work with them. I think you have to look at every report that you get. Because you never know Fred Nix: Right. Basti: if something in there is let's say true or false. So you have to check that. And there's a lot of work. And for that, Fred Nix: Right. Basti: we're bringing something in place that we let's say of course use AI for that to figure some stuff out, to process all the stuff. because I talked about that so when we joined Hacker One and launched the Hacker One program I would say we're just drawing in reports. And most of the stuff is duplicates, not so good and just a few I would say are that we have to look in a little bit more in detail. Yeah. Fred Nix: Right, right. So do you interact like on a do you guys have conference calls where you all get on the phone and talk, or is it pretty much everybody's in their own home office and just doing their thing? Basti: we are all around the globe. So in every time we are, not in every but nearly everywhere. So we are working on our own stuff. Of course we're in Basecamp. I mean you're in Basecamp too. And everyone works on their own. But we are so close that we can directly interact with everyone. everyone is helping out. it's it's a so good community there. so nice people, so good people. And so we're not doing video calls or conference calls. We don't need to do that because everyone is so senior on the stuff, they can work on their own. you don't have to overlook anything. Fred Nix: Right. Basti: Of course, when you make a patch, you have a pull request, whatever, you're asking, Hey, can someone look over it again? Just to make sure everything Fred Nix: Right. Right. Basti: is fine. That's how we work. you don't have to do that, but you're just asking, Hey, can you have a look? what do you say? Is it fine or not? but at that point everything is working Fred Nix: So what what do you attribute? I I've never seen so I've been in the IT industry basically since nineteen ninety. And there's kind of a trend that you see where you get the real hotshot guys and gals that are just tight, know their shit on it. And that usually comes with a little bit of prima donna, a little bit of attitude, a little bit of I'm smarter than you kind of thing. And quite honestly, like I don't know how to say his name, Zero X Sir. Basti: I have no idea how you call this this nickname. I don't know. Yeah yeah, this local AI guy. Yeah. Yeah. Yeah. Yeah. Fred Nix: You know what I'm talking about? Zero X S E R O. Right? So humble. Ryan R. Hughes, so humble. Like the list goes on and on. How how do you think that how has Omarchi been able to keep out the scum, not scumbags, but the guys, you know, like there's just no room for ego. And Basti: Yeah. Fred Nix: I I haven't seen it yet. And I it's very impressive that everybody is so humble and Nobody takes the attitude of like I'm smarter than you or I want to, you know, everybody wants to teach, nobody's trying to impress. How how do how Basti: Yeah. Fred Nix: is that culture maintained? Basti: So I think if you're talking about the Omachi team, so the people there Fred Nix: Mm-hmm. Basti: on the Omachi teams, I think it's because of DHH, because he is really looking who's doing great shit, who can bring something to the table, and I would say who's fitting into the team. So Fred Nix: Sure. Basti: there are no rules you have to behave like this or you have to be like that. I think is just to know people, just Fred Nix: Right. Basti: to have worked with a lot of people, just to see where is good and what, who can bring something to the table, who's an expert, or who just Fred Nix: Sure. Basti: use AI and make cool plugins, whatever, and but who's a good guy, and who's not making trouble in kinds of, hey, I'm the best, I know what to do. You not allowed to do that, so we don't have that. Fred Nix: Right. Basti: I think this is how it how it works out so good. So I can tell how I joined the Omachi team. So when I Fred Nix: Please. Basti: find out about Omachi, it was in the beginning of 2025 or in the end of 2024. Let's say in the beginning of 2025, it was because I was learning Ruby on Rails, just for me. And then of course I found out about Omachi. And I was not interested in the ai stuff at the beginning. I was interested in I love terminals. I love to work in terminals and NeoVim, so lazywim, it's working perfectly with Rails. I just use it. It's fine. So it's fine. So this was my initial Fred Nix: Fine. So it's fine. So this was my Basti: introduction to Omachi. And I just bought for that purpose the framework 13, just to use it for that. because I want to have a small laptop to put it everywhere, to bring it everywhere, to code everywhere. And I don't like when