WEBVTT 00:00:11.249 --> 00:00:16.009 Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris. 00:00:16.189 --> 00:00:16.809 My name is Wes. 00:00:17.029 --> 00:00:17.769 And my name is Brent. 00:00:18.349 --> 00:00:22.789 Hello, gentlemen. Coming up on the show this week, well, for the last three 00:00:22.789 --> 00:00:25.809 months, we've been building multiple open source agent platforms. 00:00:26.409 --> 00:00:29.589 The gains have been real, but the friction has been just as real. 00:00:29.729 --> 00:00:34.289 Our expensive, challenging, and humbling journey with open source agents. 00:00:34.469 --> 00:00:37.149 They'll round the show out with some great picks, some boosts, 00:00:37.149 --> 00:00:40.969 and a lot more. So before we get into that and the challenging journey, 00:00:41.489 --> 00:00:44.449 let's say good morning to our mumble room. Time appropriate greetings, 00:00:44.569 --> 00:00:45.609 Virtual Lug. Hello, hello. 00:00:46.449 --> 00:00:48.389 Hello. Hey, Chris, hey, Wes, and hello, Brent. 00:00:48.649 --> 00:00:52.029 Hello. Nice of you to join us on air. Hello, everybody up there in quiet listening. 00:00:52.149 --> 00:00:55.469 Nice to have you along as well. Look at them. Aren't they looking nice today, Wes? 00:00:55.609 --> 00:00:57.129 Wow, dressed up and everything. 00:00:57.409 --> 00:01:02.809 I love it when they do that on a Sunday. Also, good morning to our friends over at Define Networking. 00:01:02.889 --> 00:01:08.549 Go check out Manage Nebula from Define Networking. And go to defined.net slash unplugged. 00:01:08.589 --> 00:01:11.109 You'll get 100 hosts absolutely free. No credit card required. 00:01:11.269 --> 00:01:14.889 It's a decentralized VPN built on the open source Nebula platform. 00:01:15.069 --> 00:01:17.949 And what I like about Nebula is it's built the right way. 00:01:18.109 --> 00:01:23.389 Open source, incredibly reliable, and designed to avoid the usual points of failure. 00:01:24.329 --> 00:01:29.069 And man, is it so great for a home lab or an enterprise. It's resilient. 00:01:29.929 --> 00:01:33.029 It's incredible because you can have these lighthouses that you manage. 00:01:33.149 --> 00:01:35.749 They're public lighthouses. is you can have one system going to one system. 00:01:35.889 --> 00:01:38.429 You can have a giant mesh network. I mean, it was originally built for Slack. 00:01:39.109 --> 00:01:42.349 So you can go big, you can go big, or you can go small. And I just think that's 00:01:42.349 --> 00:01:45.949 really, really powerful. I just think it's, once you wrap your head around it, 00:01:45.949 --> 00:01:46.569 you'll see what I'm saying. 00:01:47.409 --> 00:01:52.369 So why not dip your toe in, check it out. 100 hosts for free at defined.net slash unplugged. 00:01:52.589 --> 00:01:55.669 Nothing else offers Nebula's level of resilience, speed and scalability. 00:01:55.809 --> 00:02:00.009 Get started, 100 hosts, absolutely free. Support the show, our premier sponsor, 00:02:00.609 --> 00:02:03.829 defined.net slash unplugged. Thank you very much for their support. 00:02:04.429 --> 00:02:08.949 Of the Unplugged program defined.net slash Unplugged. 00:02:11.274 --> 00:02:14.754 All right, so we want your feedback for these topics that we're about to get 00:02:14.754 --> 00:02:18.754 into. For example, this episode is based solely on questions that have come into the show. 00:02:19.154 --> 00:02:25.034 But we also get questions or maybe sentiment that is don't talk about these kinds of things. 00:02:25.974 --> 00:02:30.514 Like on March 24th, Matt wrote, and I think it's maybe his first time writing 00:02:30.514 --> 00:02:33.534 the show, You boys seem enthusiastic about AI. 00:02:33.714 --> 00:02:37.334 I recommend you create a new podcast dedicated to AI so it doesn't dominate 00:02:37.334 --> 00:02:41.534 Linux Unplugged. Half the audience can't stand AI. It's very polarizing. 00:02:41.794 --> 00:02:45.114 If you start leaning heavily into it, I'm just going to unsubscribe. 00:02:45.434 --> 00:02:48.974 I thought I'd give this feedback to let you know why you might lose some viewership in the future. 00:02:50.074 --> 00:02:52.714 Fair enough, too. Like, all kinds of feedback are appreciated. 00:02:52.894 --> 00:02:55.414 So we're not roasting Matt here. Just wanted to share, like, 00:02:55.474 --> 00:02:56.894 we get both ends of this, right? 00:02:57.154 --> 00:03:01.674 So today we're going to represent the questions that have come in about agents and whatnot. 00:03:02.194 --> 00:03:07.594 But I think it's fair to say we want to check the temperature on this just in 00:03:07.594 --> 00:03:08.354 general with the audience. 00:03:08.894 --> 00:03:11.474 So send us a boost or go to linuxunplugged.com slash contact. 00:03:12.334 --> 00:03:17.614 And I do want to also say I think sometimes people, because they are so polarized 00:03:17.614 --> 00:03:21.994 about a particular topic, don't recognize that we do intentionally try to space this out. 00:03:22.354 --> 00:03:26.154 So you consider that AI has been the number one topic for the last three years 00:03:26.154 --> 00:03:31.094 in just about every economic story, every employment story, every tech story. 00:03:32.631 --> 00:03:36.431 And so we have worked very intentionally to try to space these topics out. 00:03:37.391 --> 00:03:41.251 Last week, we talked about Ubuntu and their grub plans. The week before that, 00:03:41.431 --> 00:03:43.131 we talked about ersatz TV wrapping up. 00:03:43.291 --> 00:03:45.971 And a little bit before that, I talked about my Keeper calendar program. 00:03:46.611 --> 00:03:50.871 So we try to space it out. We try to have episodes that don't just hit AI every 00:03:50.871 --> 00:03:53.551 single week, which means, you know, we're digging whole cloth and building that 00:03:53.551 --> 00:03:55.471 stuff for you, which is part of the value we try to bring. 00:03:56.551 --> 00:04:00.291 But it doesn't just accidentally happen that we go two or three weeks not talking about AI. 00:04:00.471 --> 00:04:04.011 And then one week we talk about AI. It's not just, that's just not like the 00:04:04.011 --> 00:04:07.291 topics fall into the show by accident. It's very intentional design on our part. 00:04:07.491 --> 00:04:10.071 And if we weren't doing that, it would look a lot different. 00:04:10.791 --> 00:04:11.111 Right. 00:04:11.211 --> 00:04:13.831 Just if that filter was not in place and all you were doing was going based 00:04:13.831 --> 00:04:16.991 on like what's hot in the world and what news stories are out there and what's changing. 00:04:17.191 --> 00:04:19.791 And what would get us advertising very easily. 00:04:19.991 --> 00:04:20.151 Right. 00:04:20.631 --> 00:04:23.251 Like we could, like we could just lean into it and lose some audience and make 00:04:23.251 --> 00:04:25.031 some advertising money there, but we're not doing that. 00:04:25.291 --> 00:04:27.351 And what we're trying to do is when we talk about it, we're trying to talk about 00:04:27.351 --> 00:04:31.051 in practical ways that are here today that are the open source angle, 00:04:31.051 --> 00:04:36.631 angle because that's what we cover and really impact linux users and we try 00:04:36.631 --> 00:04:39.531 to bring something even for the skeptics even if it's an episode that's about 00:04:39.531 --> 00:04:41.731 ai you're not an ai fan we try to bring something for everyone. 00:04:41.731 --> 00:04:46.651 I would argue also solving actual problems that we have either in our infrastructure 00:04:46.651 --> 00:04:49.891 or like some of the reverse engineering that we did of that diesel heater is 00:04:49.891 --> 00:04:54.651 a good example of using a new tool to accomplish something that we've been thinking 00:04:54.651 --> 00:04:59.691 about for what a year two years something like that so attaching it to a real world, 00:05:00.291 --> 00:05:05.871 use case and problem set i think hopefully it describes how we're finding actual 00:05:05.871 --> 00:05:08.451 uses for it not just burning up a bunch of credits. 00:05:09.875 --> 00:05:13.415 Yeah, we're doing that, too. We'll talk about that. But it is a balance, right? 00:05:13.575 --> 00:05:16.595 Because we don't want to lean too heavily into it. We want to make a show that's 00:05:16.595 --> 00:05:19.495 for as many people as possible. After all, it's a show we're making for you. 00:05:20.555 --> 00:05:25.475 So we do want to get this balance right. And it's something we want to hear from you about. 00:05:25.955 --> 00:05:30.515 We think that the terrain is still being discovered, right? The map still has a lot of fog on it. 00:05:30.915 --> 00:05:34.915 And there's a lot of mixed information out there, good and bad ideas and takes. 00:05:35.455 --> 00:05:39.355 And a lot of interesting technology that is really growing this year, 00:05:39.495 --> 00:05:41.275 2026, if we're open source. 00:05:41.695 --> 00:05:45.695 And this week was another significant step for open source this week. 00:05:46.115 --> 00:05:48.815 We had another one about three months ago. This is another one. 00:05:48.915 --> 00:05:52.075 And these keep happening specifically in the open source domain. 00:05:52.355 --> 00:05:54.055 So we're trying to balance all this out. 00:05:54.375 --> 00:05:59.415 Let us know what you think with a boost or, you know, go to the unplugged.com 00:05:59.415 --> 00:06:01.875 Linux. What is it? LinuxActionShow.com? 00:06:02.735 --> 00:06:03.315 Yeah, that's right. 00:06:04.075 --> 00:06:07.355 We got the Unplugged, you know, we got the contact page over there. You can figure it out. 00:06:07.555 --> 00:06:08.995 LinuxUnplugged.com slash contact. 00:06:09.375 --> 00:06:10.335 What? No. 00:06:10.895 --> 00:06:11.415 Never heard of it. 00:06:11.475 --> 00:06:14.715 Right? Unplugged? What kind of show? What, is that a show about radio? 00:06:14.935 --> 00:06:15.535 Yes, it is. 00:06:15.595 --> 00:06:16.355 Oh, okay. All right. 00:06:16.475 --> 00:06:17.615 But internet radio. 00:06:20.755 --> 00:06:23.315 All right, so let's talk about the good, the bad, the ugly here. 00:06:23.515 --> 00:06:28.295 We've got some common questions into the show, and we're going to go through 00:06:28.295 --> 00:06:30.655 some of these and then talk about our setups and talk about some of the big 00:06:30.655 --> 00:06:31.955 stuff landing for open source. 00:06:32.655 --> 00:06:37.255 And it's not all just one particular flavor. So I think let's start with probably 00:06:37.255 --> 00:06:41.495 the one that everybody's talking about right now, OpenClaw, which is a project 00:06:41.495 --> 00:06:44.375 that's getting a lot of the attention for something you can run locally and 00:06:44.375 --> 00:06:46.755 it can use local LMs or cloud LMs. 00:06:47.547 --> 00:06:50.487 And do we really need to get much into OpenClaw? We've had a couple of people 00:06:50.487 --> 00:06:52.927 ask, but I don't really feel like we need to spend a lot of time on it. 00:06:53.047 --> 00:06:55.347 It's a node-based agent's gateway stack. 00:06:55.547 --> 00:06:57.847 Okay, let's talk about what's a gateway, Wes. Maybe that's what we could explain. 00:06:58.027 --> 00:07:00.107 Well, yeah, I mean, it kind of depends on how deep you want to get, 00:07:00.287 --> 00:07:04.847 but there's a lot of different versions now of what people are calling an agent harness. 00:07:05.847 --> 00:07:09.007 And on one hand, you have just sort of the basic model, like the first version 00:07:09.007 --> 00:07:13.307 of ChatGPT in a browser tab that you're sort of typing into and interacting with. 00:07:13.447 --> 00:07:16.067 And at one point, maybe you're copying code in, and it writes code, 00:07:16.147 --> 00:07:17.167 and you copy it out or whatever. 00:07:17.547 --> 00:07:20.667 But then we switched to this version where it's like, it's living with us in 00:07:20.667 --> 00:07:22.687 our projects, in our editors, in the terminal. 00:07:23.167 --> 00:07:25.627 And it probably has something like tools, right? 00:07:26.127 --> 00:07:30.347 And it has, so that gives it like sort of mechanisms that it can affect change, 00:07:30.547 --> 00:07:31.847 edit files, run scripts. 00:07:32.447 --> 00:07:36.067 And, but beyond that sort of core set, then you have like a lot of different 00:07:36.067 --> 00:07:38.567 other features sort of like, how is the context assembled? 00:07:38.707 --> 00:07:41.167 Well, how does the memory system work? Is there a memory system? 00:07:41.327 --> 00:07:42.307 How do you chat with it? 00:07:42.467 --> 00:07:44.867 Yeah. And then you have the inputs and the outputs, which is like, 00:07:44.987 --> 00:07:47.267 what is the control interface and control surface? It's like, 00:07:47.387 --> 00:07:49.547 how do you trigger it? Is it autonomously triggered? 00:07:49.787 --> 00:07:52.847 Does it have mechanisms to like ping you anywhere you're at? 00:07:52.947 --> 00:07:55.047 Or it just presents on like an interface on your screen? 00:07:55.227 --> 00:07:59.887 And so like on one version, you have sort of like an open code or a cloud code 00:07:59.887 --> 00:08:02.407 or codex sort of thing where, I mean, they can do more than this. 00:08:02.487 --> 00:08:05.967 But the primary thing you first see is just like a TUI interface for you to 00:08:05.967 --> 00:08:09.807 sort of be human now with better helpers embedded right there. 00:08:10.307 --> 00:08:14.427 And OpenClaw you get is a very different experience where like you just sort 00:08:14.427 --> 00:08:18.007 of are presented with a telegram chat window into this bot who lives in its 00:08:18.007 --> 00:08:20.847 own entire other sort of universe of its own. 00:08:20.847 --> 00:08:24.607 A persistent running backend called the gateway that is connected to the LM 00:08:24.607 --> 00:08:28.487 that you chose and can run some of the tools like... 00:08:29.322 --> 00:08:35.082 Could be basic bash commands could be uh other things like a model context protocol 00:08:35.082 --> 00:08:38.202 and things like that but the gateway just sort of does that via like your commands 00:08:38.202 --> 00:08:40.562 in telegram chat or whatsapp or slack or whatever it might be. 00:08:40.562 --> 00:08:43.422 Yeah so it sort of um is the organizer right it's 00:08:43.422 --> 00:08:47.302 so it monitors like the telegram api or the slack api and gets incoming web 00:08:47.302 --> 00:08:50.262 hooks or whatever and then from there it says like oh right let's trigger the 00:08:50.262 --> 00:08:54.222 lm assemble its context give it all the stuff that it needs and then it also 00:08:54.222 --> 00:08:57.722 handles when the lm comes back with like i want to run a tool the gateway is 00:08:57.722 --> 00:09:01.122 actually what goes and like executes the tool call calls the mcp server. 00:09:01.122 --> 00:09:03.862 And i want to make kind of a common confusion clear that we've 00:09:03.862 --> 00:09:07.762 seen come into the show so say you're using olama with 00:09:07.762 --> 00:09:11.142 an open source lm or you're using chat gpt54 00:09:11.142 --> 00:09:14.022 for the back end of your open 00:09:14.022 --> 00:09:19.262 claw agent and maybe you're also using open code they're the same thing you're 00:09:19.262 --> 00:09:22.642 not going to really get dramatically different results because one's open claw 00:09:22.642 --> 00:09:26.122 and one's open code other than some of this harness that wes is talking about 00:09:26.122 --> 00:09:29.802 like the memories or the skills and those things kind of make it different and 00:09:29.802 --> 00:09:32.062 they give the agent the ability to, 00:09:32.953 --> 00:09:37.073 to remember mistakes, to remember that when I say custodian, 00:09:37.253 --> 00:09:40.133 I'm referring to 172.16.0.10. 00:09:40.453 --> 00:09:44.713 And so I don't have to write 172.16.0.10. I just say SSH to custodian. 00:09:44.993 --> 00:09:49.073 And memory and, of course, DNS help with things like that, where maybe if you 00:09:49.073 --> 00:09:54.273 just went to a fresh open code session or a fresh GPT session and just said 00:09:54.273 --> 00:09:56.593 ping custodian, it'd have no effing idea what you're talking about. 00:09:56.773 --> 00:09:59.773 And that's where it's like they kind of all exist in a design sort 00:09:59.773 --> 00:10:02.613 of possible design space and they have a lot of shared components and 00:10:02.613 --> 00:10:05.833 then some of them are just optimized for different experiences like open 00:10:05.833 --> 00:10:08.553 cloud really started as this sort of personal assistant that could manage your 00:10:08.553 --> 00:10:11.593 calendar and email and interface with you and like you know help you you 00:10:11.593 --> 00:10:15.273 know chat with you as in your telegram instance and maybe something more like 00:10:15.273 --> 00:10:18.273 open code is you know they have customized prompts focused on coding and they 00:10:18.273 --> 00:10:22.353 have a different style of sub-agent implementation it's focused on orchestrating 00:10:22.353 --> 00:10:25.593 multiple agents working specifically on code i mean it can do more than that 00:10:25.593 --> 00:10:29.973 right but you can see how sort of the defaults and the shape of the interface 00:10:29.973 --> 00:10:31.533 drive what they're primarily used for. 00:10:32.298 --> 00:10:32.618 Yeah. 00:10:32.798 --> 00:10:35.138 So like all of them, right? Because it's all just an LM under the hood. 00:10:35.298 --> 00:10:38.298 They can all write scripts. They can all run tool calls. It's just kind of what 00:10:38.298 --> 00:10:39.818 you put on top to let them do. 00:10:39.818 --> 00:10:45.618 So in my case, I'm using OpenClaw, and I have five agents running through OpenClaw 00:10:45.618 --> 00:10:49.058 that have domain-specific focuses. And we can talk more about that in a moment. 00:10:50.098 --> 00:10:52.738 So because one of the questions we also get into the show is, 00:10:52.838 --> 00:10:54.198 what the hell is even the use case? 00:10:54.358 --> 00:10:57.478 What are people even using this for? I don't really get it. Like I could just 00:10:57.478 --> 00:11:01.318 write a cron job with a Python script or a Bash script and do most of what you're 00:11:01.318 --> 00:11:04.938 doing. Or I could just use ClogCode or OpenCode. I can do most of these things. 00:11:05.038 --> 00:11:07.198 Like what the hell are people actually using this for? I don't get it. 00:11:07.198 --> 00:11:12.018 And so uh i think i wanted to start with brent because brent so far i know you've 00:11:12.018 --> 00:11:14.778 been busy but i think partially too you're kind of waiting to see where this 00:11:14.778 --> 00:11:18.218 goes because it is really early days he's watching. 00:11:18.218 --> 00:11:21.738 Us uh frequently post embarrassing you know things that our agents messed. 