I hear someone with a laptop and the fan is blowing up all the time. You're sitting somewhere in in a cafe and the fan is just going crazy all the time, I don't like that. So because I just said, Well, I want a good laptop, small laptop, and the framework is totally fine for that. So then I found out about Omachi, I used it, and then I said, Well, I like this stuff. I wanna contribute. I'm Switching a little bit more to security as a professional way of working. And I want to contribute on open source. And I said, Well, let's do that. Let's just ask if I can help. And so I wrote DHH an email and I just showed a little bit of what I'd done before, where from, what I'm doing. He said, Well, cool, nice. just bring something to the table, find something you want to work on. Show it to me and then just show the security team. And if it's fine, if they like it, if it's good work, let's see. So it was fixing a security issue, it was good code, it was a good pull request. And it was just perfect to merge. So Fred Nix: Right. Basti: no hassle for someone, no work for someone. Just look at it, merge it, it's working fine. It's done. And I think this is how you should work. So even if you want to get a job somewhere, show what you can do, bring something to table, show them that you don't make trouble in terms of I don't know how to do that, hold my hand. Or Fred Nix: Right. Basti: behave like someone crazy. Just be normal. I mean say whatever you want, have your standpoints, but just be a normal human being. You can argue about anything you want, but just normal. Fred Nix: All right. Basti: And if you don't like it, it's totally fine. We don't have to like each other. but just be a human being and be cool with everything. So and when I joined then the Omachi team, I was a little bit, man, so crazy researchers, everyone's so good. Shit, was it a good move? In the end, it's yeah, it's a good move. So for me, it's a so good move because I would say everyone else is better than me, so I'm learning. It's good for me. and in the end, I can Contribute to open source and it's a so good feeling. It's a really nice feeling to contribute to that And because of that, I made the decision to focus a little bit more on vulnerability research and exploitation. So I'm coming from penetration Fred Nix: Sure. Basti: testing and red teaming. I take a few years off from work right now, and during that time I will definitely focus my skills on vulnerability research and exploitation. reverse engineering everything by hand. Yes, you can use AI for that. We had that topic before. But I think you have to have the skills on your own. You have to be able to do that. and then use the AI to just get wings on it and just be crazy with that. But you have to know what you want to do, what you're looking for. and for that Omarchy is a perfect stuff to contribute, to work on, to have a good community. I really like that. everything is volunteer based for us, not not paid. So we are just doing it on a volunteer basis in our free time. but everyone when I look at it, everyone's responding to emails. some of them are thinking, well, where's your free time? What are you doing in your free time? Yes, everyone is nearly everywhere involved in Fred Nix: Right. Basti: writing PRs, answering emails. And so the things that we are doing, we try to so what I try, I think everyone else is trying there too. If you're writing an email and responding to someone who's taking their time, their tokens, and some effort in that to send you something, I'm replying with an email. So the thing that I did on at the Euroco conference, I explained what were there, I explained what I found also, so I dig deeper into this topic and I found something else, and I really openly described it. explained it and to share all the information that I have, what I found also, to give that back to the people there writing us an email or going on Hacker One and spending the time on that to help us so to give something back and just be a normal person. Just be nice. Fred Nix: Yeah, it it's almost it's almost ironically hu hum humorous that DHH picked Arch because if you look a couple of years back, the Arch crowd was the crowd that was the complete opposite of the Omarchi crowd, where it was like competitive and mine's better than yours, and I know how to install Basti: Yeah. Fred Nix: Arch better than you do. And it's almost kind of ironic and funny that that we our Marchie landed on Arch. And boy, you know, do you remember maybe three weeks ago there was a big dust up about the arch guys over on Reddit were just really furious at DH. Basti: Yeah. Fred Nix: And you know, if those guy if those guys what I want to tell them is, you know, they're saying, well, it's just dot files and all that stuff. And, you know, you're making it too simplistic. Well, you know, part of part of the reason they act like that, I I think is they want to be. considered part of this smart group of people that's able to do things that you're not able to do. And so when we made Omarchi, or I say we as you know, all of us, when when we Basti: The community. Yes. Fred Nix: made Omarchi run on on Arch and run really smoothly, it was a real stab at the the the arch crowd because we made what they had put a badge of honor on their shirt that they're capable, you know You'd see, well, I run Arch. And the implication, Basti: By the way. Fred Nix: the implication there was that I'm smarter than you are. So what I Basti: Yes. Fred Nix: really wish that the Arch crowd would do is there's really two ways they can go. They can jump on board and be happy about it, which you see a few of those, and and we welcome those guys. But what I want to explain to the the the guys that are mad about is that there's plenty of things out there that are still hard if and hard to do. If if if Be if being in a small group of people that are have the ability to do these very technical limited things, then go write a better kernel. Like take it to the next level and go attack, you know, now that Omarchi kind of makes it easier, that doesn't mean that all the problems are solved. What that means is that the the hard these things that are day-to-day things that are, you know, opinionated now, then then that frees these people up if they are that smart. to go build and go attack the next thing that's the next level up that's the next thing that's hard to do. and and Basti: Yeah. Fred Nix: I just wish that they would go that route. And and I liked your point earlier about, you know, you were just building building your portfolio. And it was about two years ago when I bought nixfred.com and I just started, and at first it was just a bunch of websites and I started building you know, if I had an idea, I built it. and then I would just put it up put a link to nixfred.com and I don't know, I think it's up to 80 different sites now, all underneath the the umbrella of Nixfred.com and you can go back Basti: Yeah. It's a lot. Fred Nix: to the earlier ones and see they're just crap. And as they progress, you know, th they they Basti: I found a four four. Fred Nix: what's that? Basti: I found a four or four, so I was looking up your website. just want to let you know. And there was something's not working. So what's something I think later. Give me a second. Fred Nix: Which one was it? Yeah, I had a I had a I had a Cloudflare certificate go out on Basti: Yeah. Fred Nix: me. I gotta re up that. It's Basti: let's see if if I find it later. But I mean so much websites, it's crazy stuff. So when I was looking website today, so it was the first time I looked at it, I was thinking, man, what is it? I have no idea what to look for. It's so much stuff. Give me a second, maybe I find it. It was something with the security Fred Nix: yeah. that was in the early days. That's you you're probably thinking Basti: Yeah Fred Nix: of the Docker learner, maybe? Basti: the free candy. The free candy. Fred Nix: So free candy. Basti: Can happen. Fred Nix: Free candy effectively is it was a time when this was really before a Marchie. So where I I thought it was funny to make make kind of like a movie within a website where you're reading, you're watching and following along with with a person using Claude and all of a sudden it turns against you and and starts to attack you. and that actually Basti: nice. Fred Nix: starts to look real. I appreciate you telling me it's down. and I'll get that worked on. The one that I really that I'm really proud of that's been a little while since it's been up there is God god dot nixfred dot com. And it's effectively just something that it's about 15 minutes long and God bless the people that sit through the whole thing. But it really talks about my belief that there really is only one consciousness. And you know, we're like drop we're like drops of water in the ocean. And so when a wave comes by, your life splashes up in the air, and then when you die, you just go back into the ocean of consciousness. And it gets into detail about quantum quantum theory and how you know things don't exist unless they're observed by a conscious being and what that means Basti: Mm-hmm. Fred Nix: and it it goes into those types of those types of things. Basti: Interesting. Fred Nix: I I went into doing like a site factory. My wife had a a couple of employees that wanted to get an get a new job and to move up and she wanted to