00:11:21.738 --> 00:11:24.118 Up yeah and so i thought i'd give you a chance because you probably represent 00:11:24.118 --> 00:11:25.598 where a lot of the audience is at on. 00:11:25.598 --> 00:11:26.078 This and just. 00:11:26.078 --> 00:11:29.918 Kind of talk about like where you feel it is on your adoption curve. 00:11:29.918 --> 00:11:32.518 Yeah i'm typically a little slower on the 00:11:32.518 --> 00:11:35.178 adoption curve than you boys which is beautiful because i get to 00:11:35.178 --> 00:11:38.038 watch you intimately screw things up and then 00:11:38.038 --> 00:11:40.738 learn from your mistakes so that's lovely but we 00:11:40.738 --> 00:11:45.538 do it for the show right boys but my hesitation always around new technologies 00:11:45.538 --> 00:11:51.738 tends to be you know of course privacy but also security because i might not 00:11:51.738 --> 00:11:57.098 have the same confidence as either of you to either not make the mistakes that 00:11:57.098 --> 00:12:00.158 i will regret later or to recover from them gracefully. 00:12:00.398 --> 00:12:05.938 So I like to wait just a little longer to see, let's say like a project like 00:12:05.938 --> 00:12:10.478 Open Claw reach more maturity than adopting it, you know, on week one, 00:12:10.678 --> 00:12:13.418 as so many people have done throughout the internet. 00:12:14.158 --> 00:12:18.338 So I would say that's probably pretty accurate that some of our audience fall 00:12:18.338 --> 00:12:22.338 into the category that I sit in, or maybe wait even longer. 00:12:22.878 --> 00:12:27.198 And I would say that's not a bad thing. That's okay. It means you're falling 00:12:27.198 --> 00:12:33.198 a little bit behind because the tools are moving so fast these days like every other day it seems. 00:12:34.264 --> 00:12:41.044 I just add to my list of things to learn. But that said, it's an evolution that 00:12:41.044 --> 00:12:44.224 you still need to keep up on, in my opinion. 00:12:44.464 --> 00:12:48.744 And I didn't have this opinion several months ago, because I was still kind 00:12:48.744 --> 00:12:50.404 of pausing and waiting to see. 00:12:50.624 --> 00:12:55.784 But having put my own pause on some of these tools, I got to say, 00:12:55.904 --> 00:12:59.804 it's made me a better open source software user. 00:12:59.804 --> 00:13:05.284 And allowed me to accomplish a bunch of projects that I've had on my to-do list 00:13:05.284 --> 00:13:08.744 for years and do them at speeds that I never would have anticipated. 00:13:09.304 --> 00:13:17.044 So that part, even though I hesitated to start to adopt, it's incontrovertible 00:13:17.044 --> 00:13:20.504 to me now that it's useful if you point it at the right thing. 00:13:21.584 --> 00:13:25.244 Yeah, so I think your take is pretty spot on. 00:13:25.404 --> 00:13:31.024 I don't know if you agree, Wes, But I think like it is, it is breaking and moving fast. 00:13:31.224 --> 00:13:35.744 And if you're not comfortable going in there and using something like a codex 00:13:35.744 --> 00:13:40.024 or an open code to sometimes fix it, you're probably going to have a bad time. 00:13:40.584 --> 00:13:44.324 Yeah. And I think that's to where like the difference in model sort of matters, 00:13:44.424 --> 00:13:47.944 right? Like open code is I tend or cloud code. 00:13:48.084 --> 00:13:51.684 I tend to like, you know, you, you open it, you run it on your computer or you run it somewhere. 00:13:51.924 --> 00:13:54.184 It sort of has a, maybe you run it for a long time and it runs persistently 00:13:54.184 --> 00:13:55.604 for days and weeks or whatever. But yeah. 00:13:56.757 --> 00:13:59.637 Versus the gateway for open claw is a systemd service that 00:13:59.637 --> 00:14:03.197 runs on a server and so um just the 00:14:03.197 --> 00:14:06.197 models are very different and the introspection and the default of how much info 00:14:06.197 --> 00:14:08.937 and sort of the inner state that you're exposed to 00:14:08.937 --> 00:14:11.877 of the system is very different and then on top of 00:14:11.877 --> 00:14:14.637 that you know you're so there you're sort of pressing the bounds 00:14:14.637 --> 00:14:17.337 of like how little interface can you have and 00:14:17.337 --> 00:14:20.377 still have this thing manage productive work which is its own question but 00:14:20.377 --> 00:14:23.797 that sort of imposes a lot on the whole model and just 00:14:23.797 --> 00:14:27.437 the nature of the project yes like it just 00:14:27.437 --> 00:14:30.257 moves crazy crazy fast so fast that we've both now had 00:14:30.257 --> 00:14:33.217 to sort of fork the up the upstream nix 00:14:33.217 --> 00:14:36.677 code which wasn't updating fast enough to keep up with the proper upstream 00:14:36.677 --> 00:14:39.337 source code and it's just you know it's trying to 00:14:39.337 --> 00:14:42.237 do a lot there's a ton of features so i think 00:14:42.237 --> 00:14:45.257 we've maybe both been continuing to run it because we have been curious 00:14:45.257 --> 00:14:49.677 about i mean we run other things but just because it has been sort of the locus 00:14:49.677 --> 00:14:54.237 of a lot of the frontier but if you don't care about that aspect you can still 00:14:54.237 --> 00:14:58.637 even have this model style of approach and have much more stable things or things 00:14:58.637 --> 00:15:02.157 that are moving slower or you know aren't based on the node ecosystem but have 00:15:02.157 --> 00:15:04.577 like a go core or a rust core python i want. 00:15:04.577 --> 00:15:07.377 To i want to talk about a couple of those tools because because it's not i think 00:15:07.377 --> 00:15:11.037 you're touching on a good point is it's not going to be just open claw and some 00:15:11.037 --> 00:15:13.677 of them are going to be more stable more lts style. 00:15:13.677 --> 00:15:17.337 Yeah there's going to be enterprise versions and like debian style versions 00:15:17.337 --> 00:15:18.757 and you And I think I'll manner of them. 00:15:18.757 --> 00:15:24.717 Here's how I boil it down, is right now, it's not worth burning a lot of money 00:15:24.717 --> 00:15:26.297 on tokens to run OpenClaw. 00:15:26.717 --> 00:15:31.037 I just, I genuinely don't think it is, because you'll spend a lot of those tokens fixing it. 00:15:31.717 --> 00:15:35.857 And if you have a plan where you have access to a lot of AI tokens, 00:15:35.857 --> 00:15:41.117 or you have local AI hardware, where it costs you nothing, then go for it. 00:15:41.137 --> 00:15:44.497 Because if you do, you will learn so much. 00:15:44.637 --> 00:15:47.077 It is an incredible learning experience, but also... 00:15:47.961 --> 00:15:51.561 You do become a better operator. Like I know where their deficiencies are now. 00:15:51.661 --> 00:15:54.701 So I prompt better and better and better and I get better and better results 00:15:54.701 --> 00:15:57.061 and I have them doing more and more things, which we'll get into a couple of 00:15:57.061 --> 00:15:58.661 those because I do want to talk about use cases. 00:15:58.841 --> 00:16:03.181 But that's my hot take is I don't think it's worth spending a bunch of money 00:16:03.181 --> 00:16:08.921 on Open Router or going and getting some $100 a month plan to run Open Claw right now. 00:16:09.061 --> 00:16:12.901 I think especially if you're going to try to use it as like the only thing you're 00:16:12.901 --> 00:16:16.541 doing, like if you're trying to push through building the entire thing through that. 00:16:16.701 --> 00:16:17.701 It's going to be a bad time. 00:16:17.701 --> 00:16:18.561 It's an interesting experiment. 00:16:18.721 --> 00:16:19.421 Right? It's just not there yet. 00:16:19.521 --> 00:16:23.841 But yes, versus like if you're kind of having it orchestrate battle-tested open 00:16:23.841 --> 00:16:26.621 source services or scripts that you have open code right for you. 00:16:26.741 --> 00:16:27.941 Like that is a very different experience. 00:16:28.021 --> 00:16:32.461 That's it. So let's get into that's it because that I think gets us to our use cases and our setups. 00:16:33.021 --> 00:16:37.821 If you're comfortable using a superior model and a superior tool to build the 00:16:37.821 --> 00:16:42.261 infrastructure around these things and then have them operate it with guardrails, 00:16:42.401 --> 00:16:44.521 you're going to get great results. 00:16:45.021 --> 00:16:48.221 If what I just said doesn't make sense to you, it's going to be a hard time. 00:16:48.321 --> 00:16:49.421 And that's where we're at. 00:16:49.641 --> 00:16:54.461 And I just, if I, it is a really tough thing. It's like, it's like when it took 00:16:54.461 --> 00:16:56.461 you days to get Linux running. 00:16:56.601 --> 00:17:01.601 I mean, this has a lot of this same energy I've been putting into this where like, I don't have time. 00:17:02.101 --> 00:17:05.741 But yet, just like I didn't have time 20 years ago when I was learning Gentoo, 00:17:05.981 --> 00:17:09.861 I somehow did it, right? Like I just, the drive is there. 00:17:10.021 --> 00:17:13.521 Yes, because you can see there's potential. There's a lot of fun in it and a 00:17:13.521 --> 00:17:16.501 lot of frustration. And so I think it's worth knowing that and like, 00:17:16.681 --> 00:17:18.661 it's almost like a pet, you know, it's like a pet. 00:17:18.841 --> 00:17:22.921 It's a big side project, but don't go in necessarily expecting to convert it 00:17:22.921 --> 00:17:25.681 into a production thing that's going to be rock solid that you forget about. 00:17:25.801 --> 00:17:29.241 I think pet is one, you know, like the Tamagotchi is one thing and that's fine 00:17:29.241 --> 00:17:32.341 if that's what you want. And you've got a way to do that economically. 00:17:32.821 --> 00:17:38.861 I kind of look at it more as an intern or a really kind of basic producer. 00:17:39.021 --> 00:17:40.901 And I, I've gotten it pretty close. 00:17:41.887 --> 00:17:47.027 Um, so in my case, I have a doing a lot of analysis for the show. 00:17:47.647 --> 00:17:50.587 Every episode of pulls down the transcripts, it does sentiment analysis. 00:17:50.587 --> 00:17:53.247 It does, it keeps track of everything we've talked about. 00:17:53.467 --> 00:17:58.327 I can also pull down emails and like that email we got earlier that I read in the show. 00:17:58.527 --> 00:18:01.167 I, it matched them the sentiment analysis of the things we actually talked about 00:18:01.167 --> 00:18:03.587 in a surface that email, because, Hey, this is actually, you guys, 00:18:03.707 --> 00:18:05.347 you know, this email is kind of on point. 00:18:05.507 --> 00:18:08.727 Like, I don't think Matt likes to hear this, but it was the agent that surfaces 00:18:08.727 --> 00:18:11.867 email to me that said, Hey, he's complaining and you have been talking about it in this episode. 00:18:11.887 --> 00:18:14.747 Episode this episode episode you mentioned this and it's like he might have 00:18:14.747 --> 00:18:17.087 a point here it's worth reading this and that's why his email made into the 00:18:17.087 --> 00:18:22.087 show today and that's something i would task a producer to do if i had uh you know i don't know a. 00:18:22.087 --> 00:18:22.787 Budget for producers. 00:18:22.787 --> 00:18:25.207 What would it be in washington state a hundred thousand dollars i don't know 00:18:25.207 --> 00:18:28.767 it's crazy here so it's just not going to happen i'm not not even paying myself 00:18:28.767 --> 00:18:32.867 at the moment so it's not that's not going to happen it also generates news 00:18:32.867 --> 00:18:38.567 briefings for me every day both in text with sources but also in audio that 00:18:38.567 --> 00:18:41.027 uses my fresh RSS server feeds. 00:18:41.287 --> 00:18:45.447 So the feeds I've curated for the last decade that I have in my local fresh 00:18:45.447 --> 00:18:49.967 RSS server, every morning it goes, it does analysis on that, 00:18:50.127 --> 00:18:55.567 and it generates me a seven to 15 minute long report of the stories that are relevant to our shows. 00:18:55.667 --> 00:18:57.967 And then it marks them red in my fresh RSS feed. 00:18:58.067 --> 00:19:00.887 So then later when I go to read my news stories, because I'm always trying to 00:19:00.887 --> 00:19:03.927 stay on top for the shows, the ones that have been in my audio briefing that 00:19:03.927 --> 00:19:06.187 I listened to on the drive to the studio are marked red for me now. 00:19:06.927 --> 00:19:10.367 Little things like that. Or when a sponsor emails me and I'm trying to get a 00:19:10.367 --> 00:19:13.407 sponsor going, I have that surface, that alert to me using GWS. 00:19:13.567 --> 00:19:18.587 So GWS CLI lets me check in on these things without going crazy with permissions in my inbox. 00:19:20.243 --> 00:19:23.563 And these are little tasks that I have it do, but I think the stuff that I would 00:19:23.563 --> 00:19:28.203 find if you're asking what's the use case, it's for giving you an easy interface 00:19:28.203 --> 00:19:30.043 to manage all the crap you've set up. 00:19:31.143 --> 00:19:33.583 There's a lot of, you know, Home Assistant's a great example. 00:19:33.763 --> 00:19:39.143 It really can be a great accelerator to your setup there. So at home, I have the Frigate DVR. 00:19:39.363 --> 00:19:43.363 The Frigate DVR, when it notices an event or a face via MQTT, 00:19:43.743 --> 00:19:44.923 sends an alert to Home Assistant. 00:19:45.723 --> 00:19:48.583 Home Assistant has an automation that wakes up my agent Uhura. 00:19:50.183 --> 00:19:55.263 Uhura analyzes the image from Frigate and then sends me a report with the faces, 00:19:55.263 --> 00:19:58.003 the people identified, and a description of the situation. 00:19:58.203 --> 00:20:02.003 And it's estimate if it's the severity level of the situation. 00:20:02.183 --> 00:20:06.123 I just have a telegram chat. And whenever I'm away, because home assistant automatically 00:20:06.123 --> 00:20:10.083 activates this detection system, home assistant's doing the lift here. 00:20:10.403 --> 00:20:13.823 But my open claw agent is doing the final analysis and report. 00:20:14.383 --> 00:20:18.003 The wiring, the hard stuff, is the infrastructure with Frigate, 00:20:18.203 --> 00:20:19.883 MQTT, and Home Assistant. 00:20:20.323 --> 00:20:25.423 Then OpenClaw just sits on top of that as a layer to give me access to all these APIs and features. 00:20:25.623 --> 00:20:29.863 So when I wanted to start getting these alerts, I tasked the OpenClaw agent 00:20:29.863 --> 00:20:33.763 to finish up the YAML in Home Assistant to add the sensor for face detection, 00:20:33.943 --> 00:20:37.563 and then to expand its reports that come back to me with image analysis. 00:20:38.243 --> 00:20:41.043 But I didn't have it build the entire system from WholeCloth. 00:20:41.143 --> 00:20:45.003 I had it do the last 10%. And it's working great. I've been sending West the results all week. 00:20:45.123 --> 00:20:45.923 It's been a lot of fun. 00:20:46.043 --> 00:20:49.643 It's a lot of fun to have it analyze and learn the family and pass silent LM 00:20:49.643 --> 00:20:51.763 judgment on like a room being cluttered. 00:20:52.003 --> 00:20:54.983 It likes to think you're, I mean, does it not get that you're in an RV? No, it does. 00:20:55.123 --> 00:20:59.083 That's what's also interesting is it does recognize, it's figured out that it's an RV. 00:20:59.263 --> 00:20:59.343 Oh. 00:20:59.743 --> 00:20:59.963 Yeah. 00:21:00.363 --> 00:21:00.903 That's great. 00:21:01.063 --> 00:21:04.983 Yeah. So that's probably one of a million use cases. But again, I wouldn't do it. 00:21:05.738 --> 00:21:09.258 Unless you could find a way to economically get access to the AI, 00:21:09.338 --> 00:21:10.598 which we're going to talk about more about in a moment. 00:21:10.678 --> 00:21:11.698 But I'm curious if you want to share 00:21:11.698 --> 00:21:14.118 any of the things that you're kind of routinely using these things for. 00:21:14.318 --> 00:21:17.838 Well, yeah, you're just making me think like, well, I mean, any anything you 00:21:17.838 --> 00:21:21.458 do want to orchestrate that you don't want to have to go sit at a computer to do. 00:21:21.618 --> 00:21:25.458 So it could be stuff that's routine is like monitoring your systems or reporting 00:21:25.458 --> 00:21:28.778 on, you know, you don't need it to go collect the metrics necessarily. 00:21:28.778 --> 00:21:31.358 But you can have something that's collecting metrics and it can look at it and 00:21:31.358 --> 00:21:33.918 give you an assessment about things or look at anomalies. 00:21:34.098 --> 00:21:36.558 Or you could have it, if it has permissions and you're willing to do this, 00:21:36.658 --> 00:21:40.418 go run updates on some of your servers and report back on how that goes. 00:21:40.818 --> 00:21:44.598 And it also is just useful to have it flip the script. And I think many of us 00:21:44.598 --> 00:21:50.358 recognize that if you leverage web search with LLMs, they can be very useful researchers. 00:21:50.538 --> 00:21:54.338 If you just rely on what's in their pre-trained data set, then you can get much 00:21:54.338 --> 00:21:57.278 different results. But if it has access to fresh, good, you know, 00:21:57.358 --> 00:22:00.078 reg style sort of data and information, it can be quite useful. 00:22:00.358 --> 00:22:03.558 And so instead of having to go sit at your computer and pull up some web interface 00:22:03.558 --> 00:22:08.978 to go do all that, you know, I can just like I spent a while making sure that 00:22:08.978 --> 00:22:12.718 my bot had really good access to whisper transcription so that, 00:22:12.798 --> 00:22:13.798 you know, I could just send it. 00:22:13.798 --> 00:22:16.858 I could record voice in Telegram and shoot that over and then it could go spin 00:22:16.858 --> 00:22:20.538 up and have sub agents that go use search engines to go pull a bunch of stuff 00:22:20.538 --> 00:22:25.118 and then sort of recursively analyze that and then at the end produce a markdown report. 00:22:25.658 --> 00:22:28.298 And then, you know, kind of like you've been doing, I can have that then spit 00:22:28.298 --> 00:22:32.958 back out into sort of like a podcast form with a pocket TTS voice. 00:22:33.158 --> 00:22:34.938 Which is open source, runs on your CPU. 00:22:35.218 --> 00:22:38.258 And that's all stuff I could do myself or sit at a thing and task an element to do. 00:22:39.113 --> 00:22:42.173 And I might not even bother if it's like, oh, I got to go interrupt what I'm doing. 00:22:42.213 --> 00:22:46.913 But if all I have to do is fire off a quick voice request, but you can also 00:22:46.913 --> 00:22:49.073 burn a lot of time if you're trying to set that up fresh every time, 00:22:49.093 --> 00:22:53.113 or if you haven't properly sort of ossified it and tested it and made sure that it works reliably. 