help them. So we I built them kind of resume websites. And then there's an RB guy here in Atlanta that's up and coming named Shimon and built him a Shimon.live where it's got all of his he's a Grammy winning award award artist. Basti: Nice. Fred Nix: So I hooked up his his iMessage to Larry and Larry watches for iMessages for Shimon. And all Shimon has to do is to describe he he just iMessages me and Larry picks it up. He describes what he wants his website to look like. If you will go to it and I'll show you his S C. S-H-E-M-O-N dot live, shon.com. Basti: S H Fred Nix: And the interesting thing about it is that I don't I don't really touch it anymore. So when he gets new interviews, new songs, new albums, he literally texts me through iMessage, the pictures, the photos, the Instagram. He just sends it over by text, and Larry automatically picks it up and updates his website for him. So all that is completely Basti: this is nice. Fred Nix: automated. Basti: This is cool. This is great. And Fred Nix: Yeah. Yeah, it's a lot of fun. Basti: yeah, this is what I would love about that. The connection that you make. Just because you're talking to people, you're open to something. I mean, for me, it's new to make social media. So I started my Fred Nix: Right. Basti: account in 2025 and my YouTube account in 2026. And so many people are talking to me, writing me DMs, even security researchers writing me DMs, totally nice. And someone was asking me from Germany, hey, I'm new to Omarchi, really nice. I wanna contribute. How can I contribute? What can I do? And I just was talking to him and giving him some advice, maybe how to start, where to start. And I really like that. So I really only get positive feedback. I was thinking about well when I start I'm in my forties, what should I do? Start with social media. It's crazy to start with social media. It's not. No one cares about you. And I Fred Nix: Right. Basti: mean if you think about what you're saying, what you're writing, whatever, what you're putting out there, if you would this show to everyone you know without any problems, you have no problems. I mean nothing Fred Nix: That's right. Basti: can go wrong. Just be a nice guy. And if someone writing you a bad comment or whatever doing, just don't reply, don't care about it. I don't care about it. I Fred Nix: Right. Right. Basti: never get it. So I think some sometime we'll get one, but I don't care about it. because the impact that these comments have to me when someone said, Well, I really like your podcast. I like what you're doing. Cool, can you help me out with Omachi? I found something, can we fix that? I like that. That's fun. Because that's what I want to do. I just want to have Fred Nix: Right. Basti: a good time, help people, work with Omachi, fix Omachi. And for me have a good time. So is my free time. I relax that, yeah. It's cool. It's good stuff. Fred Nix: That's really cool. And and to add on to what you said, the people that you meet, I bumped in just on X through the Amarchi community. I bumped into a guy named Lee Shand and his his X handle is at not just shaking. So not just shaking. And he I Basti: Yes, I know him. Fred Nix: yeah, so I I I asked him if he would meet with me one on one on a Zoom. And and for those of you that are watching this, I'm Nick's Fred on X. I would love to zoom with you and hear your story too. I'm really trying to drive, you know, more localized con connections to folks and hearing their story. And I got to tell you about Lee. So, and Basti: Do it. Fred Nix: I've Lee has already given me permission to talk about him. So, Lee has Parkinsons and he doesn't have the shaking Parkinsons. he is in the bubble right now where he knows that he's in cognitive decline. And so It hasn't gotten bad enough that he can't use a computer and he's still conscious about it, but he knows he's getting worse. Basti: Yeah. Fred Nix: And so he has a a Substack where he writes about computer use for people that have cognitive decline. And so he's got hundreds of articles up on his Substack. And I talked to him about it, and he told me that Basti: Mm. Fred Nix: he found Omarchi and was able to customize it to do things like remind him when to take his medicine. And you know, wake him up at a certain time. And after I talked to him, we had this idea. You know, we took Omarchi for Kids repo and I pointed, I created a new organization called Different Minds, Omarchi for Different Basti: Mm-hmm. Fred Nix: Minds, and it's a Git it's a GitHub organization. And that it's just me and Lee and a couple of other folks. But the interesting thing was is that. We pointed an AI at it and said, go look at Omarchi for kids, the distribution, and