00:22:53.653 --> 00:22:56.593 And this is where you're talking about being a good operator is understanding 00:22:56.593 --> 00:23:01.733 because by default, the LLM that's sort of operating within this harness doesn't 00:23:01.733 --> 00:23:04.713 necessarily, and this varies per harness, but OpenClaw's not great at this, 00:23:05.033 --> 00:23:07.553 doesn't necessarily have a very good understanding of how it works. 00:23:07.553 --> 00:23:10.733 Like how, and it's a moan model of what it, how it functions. 00:23:10.913 --> 00:23:13.953 And so if you don't have that model and aren't sort of infer, 00:23:14.053 --> 00:23:16.593 you know, passing some of that info or haven't spent time doing that, 00:23:16.773 --> 00:23:18.033 then it can get really confused. 00:23:18.113 --> 00:23:20.353 And you have situations where it works great a couple of times. 00:23:20.513 --> 00:23:24.013 And then a week later, you're like, oh, it has no idea that it was ever able to even do that. 00:23:24.193 --> 00:23:27.973 Yeah. I think one of the benefits of playing with this now is I've learned how 00:23:27.973 --> 00:23:30.853 both capable and dumb LLMs are, right? 00:23:30.913 --> 00:23:33.433 They're starting with a fresh world every single time. 00:23:33.693 --> 00:23:36.793 And when you experience them without an agent harness around them 00:23:36.793 --> 00:23:39.653 that's always your primary interface and so once 00:23:39.653 --> 00:23:42.313 once you have an agent harness and it starts to become a little 00:23:42.313 --> 00:23:45.453 more personalized i'll give you one more use case example that happened on 00:23:45.453 --> 00:23:48.113 saturday yesterday uh the boy and 00:23:48.113 --> 00:23:51.133 i were sitting there and we went to go to tunar which i talked about 00:23:51.133 --> 00:23:55.433 recently to watch some streaming tv and we went to the regular show channel 00:23:55.433 --> 00:23:59.153 and it's one of those dad moments you hate where you hit the button and you 00:23:59.153 --> 00:24:02.953 get stream failed all right oh i've just been telling him i just set up a regular 00:24:02.953 --> 00:24:08.493 show channel stream failed oh i'm sorry dylan i don't know why this isn't oh wait a minute. 00:24:09.853 --> 00:24:12.613 Lore go see why tunar says there's uh no 00:24:12.613 --> 00:24:15.813 episodes for a regular show i've got the entire series there's eight seasons 00:24:15.813 --> 00:24:21.633 there's plenty of episodes fix it right and what lore identified was is that 00:24:21.633 --> 00:24:25.993 ersatz apparently didn't care but tunar does care how the files are organized 00:24:25.993 --> 00:24:29.853 on disk even though it's getting the information from jellyfin it's still tunar 00:24:29.853 --> 00:24:31.393 is still sensitive to how the files, 00:24:32.073 --> 00:24:36.673 are organized and if you don't have them in individual season folders it doesn't see any episodes, 00:24:37.708 --> 00:24:41.408 So there's eight seasons and they're all, it's all, you know, season one. 00:24:42.108 --> 00:24:45.108 And there's some of these are, there's like 36 episodes a season and they're 00:24:45.108 --> 00:24:46.708 all just in the root directory. 00:24:47.748 --> 00:24:50.128 So Laura identified that and he said, well, here's the problem. 00:24:50.688 --> 00:24:54.368 Do you want me to SSH into custodian, create the season folders and move them 00:24:54.368 --> 00:24:59.168 all for you and organize them and then ping the TUNAR API and have it rescan? Yeah, do it. 00:24:59.768 --> 00:25:03.248 And then five minutes later, we go back to the channel, hit play and it works. 00:25:04.048 --> 00:25:06.908 And I never had to get off the couch. I never had to stop interacting with my 00:25:06.908 --> 00:25:10.208 son. I just sat there and tasked the machine to go fix it for me because that 00:25:10.208 --> 00:25:13.888 harness tells it all the information it needs to know to go do those things. 00:25:14.108 --> 00:25:15.988 And, of course, it can execute tools. 00:25:16.148 --> 00:25:20.808 It's so useful as an organizer and sort of default sysadmin interface for a 00:25:20.808 --> 00:25:22.968 lot of stuff like that, where it's like, I know what needs to be done. 00:25:22.968 --> 00:25:25.088 I just don't really want to do the tedium of doing it. 00:25:25.148 --> 00:25:29.848 And it can present me like a plan that I can audit and approve and then tell it to go do that. 00:25:30.068 --> 00:25:35.208 So I think the reason why it kind of shifts this week is Gemma 4 has landed. 00:25:35.548 --> 00:25:40.988 This is from Google DeepMind. It's their open source Apache 2 licensed model. 00:25:41.308 --> 00:25:46.468 And so it's based on their bigger commercial Gemini model, but this is their open source play. 00:25:46.568 --> 00:25:50.928 So they have a high end commercial play and now they're trying to have a high end open source play. 00:25:51.208 --> 00:25:55.348 And it does seem to have pretty advanced reasoning. It's been trained with agentic 00:25:55.348 --> 00:25:58.568 workflows and people have been running it on their iPhones. 00:25:59.308 --> 00:26:03.588 And it seems to be performing more than 20 X models, its own size. 00:26:04.468 --> 00:26:08.048 It's, you know, competitor models. And you can try it out right now in LM studio. 00:26:08.068 --> 00:26:13.288 I downloaded it on my 2016 era Nick station upstairs, which is it's got like 00:26:13.288 --> 00:26:16.528 an AMD Radeon, maybe from 2018 at best. 00:26:16.808 --> 00:26:20.428 I think I upgraded it once from like an old NVIDIA card to a Radeon just for compatibility. 00:26:22.416 --> 00:26:25.896 Maybe. And it's got a, you know, an i7 from 2016. 00:26:26.296 --> 00:26:29.196 And it's got 32 gigs of RAM. Oh, maybe 64. 00:26:29.536 --> 00:26:34.816 And I was able to load Gemma 4 on that 2016 era system and actually have it 00:26:34.816 --> 00:26:38.396 do accurate image analysis in, you know, under a couple of minutes. 00:26:38.596 --> 00:26:39.716 It's slow on that old system. 00:26:39.996 --> 00:26:46.036 But people on iPhones and even newer boxes are seeing incredible results with 00:26:46.036 --> 00:26:50.296 tool call capability, reasoning, and you can run the entire thing on your local machine. 00:26:50.576 --> 00:26:53.956 PJ, you were trying it out earlier on your machine. What are the specs for the 00:26:53.956 --> 00:26:54.856 computer that you were running it on? 00:26:55.256 --> 00:27:02.216 The machine I have is a, we'll see, AM4 5700, I think. 00:27:02.696 --> 00:27:10.136 The nicer one, the one with 3D cache. And it's got a 1070, GTX 1070. So another older GPU. 00:27:10.476 --> 00:27:11.956 Could you tell if it was using GPU acceleration? 00:27:12.476 --> 00:27:17.316 Yeah, it was. I loaded up NVTOP and I know that the Olama Web UI that i've got 00:27:17.316 --> 00:27:21.336 loaded up is it's all set up for gpu use and i just threw a picture of my dog 00:27:21.336 --> 00:27:24.716 put it in the chat and then it described my dog within just a few seconds. 00:27:24.716 --> 00:27:26.956 Oh really that fast yeah. 00:27:26.956 --> 00:27:31.856 I'm even using a the 12b model which is probably too big for the amount of eram 00:27:31.856 --> 00:27:35.616 i have but tends to work fine when i when i throw stuff like that at it. 00:27:35.616 --> 00:27:38.436 Yeah this is i think beginning to be at a 00:27:38.436 --> 00:27:41.276 model scale where you could play with his agentic stuff and not 00:27:41.276 --> 00:27:45.076 blow a bunch of money money on big tech cloud tokens then 00:27:45.076 --> 00:27:48.376 something else that landed this week which we 00:27:48.376 --> 00:27:51.436 haven't had a chance to try it because we want to give it a proper try if you 00:27:51.436 --> 00:27:55.476 guys are interested is mesh llm and 00:27:55.476 --> 00:28:00.416 it lets you pool your spare gpu capacity across your lan and then expose it 00:28:00.416 --> 00:28:06.216 as one open ai compatible api this is from block and it's a peer-to-peer system 00:28:06.216 --> 00:28:11.556 that lets anyone pool spare gpu compute and what it does is the system loads an LLM, 00:28:11.676 --> 00:28:14.496 and when it consumes the VRAM of the local host, 00:28:14.636 --> 00:28:17.596 it goes out to the mesh, and it distributes across the mesh, 00:28:17.736 --> 00:28:19.296 and then continues loading the model. 00:28:20.350 --> 00:28:21.710 I really want to try this. 00:28:22.830 --> 00:28:27.250 You know I have been saying this is coming, and I am so excited to try this 00:28:27.250 --> 00:28:30.730 because it really opens it up for folks like us that have spare hardware. 00:28:31.250 --> 00:28:34.850 They're not great, but if you really pull it all together, I don't know, 00:28:34.910 --> 00:28:38.870 or maybe even just a couple of cheaper VPSs, we'll see. We'll see how much it 00:28:38.870 --> 00:28:40.170 really needs a great GPU or whatever. 00:28:40.390 --> 00:28:45.330 But I'm really excited about Mesh LM, and it's open-sourced by Block, 00:28:45.510 --> 00:28:48.590 and I think it's going to be really great. They also have an agent harness they've open-sourced. 00:28:49.210 --> 00:28:53.010 So if this we've talked a lot about open claw uh loeb 00:28:53.010 --> 00:28:55.970 hub is another one i've played with i think it's out of china it is 00:28:55.970 --> 00:28:59.750 open source and it's a lot more user-friendly it's 00:28:59.750 --> 00:29:02.650 it is the chat interface itself it's also the skill 00:29:02.650 --> 00:29:07.030 store the mcp store eight different agent personality stores and it connects 00:29:07.030 --> 00:29:11.170 to every freaking kind of provider that i ones i've never even heard of i had 00:29:11.170 --> 00:29:14.630 no idea there were that many plus all the local stuff and all the free stuff 00:29:14.630 --> 00:29:19.830 it's crazy and it's just a built-in agent orchestration platform in one UI, it's called LoebHub. 00:29:20.350 --> 00:29:22.290 Goose is another one that's out there. 00:29:22.830 --> 00:29:26.670 And then there's the Hermes agent, which is being described as a self-improving 00:29:26.670 --> 00:29:28.970 AI agent built by newest research. 00:29:29.290 --> 00:29:32.450 And they say it has a built-in learning loop, which I've also created for my 00:29:32.450 --> 00:29:34.810 own. I call it the reflection loop, which I recommend. 00:29:35.070 --> 00:29:40.070 It's a nightly job that scours the JSON chat sessions and looks for mistakes 00:29:40.070 --> 00:29:43.330 and corrections that the agent made and then documents those in the memory. 00:29:43.470 --> 00:29:44.810 And they've built that in. 00:29:46.148 --> 00:29:49.268 So it self-improves, it nudges itself for persistent knowledge, 00:29:49.328 --> 00:29:53.068 it searches its own past conversations, and it builds a deepening model of who 00:29:53.068 --> 00:29:55.788 you are across sessions, and you can run on a $5 VPS. 00:29:56.548 --> 00:30:00.248 Love it. Yeah, there's so many cool options. I put two in here that are sort 00:30:00.248 --> 00:30:01.768 of on opposite ends of the spectrum. 00:30:01.948 --> 00:30:05.028 So one is called LibreFangs. This is like a Rust one. 00:30:05.268 --> 00:30:09.488 It's an agent operating system, a full platform for running autonomous agents, 00:30:09.688 --> 00:30:13.288 built from scratch in Rust, not a chatbot framework, not a Python wrapper. 00:30:13.288 --> 00:30:16.328 And as they put it, traditional agent frameworks wait for you to type something. 00:30:16.688 --> 00:30:19.888 This thing runs stuff that just is supposed to work for you 24-7. 00:30:20.388 --> 00:30:24.328 So it comes with a researcher, a collector, a predictor, a strategist, 00:30:24.408 --> 00:30:29.528 a trader, a Twitter personality, LinkedIn, browser, API tester, DevOps personalities. 00:30:29.828 --> 00:30:32.448 And so you can spawn all these things they can persistently run. 00:30:32.668 --> 00:30:36.448 You've seen similar things with Gastown. There's a bunch of it. 00:30:37.028 --> 00:30:41.108 But then on the other side is something called OpenHarness, which is a lightweight 00:30:41.108 --> 00:30:42.648 agent infrastructure, sort of 00:30:42.648 --> 00:30:46.568 tool use, skills, memory, multi-agent coordination, but that's kind of it. 00:30:46.888 --> 00:30:50.988 It can connect to other things too, but it's intentionally built to be an open 00:30:50.988 --> 00:30:53.668 source Python implementation designed for researchers, builders, 00:30:53.708 --> 00:30:57.368 and the community to help you understand how it works, experiment, extend. 00:30:57.728 --> 00:31:01.088 So rather than something that's like a product or a crazy fast open source moving 00:31:01.088 --> 00:31:03.688 thing, here's something maybe you could get more comfortable with and actually 00:31:03.688 --> 00:31:05.248 play with and like build understanding. 00:31:05.248 --> 00:31:10.308 Okay, nice finds I think a couple other trends I also I have like my biggest 00:31:10.308 --> 00:31:14.168 takeaway for both the haters and and the hypers that I want to get to but, 00:31:15.322 --> 00:31:19.962 I think we're about to see this massive enterprise shift towards agents. 00:31:20.242 --> 00:31:24.382 And Microsoft's leading here, I predict, here on the show. You're also going 00:31:24.382 --> 00:31:27.142 to see Red Hat make a big deal about this at Summit this year. 00:31:27.282 --> 00:31:29.862 You're already seeing it sort of in, like, job postings and stuff. 00:31:30.042 --> 00:31:34.062 You know, way more stuff like MCP tool mentioning and just, like, agent pipelines. 00:31:34.142 --> 00:31:38.522 That's what I'm—yeah, you and I—that's one of the things we watch for the show, and it's clear. 00:31:39.082 --> 00:31:42.762 So Microsoft has released the Agents Governance Toolkit. 00:31:44.162 --> 00:31:47.122 Really the takeaway here isn't this particular product 00:31:47.122 --> 00:31:49.982 it's that this is a thing that these companies are going to make 00:31:49.982 --> 00:31:56.122 now and it's their new open source project it's a toolkit that has identified 00:31:56.122 --> 00:31:59.802 all kinds of different types of risks and tried to create essentially a sandbox 00:31:59.802 --> 00:32:05.922 an os system around it that does governance monitoring communications monitoring 00:32:05.922 --> 00:32:07.642 all that kind of stuff around imagine right. 00:32:07.642 --> 00:32:10.982 Like when these things have access to various tools and various implementations 00:32:10.982 --> 00:32:15.202 of them and you're going to want to have some person over here in sales be able to do certain things. 00:32:15.202 --> 00:32:15.882 And another person. 00:32:15.882 --> 00:32:19.582 In IT do more things and you're going to need to want to tie it into all your 00:32:19.582 --> 00:32:21.602 existing ancient enterprise tools. 00:32:21.882 --> 00:32:26.722 It ties into your policy engine. It intercepts every agent action before execution. 00:32:27.062 --> 00:32:32.402 There you go. And also interestingly enough includes a mesh network component 00:32:32.402 --> 00:32:34.262 so you can have agent to agent communication. 00:32:34.962 --> 00:32:36.662 But presumably with policies. 00:32:36.922 --> 00:32:37.182 Yeah. 00:32:37.322 --> 00:32:37.822 That's interesting. 00:32:37.942 --> 00:32:42.002 Yeah, yeah. I think you're going to see other companies lean into this a lot. 00:32:42.342 --> 00:32:47.022 And then there are, I think, probably two big takeaways that we should probably talk about. 00:32:47.142 --> 00:32:52.222 But first, thank you to our members for making this possible. 00:32:52.362 --> 00:32:54.702 You can support the show at linuxunplugged.com slash membership. 00:32:54.702 --> 00:32:58.142 We just have to find networking right now. We could really use the support to keep us going. 00:32:59.162 --> 00:33:02.082 And if you like where we're going, where the direction we're going and how we 00:33:02.082 --> 00:33:04.862 go about it, your support's one of the best ways to keep it that way, 00:33:05.062 --> 00:33:07.862 linuxunplugged.com slash membership or jupyter.party. 00:33:08.302 --> 00:33:12.102 And a big thank you to our members for supporting the Unplugged program. 00:33:15.510 --> 00:33:20.090 Now we've seen a lot of open source projects recently describe how they want 00:33:20.090 --> 00:33:26.050 AI to be used in their projects. We've even seen some projects just outright try to block AI. 00:33:26.550 --> 00:33:29.830 But the real question is, is that a futile gesture? 00:33:30.030 --> 00:33:34.870 Or are they just going to have to find a way to play nice with these AI agents 00:33:34.870 --> 00:33:36.930 who are trying to make our worlds a better place? 00:33:38.090 --> 00:33:40.270 Yeah, I mean, there's a few different ways to think about that. 00:33:40.410 --> 00:33:43.550 And, you know, sort of the game theory or strategy side and the real, 00:33:43.590 --> 00:33:46.030 you know, sort of whatever happens in the market and where things go. 00:33:46.290 --> 00:33:50.470 But I think it's interesting that like the downsides and sort of like, 00:33:50.670 --> 00:33:54.170 who are you fighting and who are you trying to serve maybe changes. 00:33:54.170 --> 00:33:58.750 Like there's one version where like it's a bunch of scrapers sucking down the 00:33:58.750 --> 00:34:03.430 entire internet to train things and, you know, is making it hard for you to run your website. 00:34:03.510 --> 00:34:04.930 And like, we see that and that's a real problem. 00:34:05.050 --> 00:34:09.370 Yeah, for sure. Um, and the other version is like, I'm trying to compare some 00:34:09.370 --> 00:34:13.250 open source projects for the show, and I want my bot to go clone each of them 00:34:13.250 --> 00:34:17.210 so we can write a report about how to, you know, make me a table to think about 00:34:17.210 --> 00:34:19.370 the different ones and see which ones I want to actually try. 00:34:20.010 --> 00:34:26.030 And, you know, I want it to respect rules and rate limits and all those things, but it's also sort of, 00:34:27.036 --> 00:34:30.936 doing what I would do myself. And so having it blocked just sort of feels like 00:34:30.936 --> 00:34:34.336 it's limiting the stuff that I want to do, which is probably to go on and like 00:34:34.336 --> 00:34:37.216 highlight an open source project that I've tried and I'm excited about or maybe 00:34:37.216 --> 00:34:40.116 contribute to or, you know, engage with in some way. 00:34:40.336 --> 00:34:43.596 Yeah. And so I think what we're seeing, and this is going to be a tough one 00:34:43.596 --> 00:34:47.256 for the community to wrap their noodles around because it's a big transition 00:34:47.616 --> 00:34:49.956 because for as long as I've been on the internet, bots are bad. 00:34:50.276 --> 00:34:51.576 And, you know, maybe you, maybe 00:34:51.576 --> 00:34:54.496 you want your site indexed for Google or something, but that's about it. 00:34:54.616 --> 00:34:58.816 Otherwise bots are just traffic and putting load on your system that you don't 00:34:58.816 --> 00:35:04.856 want, but we're shifting from mass web scraping bots, which still exist and 00:35:04.856 --> 00:35:08.836 maybe are even worse than they've ever been in some regard to something that's a lot more personal, 00:35:09.136 --> 00:35:14.476 something that's an extension of the user's hand acting on the direct intent of the user. 00:35:14.756 --> 00:35:19.636 And so if you block that, you're blocking new users to open source software. 