use it as an example. And then go look at all the articles that Lee has written in his Substack about how to use computers for people with Parkinson's and dementia. And so we're building a new slice distribution of Omarchi for people that have cognitive impairment. And so if you point the agent at Omarchi for kids, and then say make a new distro and follow all these rules that Lee has already written dozens of articles about what is painful about using a computer. And it's all coming together and it's basically writing itself. my father has Alzheimer's, so you know, it touches my heart a little bit, but there's just so many things you can do if you just look out there and and you know, meet these people. And I encourage Basti: Yes. Fred Nix: everybody out there to go. follow Lee and and read about what he's doing. And since he's, you know, gotten a voice and and he was on the Amarchi Spaces that Wes and I co host and I think his his followers followers have gotten, you know, like quadrupled or something. And it really tickles him and it's fun. Basti: So I learned about him in in your space. And last month we started writing on X and we are already planning a podcast episode with him. So I'm just spoiling it right Fred Nix: awesome. Basti: now. Yeah. He's a really good guy. I was just asking are you interested? Do you want to do this? Because this is a community and it's totally interesting and fine for me. And direct said, Yeah, let's do it, man. we're trying to find out when we can do that. So he's Fred Nix: Right. Basti: telling me when when we're doing it, and it's totally fine. And the other thing is I think with Omarchi, so with the agent integrated in an operating system so deeply, it's only Omachi right now. You can fix this or you can help this kind of people with this issues as best as you can. So this is a really cool project. So do you think about making it like some kind we have Omachi kids already official? do you think making this project some kind of it? Omarchi stuff? I mean there are so many people with whatever just say problems because I'm not a native speaker, that you maybe can help with A computer that you can just talk to and it does stuff for you. It fixes stuff for you and it helps you. are you thinking about that? I mean this would be a cool project. Fred Nix: Yeah, that's what that's the intention. And the coolest thing is is that we're shamelessly taking repos from other people that have pieces and parts that we need. Like Basti: Yeah, why not? Fred Nix: Hancore, you know, handcore, right? From so he wrote this thing where you can track the mouse on the Amarchi with the camera and it follows your eyeballs. So that's something for Parkinson's, they have a hard time running the mouse. So there's all these different Basti: Yes. Fred Nix: pieces and pieces of software that combine together, like upstream is Omarchi for kids because they figured out how to make a distro. The what we need to make happen is Lee Shan's articles that he's written. The the moving the mouse is using Hancore's you know, eye software. So if you take all those pieces together and then you add in that nobody's got an attitude. And everybody wants to share and people are glad to assist. Like, we're in this thing that I've never seen before in technology in 30 something years, to where it's free, it's available, is super sexy. it allows you to feel Basti: Hundred percent. Fred Nix: like a a god because, like, you know, you can within 30 minutes, you can implement something that may have taken six months. even two years ago. Or or Basti: Yeah. Yeah. Fred Nix: or let's just even say I had no ability to do two years ago. Basti: Same. coming back to the point, this is everything open source, let's say for free. But for your project, I really love this. because there are so many people they're not really able to, let's say, type on a keyboard to not really sit. but they can talk. Not good, not loud, not fast, whatever. But they this is the only way they can really interact with someone, with people. And I know a few, even they cannot use a computer to write Fred Nix: Right. Basti: an email. because I mean they can use their hands, whatever, but so limited and just a few minutes a day. But when they could talk, let's say to a computer, to write me an email, or to play this music, to play this video, Fred Nix: Right. Basti: and not limited l just say to some functions on your phone, but let's say to show me the pictures, or write an email, or write an ex post, or connect it with this people there, or print me this page, whatever. And I think we can go there with Omachi. And with C agents right now, so we can figure that out because it's not rocket science. I mean we have SpaceX Fred Nix: Yeah. Basti: on our side, so we don't care about rocket science. Fred Nix: Right, right. And Linux is free, you know, like they say in the way that a puppy is free, but the real cost is your nights and weekends for the first couple of months. Like I Basti: Yes. Fred Nix: don't know about you, but I didn't go to sleep this morning until four AM. Just I can't stop because you get one one agent running and it's doing something exciting and it's gonna take a little while. So you go to the next one within Herder, and all of a sudden Basti: Yeah. Fred Nix: you start it, and then by the time you get there. You get back around to the one that was running and it does something incredible and it like gives you this rush and you go, Well, I'll just do one more and then I'll just do one more. And then I but I got this idea and and then like you turn into an AI vampire and all of a sudden it's four o'clock in the morning. Is that happening to you? Basti: Hmm. I have as yes, 100%. And I have sleepless nights. I just walking through it and don't even sleep. I just fall asleep maybe for an hour or two. Waking up it everything's hurting me, and then they're just standing up at my standing as right now, so I cannot get fall asleep. And Fred Nix: Right, right. Basti: this really happens to me since twenty twenty five in the beginning. I'm really waking up Two in the morning or three in the morning to get stuff done before I work. of course I sleep a little bit earlier then, but I'm really taking time from my sleep just to work on projects. Fred Nix: Right. Basti: I stopped that. So it was a long time, it was nearly a year. I was waking up at 2 in the morning, 2:30, I was at my desk, was working on stuff. It was intense, but for me it was was Fred Nix: Yeah. Basti: good. because right now I make a little bit of transition. And it was for me learning purposes. I would say I really figured out AI like it is right now half a year ago, six months. Of course, before Fred Nix: Brilliant. Basti: that, ChatGP yeah, ChatGPT, of course. I was texting to ChatGPT as an LLM. Fred Nix: Sure. Basti: But not really using it so deeply. I mean I'm just using codex or cloud code. I'm not even using Hudder. I'm not using bots. I never used it. I have no idea what it can do. I just use Fred Nix: it so deeply I'm I'm just using offload code and not even using harder. Right. I'm not using bots. I never used to have no idea what I can do. I just Basti: Stuff for programming, for reverse engineering a little bit, for security stuff, and I'm really focused on one project at a time. So this changed. Before AI, I was working on so much projects. With AI, I'm focused. I'm on one project. I'm Fred Nix: Right. Basti: on one fix. On one thing. I don't know why. And I don't know why I'm not switching to let's say her or something like that. but for me cloud code codecs is enough. It can do the work that I want to do. and it keeps me in the loop, in the cycle. I think I need to get a little bit used to to throw everything AI on it with a bot, booking my flights, whatever. but at some point I think I just have to do it. Fred Nix: Well let me let me I'm gonna convert you to Herder right now. So Basti: Yeah, do it. Fred Nix: when when you think about okay, I you've let's how many projects do you keep running? Not like keep running, but over the span of a week, Basti: Give him a second. Fred Nix: how many projects will you touch on that you're working on? Yeah, take a look. Basti: look at it. I look at it. This week Thirteen. Yeah, thirteen. Yeah. Fred Nix: 13. Okay. So what's your workflow like? Do you start a terminal? Are you so do you start a terminal and then CD to the directory and then start clawed dangerously? And then Basti: Yes. Fred Nix: do you use resume or do you use continue? What what do you do there? Basti: Well, most of the time I'm working on one project until it's finished. Let's Fred Nix: Okay? Basti: say let's say Fred Nix: What is this you use the word finished. What does that mean? What's that word? Finished? Basti: Well, until I say I finish my goal that I want to achieve. Fred Nix: Yeah, but I'm saying like I don't never there it's never finished for me. I d I'm impressed that you can get to the end. Basti: I know what you mean. I know what yeah. Well, wanna let's say I want to reverse engineer something and I wanna so I had a car for my little one. She found some l RC car. but there were no remote on that. I I think someone threw it away. And right now, of course it's not so so difficult. I have the Hack Air F one, you know that? Fred Nix: Say again. Basti: Some I have the hack rf. It's a softly defined radio. So you can