00:35:19.976 --> 00:35:24.196 You're, you're kind of saying, well, we want users, but we don't want the users 00:35:24.196 --> 00:35:25.856 that are asking the machine to do it for them. 00:35:26.730 --> 00:35:31.330 Which is kind of a moral call there. And I don't think we've really thought through that as much. 00:35:31.770 --> 00:35:35.530 And I think by just saying what we're going to do is we're going to just put 00:35:35.530 --> 00:35:38.230 a proof of work, you know, anime bot thing up first. 00:35:38.610 --> 00:35:41.210 And then you look at the cute little anime thing and you do proof of work and 00:35:41.210 --> 00:35:42.950 then you get to, you know, view the source code. 00:35:43.510 --> 00:35:50.050 And while that is quaint, it's easily defeatable. I mean, there are so many projects now. 00:35:50.270 --> 00:35:53.750 Even a lot of these aren't harnesses are just coming with things that just defeat it built in. 00:35:54.590 --> 00:35:59.390 And so you're just creating more load and more work and burning more CPU cycles 00:35:59.390 --> 00:36:02.610 and using more electricity for no result. 00:36:02.810 --> 00:36:07.010 And I'm not saying that that's a good thing. I'm saying the way to solve it 00:36:07.010 --> 00:36:10.230 is instead of burying your head in the sand, it's to engage with the process. 00:36:10.530 --> 00:36:15.490 You know, free software didn't get as far as it did because we never engaged 00:36:15.490 --> 00:36:17.890 with licensing debates and legal actions. 00:36:17.890 --> 00:36:22.390 We had organizations that sprung up around free software to protect and fight 00:36:22.390 --> 00:36:27.150 for it in the legal system to give it a legal space where it had to be respected. 00:36:27.350 --> 00:36:31.530 And now you have organizations that are ginormous the size of countries all 00:36:31.530 --> 00:36:34.570 around the world that are following free software licenses for the most part. 00:36:35.516 --> 00:36:36.676 It didn't happen because we 00:36:36.676 --> 00:36:39.616 just ignored commercial software in the legal system and buried our head. 00:36:39.816 --> 00:36:44.916 That happened because parts of the community intelligently engaged and advocated. 00:36:45.116 --> 00:36:49.836 And that's what has to happen here. You're not going to get away from these bots. 00:36:49.996 --> 00:36:53.976 They're going to just walk right around your cute little anime proof of work splash page. 00:36:54.116 --> 00:36:57.036 And they're just going to get the information they want. Or they'll go to a 00:36:57.036 --> 00:37:01.596 project that makes it available via an API JSON markdown file or just doesn't block them. 00:37:02.296 --> 00:37:05.536 And because software is going to be easier to create, easier to extend, 00:37:05.676 --> 00:37:10.156 and easier to patch, it's going to be easier to make the projects that block that more irrelevant. 00:37:10.556 --> 00:37:13.496 And because it's going to be the new generations coming on board that are going 00:37:13.496 --> 00:37:15.756 to be using these tools, you're going to block new adoption. 00:37:15.956 --> 00:37:19.376 And we sure hear a lot of lip service about trying to draw in new users to free 00:37:19.376 --> 00:37:23.896 software, but when the opportunity actually comes along, we're actually gatekeeping, 00:37:24.076 --> 00:37:28.096 and we're blocking them and preventing them because we don't like the tools they're using. 00:37:29.336 --> 00:37:32.396 I think it's you know there are there are certainly valid concerns and 00:37:32.396 --> 00:37:35.296 arguments around sort of impacts to and risks to sort 00:37:35.296 --> 00:37:38.436 of the commons from some of these developments but i 00:37:38.436 --> 00:37:41.916 think what we don't talk about enough is the extent that these same tools can 00:37:41.916 --> 00:37:45.656 enable that like i'm making and publishing more open source these days with 00:37:45.656 --> 00:37:47.996 some of the help of these tools than i ever have before now if anyone wants 00:37:47.996 --> 00:37:51.356 to run it that's up to them but uh like i think there's a version of this where 00:37:51.356 --> 00:37:56.236 we can sort of embrace the good parts and use that to build more of the open 00:37:56.236 --> 00:37:57.616 source stuff that we need. 00:37:57.876 --> 00:38:01.796 But I think, you know, your story is a good one and you really have like leveled up. It's incredible. 00:38:01.956 --> 00:38:06.196 But my story, I think, is maybe more what the community should think about is 00:38:06.196 --> 00:38:09.536 I have been using computers for 40 years. 00:38:11.390 --> 00:38:16.250 I'm getting to be an old man and I've been using computers since like that you 00:38:16.250 --> 00:38:21.490 hooked them up to TVs and you know, like, like a long time I used cartridges 00:38:21.490 --> 00:38:23.430 and like, it's been a while. 00:38:23.750 --> 00:38:29.170 And, um, in 43 ish years of using computers, cause I'm getting old. 00:38:29.510 --> 00:38:35.550 Um, I never once wrote anything more than a line of bash code that I use myself. 00:38:35.830 --> 00:38:39.650 Now you can go to my GitHub repository and I'm releasing software like crazy. 00:38:39.650 --> 00:38:40.910 Now, a lot of it's for myself. 00:38:41.170 --> 00:38:44.430 Some of it are upstream patches. Some of it's for the JB infrastructure or something like that. 00:38:44.910 --> 00:38:48.930 But I am now writing open source code that people are using. 00:38:49.050 --> 00:38:51.050 And it's good. It works for me. 00:38:51.350 --> 00:38:54.170 And it's been a lot of it's been running for months in production. 00:38:54.830 --> 00:38:58.610 So I never created software for 40 years until these tools came along. 00:38:58.730 --> 00:39:00.830 And now I'm creating GPL software. 00:39:02.365 --> 00:39:04.825 And like, we just see like, okay, there are proprietary things, 00:39:04.945 --> 00:39:06.685 right? Like, you know, cloud code and et cetera. 00:39:06.825 --> 00:39:10.185 But there's just, as you expect in our wonderful community, 00:39:10.405 --> 00:39:14.905 like immediately you see all of these various open source harnesses and, 00:39:15.025 --> 00:39:18.245 you know, whatever you think of GitHub itself, like if you just go look at activity 00:39:18.245 --> 00:39:21.065 and various things on GitHub, it's clear that there's just a lot of people excited 00:39:21.065 --> 00:39:24.525 to work on and tweak and share ideas and get inspired. 00:39:24.725 --> 00:39:28.765 It's like, there is a sub aspect of this that is all of the great things we love about open source. 00:39:29.605 --> 00:39:33.265 And I think that probably doesn't get enough attention. But so that's, 00:39:33.325 --> 00:39:36.005 I think, one aspect of this I want to talk about with you guys before we move on. 00:39:36.525 --> 00:39:41.545 And the last one is I really want to stress this point. 00:39:42.545 --> 00:39:45.885 The agent is not the magic part. 00:39:46.305 --> 00:39:50.165 The real reward in this is the infrastructure you build along the way. 00:39:50.945 --> 00:39:54.905 And the more I work on these AI agents like OpenClaw, the more I think the really 00:39:54.905 --> 00:39:57.765 valuable part is the infrastructure I'm setting up around it. 00:39:57.765 --> 00:40:02.565 Because you got to remember these things are often starting with a blank brain for the most part. 00:40:02.665 --> 00:40:06.225 You can't trust them with complete tasks or jobs. You can come. 00:40:06.345 --> 00:40:07.345 Actually, that's the breakdown. 00:40:07.625 --> 00:40:10.645 I don't think you can come. You can trust them with a complete job, 00:40:10.685 --> 00:40:14.965 but you can trust them with a complete task if you give them the skills and 00:40:14.965 --> 00:40:17.545 the guardrails. And what is that when I say that it's generally. 00:40:19.233 --> 00:40:22.153 A Python script that they call with certain flags, depending on the task. 00:40:22.413 --> 00:40:23.813 It's maybe a CLI wrapper. 00:40:25.293 --> 00:40:30.693 So, like, for example, I use GWS. And GWS is a command line client that Google 00:40:30.693 --> 00:40:33.553 has put out to interact with Google Workspace in an agentic safe way. 00:40:34.653 --> 00:40:41.093 And I have like six GWS inboxes. And I have created a wrapper for GWS Unplugged. 00:40:41.613 --> 00:40:44.973 GWS, you know, X, Y, Z. And so when the agent goes to check the inbox, 00:40:45.133 --> 00:40:49.153 there's no ambiguity of what inbox they're checking because they're calling 00:40:49.153 --> 00:40:51.513 the unplugged wrapped GWS client. 