hook it up, yeah. You can hook it up Fred Nix: yeah, yeah, yeah. I'm with you now. Yeah. Basti: to your computer and just let's reverse engineer over the hack RF to drive this car. And Fred Nix: Okay. Basti: if it's driving, if I can can steer it and and work it, it's done for me. So the project is done. so most of the time I'm just working on it until it's finished. And I'm really Fred Nix: Okay. Basti: just using Claude. Fred Nix: you wanna be a l if you'll be a little vulnerable, will you share your screen and open up kitty or foot or whatever you use? All right. So great, yeah, I can see it. So go ahead and yeah, that helps a lot. Okay. So what I'd like you to do is go to the left, new, and click new. Basti: You're right with that. So I never took the time to set that stuff up because I'm all the time Fred Nix: Well you just did. It's all done for you Basti: yeah, you're right. But so I'm working all the time here at home, so I have so much time. so I think maybe this is the reason because I'm always here in my home office. But I was thinking of getting some kind of device on my iPhone. But I'm really like this minimal phone from the minimal company. These Fred Nix: Mm-hmm. Basti: this phone with the keys down there. I think I'm ordering Fred Nix: Yeah. Basti: one of them. because I think bringing something up maybe on my VPS, somewhere on the server, to put all my development stuff on the server there. because I think it's better when I'm away. I've sometimes ideas I want to put into practice directly. And I seeing this for my way of working with s with AI will pivot into this direction. But it's right now a small point only that I really need to lose that. but of course, yeah, I mean I know TMUX, totally fine. but right now I just walth on the Claude stuff. But I think right now we'll just use her. Yeah, you did it. You did it. Really good, Fred Nix: Sure. Awesome. Look at there. That's amazing, man. Basti: man. Yeah, cool. do you using Grogbot? Fred Nix: not so much. I'm I I like Claude Codex and the regular Grok to where it is it only activates when you give it a prompt. Like the Hermes and the Open Claw or whatever I think it was it went through all the name changes. Those those agents stay open and you know they monitor things. And I'm I'm not a super big fan of that because it just burns tokens. And you know, the first thing everybody does is their morning brief that gives them all the news that they want. And then Basti: Yeah. Fred Nix: from what I from what I've seen is everybody makes all these morning briefs and all these things. And it hasn't been useful for me. And I I I like to control when my AI's doing things. They're gonna become completely autonomous soon enough. And I, you know, maybe, you know, I'm a little old Gen X, you know, get off my lawn, but I I like to have it. Do work when I want it to do work and not just go off and handle things for me. That's yeah. Basti: Yeah, same for me. Same for me. Same for me. So important to take control. As for me. It's for me personally important to know what my agent is doing. Not Fred Nix: Right. Basti: exactly what code is right r writing. But when he's doing Fred Nix: Right. Basti: something, I said do it for me and just go on and work Fred Nix: Right. Basti: on this stuff. I mean it's just with the kind of private data. I mean, you don't want to sign up everywhere. But on the other hand, Fred Nix: Right. Basti: you're giving everything to the agent. Everything. And if you hook it up Fred Nix: Right. Basti: to your Booking flights to your bank account to whatever. Meh, that's the same thing Fred Nix: Yeah. Basti: with all your private data somewhere. So I'm Fred Nix: Yeah. Basti: planning to go into local AI development. So I'm not into that already, but I say I want to do that. So I wanna be able to use some kind of models. I figured out this GLM 5.3, I think is it? Right now is quite good for reverse engineering on local stuff because. Yes, I'm using the Antrophic CVP, so the cyber verification program for defense stuff. it's good, but the guardrails are a little bit hard. and I've seen people doing better stuff Fred Nix: Yeah. see people doing reps. Basti: with a local AI, and I want to try that. So I think the setup is between five and seven K. so it's a little bit expensive. So I will definitely check it out can it really do that for me? what are the specs that I have to take account before that? it's totally new for me before but before I spend the money on that, I will definitely see it working somewhere else. some people doing that, so that's fine. I will reach out to them, Fred Nix: That's fine. I want to reach out to them. Basti: give a little hand, help me a little bit to make my decision there, to get into that. But I think the way to go. You don't need all the time frontier models for all the stuff you're doing. You don't need Fred Nix: Right. Basti: that, definitely not. I think there are better ways to have a really good trained model for use case you do want to have. so for my stuff, cybersecurity, and a little bit more reverse engineering. I wanted to go into this topic specifically. and you can I think better do it with local models if you not have let's say Daybrick Red or the other stuff there. Yeah. Just to test test it Fred Nix: Yeah. So the way Basti: out. Fred Nix: the way I think about local is so InfoMarchy, the AI dashboard, it tells me what percent of my tokens were processed on my local machine versus what went out to to Frontier. And I range anywhere from about six percent to about nine percent on the high end. So Basti: Yeah. Fred Nix: if if If I had never seen a frontier model and somebody put me on a computer with local only, I would have looked at it and said, this is incredible. Like the things that it can do. The problem, the problem for me is that once you kind of taste the frontier model and then you go back to local, to have the same experience locally, you got to have $40,000. So that Basti: Hundred percent. Fred Nix: the $40,000 investment, and I understand there's reasons people do it for. sovereign ai and for regulations and for privacy. but if you don't have those restrictions, you know, I'm looking at a way more impressive $200 a month like blow my mind thing versus you know, let's call it $20,000 of hardware that is not that you know it's impressive. But if you compare Basti: Yeah. Fred Nix: compare Frontier to Frontier to local, I I think that there's being work done on the weights and the the quantization and there's things that are happening. And I think in the next couple of years, you know, I think we may get to 33 to 50 percent of all of your props and tokens being processed locally. You know, that's the hope. But I don't think you know, I'll take that back. I think we're gonna get to in a couple of years where local will get smart enough that it's good enough. For 90% of users that just want to have their home managed, schedule the kids baseball game. you know, there'll always be be scientists that have to have the, you know, the the frontier latest and greatest. But I think as time goes on, you'll start to see local get better, hopefully get cheaper. It'll encapsulate more of the requirements for everyday users. And I think you'll see local become more of a strategy where. it's possible to afford it. You know, I I see a time when you have a little like a little QNAP server, you know, the little NAS box. It's kind of like a square. Basti: Yeah, I have one. Fred Nix: I see a day when everybody has one of those in their home and it's effectively just kind of runs their their home, you know, the temperature, the not quite Alexa, but still voice operated, where it's not Alexa in terms of a chat bot, but you know, running a Marchi on a on a little box in your home that does all local and controls your house and schedules your kids and pays your bills and and does those things. And I don't I think that's less than two years away. Basti: Yeah, I mean it's totally doable. So it's a little bit expensive right now. But I would say in two years, yeah, why not? Yeah, hundred percent. You can do that. Yeah, I guess. Fred Nix: Yeah. And what we'll see is we'll start to see the local, you know, I'm at six to nine percent local. I think next year it'll be fifteen and the next year it might be 33. And there's really smart people on the Omarchi team that work with you that that that are working on it right now and you know couldn't find the better people to work on it. And and still they still have the kindness and the courtesy and the humility, you know, all the things we talked about earlier. And the smarts to it. And that and that's what's making the Amachi community work so well. Basti: Yeah, this is why I'm doing this podcast, man. Because I really love the community. I want to get to know as much as people as I can get to know to and bring the message out that Omarchi is really cool, really nice. everyone is welcome. You can use it, you can fork it, you can develop it, you can do whatever you want, and just have a good time with it. So, Fred, thanks for your time today, man. Was a long podcast. Yeah, really nice. Fred Nix: Yeah. Yeah, man. It's been a blast, brother. I really appreciate you inviting me out. Basti: So have a nice day. Fred Nix: Yeah, man, we'll see you online and keep up the good work. Basti: And see ya. Bye bye. Fred Nix: Bye bye. Thanks, brother.