00:40:52.013 --> 00:40:55.553 And then there's a skill, which is a markdown file that just says to check the 00:40:55.553 --> 00:40:57.593 inbox, do these steps. I wrote that once. 00:40:58.833 --> 00:41:04.333 And then I can, for the rest of eternity, just ask the agent to go execute that task. 00:41:04.513 --> 00:41:08.153 And that's kind of when I say you need to build the infrastructure around it. Maybe it's a script. 00:41:08.333 --> 00:41:13.333 Maybe it's a CLI. It's probably a skill, something that gives the agent some 00:41:13.333 --> 00:41:15.633 instructions from a completely blank brain. 00:41:15.813 --> 00:41:18.113 And when you build this little bit of scaffolding around it, 00:41:18.193 --> 00:41:19.213 you get incredible results. 00:41:19.373 --> 00:41:22.833 This can also be where it's helpful to say maybe you start in the browser talking 00:41:22.833 --> 00:41:25.593 to something like Cloud or your favorite assistant and you build the spec. 00:41:25.753 --> 00:41:29.553 And then you use something like OpenCode to actually implement and test it and get it ironed out. 00:41:29.653 --> 00:41:32.433 And then you can load it in and run it in your open. There's a lot of ways to 00:41:32.433 --> 00:41:35.573 sort of combine these and not just shove it all through the claw as well. 00:41:35.713 --> 00:41:38.853 I think actually, you know, for you, most of your work isn't in a claw, 00:41:38.973 --> 00:41:40.373 right? you're generally interacting 00:41:40.373 --> 00:41:43.693 through some other application or your app or some interface. It's, 00:41:44.574 --> 00:41:47.094 It's really just what you're trying to get out of it. I mean, 00:41:47.194 --> 00:41:50.374 I think that's interesting. Like, my primary interface is probably the OpenClaw 00:41:50.374 --> 00:41:51.414 agent, but I don't think yours is. 00:41:51.614 --> 00:41:55.194 No, I mean, I usually have OpenCode going on a couple of things, and then I'll have Yap. 00:41:55.374 --> 00:41:59.134 I especially like using Yap for, like, getting skills going because it's a very 00:41:59.134 --> 00:42:03.054 clean environment, so it's just what I've put into the prompts and the history 00:42:03.054 --> 00:42:06.614 that it has, but it has full access to a lot of tools, especially now that it 00:42:06.614 --> 00:42:08.354 has, like, direct search built into it. 00:42:09.434 --> 00:42:12.354 So that helped a lot. And then, yeah, right, and then once I've sort of, 00:42:12.454 --> 00:42:14.374 often I'll build a lot of services, 00:42:14.574 --> 00:42:17.454 maybe it's a new mcp server maybe it's some new scripts like 00:42:17.454 --> 00:42:20.514 i just worked on something to better as a fallback to 00:42:20.514 --> 00:42:24.454 sort of the public convert sites to markdown stuff like markdown.new i wanted 00:42:24.454 --> 00:42:29.434 a mechanism i could run on the command line just as a fallback um and so that 00:42:29.434 --> 00:42:32.554 was useful for the agent as a fallback it's built into one of the search mcp 00:42:32.554 --> 00:42:37.574 servers i'm using now uh and i have it as a tool that i can just also run so 00:42:37.574 --> 00:42:39.774 yeah you can kind of like you know shop them all around. 00:42:39.774 --> 00:42:44.594 I think another thing to remember is they're they're probably going to disappoint 00:42:44.594 --> 00:42:48.334 you the first couple of times you task them to do something because there's 00:42:48.334 --> 00:42:51.754 going to be little bits that you've missed in your scale or in your script. 00:42:51.974 --> 00:42:56.614 And so I when I when I designed to do something new that's going to do routinely, 00:42:56.614 --> 00:43:00.334 I expect the first couple of times it's going to screw it up. 00:43:01.130 --> 00:43:04.930 Because I think of it as a new hire that I've just trained to do something for the first time. 00:43:05.270 --> 00:43:09.290 And you got to expect the new hire is going to need a little handholding a few times they do the task. 00:43:09.510 --> 00:43:12.250 So the first time the agent runs to the task and they screw it up, 00:43:12.290 --> 00:43:16.930 I then use something like open code to go review the logs and figure out where 00:43:16.930 --> 00:43:20.450 the agent went wrong and then go harden up the skill, quote unquote, 00:43:20.730 --> 00:43:23.990 to address that. And then I have the agent run through the cycle again. 00:43:24.310 --> 00:43:27.750 And one of the tricks I do here is I reset the session. 00:43:27.890 --> 00:43:31.230 So it's always a fresh context. so it's not using memory because you always 00:43:31.230 --> 00:43:32.770 want to plan for a fresh context. 00:43:33.010 --> 00:43:36.630 So I'll reset the context and then I'll run through the process again. 00:43:36.830 --> 00:43:40.350 And if it makes a mistake, I'll have OpenCode analyze the session logs, 00:43:40.670 --> 00:43:42.790 figure out where it went wrong and harden the skill again. 00:43:42.930 --> 00:43:45.650 And I iterate on that a few times and usually by the third pass, 00:43:45.710 --> 00:43:46.890 I've caught all that stuff. 00:43:47.010 --> 00:43:51.290 And from that point forward, the thing just runs on its own forever until I want to modify it. 00:43:51.530 --> 00:43:55.050 Or OpenClaw screws something up with some massive update. That could always happen. 00:43:55.230 --> 00:43:58.030 You know, it did strike me. It was very slow, but I was playing around with 00:43:58.030 --> 00:44:01.070 Gemma 4. And even just on the CPU, I could get it to run. 00:44:01.770 --> 00:44:04.610 Not super fast. But, like, I've been doing this parallel work, 00:44:04.630 --> 00:44:05.710 and you're talking about the infrastructure. 00:44:05.930 --> 00:44:07.710 And something that was really clicking for me was just this, 00:44:08.130 --> 00:44:10.630 like, I've been using, searching, SearchXNG more. 00:44:10.850 --> 00:44:13.930 Yeah, yeah, that's really handy. It's a great example of building on something 00:44:13.930 --> 00:44:15.410 that's open source and self-hostable. 00:44:15.530 --> 00:44:18.950 And so at first I was just using it, right? But then I needed to search. 00:44:19.050 --> 00:44:22.130 I didn't want to sign up for a Brave API key, which is the one built into OpenClaw. 00:44:22.270 --> 00:44:24.290 You know, there's various mechanisms to do it. 00:44:24.450 --> 00:44:28.090 I was like, but I have this infrastructure. And so I got OpenCode to help me 00:44:28.090 --> 00:44:31.550 develop an MCP server for it, and that's baked into my injector setup. 00:44:31.730 --> 00:44:36.510 And so now anything, any LLM that makes a call that uses the injector has access 00:44:36.510 --> 00:44:39.270 to search automatically that's routed through my local infrastructure. 00:44:39.590 --> 00:44:42.990 And then so I hooked the Gemma 4 model up to that. 00:44:43.150 --> 00:44:47.710 And so then I was able to have this local model doing direct calls to my local 00:44:47.710 --> 00:44:52.010 search engine provider to then go prepare the report for me or whatever I had to do as a test desk. 00:44:52.110 --> 00:44:56.150 Now, it took four minutes because it was running on the CPU to do a handful of tool calls. But- 00:44:56.816 --> 00:44:57.776 That's just going to get better. 00:44:57.916 --> 00:45:01.036 That was all local using local self-hosted search. 00:45:01.516 --> 00:45:03.736 Now, of course, that search reaches out to DuckDuckGo and various libraries. 00:45:04.116 --> 00:45:06.936 But like, yeah. Yeah, but you're controlling that aspect of it, right? 00:45:06.936 --> 00:45:09.996 I get to set all of that. That's configured declaratively in NixOS, right? 00:45:10.956 --> 00:45:13.116 And that's what I'm trying to come back to. It's like, oh, yeah, 00:45:13.176 --> 00:45:16.436 you have a search XNG, whatever it is, instance. Well, guess what? 00:45:16.536 --> 00:45:17.616 It just got a lot more useful. 00:45:17.796 --> 00:45:20.036 You got Home Assistant. Guess what? It just got a lot more useful. 00:45:20.276 --> 00:45:23.616 You got Frigate DVR. Guess what? You got Tunar. You got Jellyfin. 00:45:23.616 --> 00:45:28.876 You got anything that has an API or a config file just got more useful. 00:45:29.036 --> 00:45:29.176 Yep. 00:45:29.416 --> 00:45:34.976 That's how it works. And so that's the exciting part. But it is very early days. 00:45:34.976 --> 00:45:37.776 And I think you should wait if you can. And things like Gemma 4 are going to 00:45:37.776 --> 00:45:38.656 make it a lot more possible. 00:45:38.816 --> 00:45:43.076 And then just because I can't not, but I think NixOS or some kind of declarative 00:45:43.076 --> 00:45:48.556 infrastructure is ways and keeping everything in Git that you can all very useful 00:45:48.556 --> 00:45:50.756 for this stuff because they can mess things up. 00:45:50.896 --> 00:45:54.936 You can mess things up. config files change a lot or new things happen and just 00:45:54.936 --> 00:45:58.836 having having a lot of snapshots you can roll back a reference is great. 00:45:58.836 --> 00:46:02.036 The way i the way i try to do it with a budget is uh 00:46:02.036 --> 00:46:04.936 i did subscribe to the minimax token plan which is 00:46:04.936 --> 00:46:09.496 a one-time annual charge and then it's so much capacity that i've been throwing 00:46:09.496 --> 00:46:13.136 everything i can at it to try to use it up like audio generation image analysis 00:46:13.136 --> 00:46:15.896 everything and i just i cannot use up the tokens so it's a great problem to 00:46:15.896 --> 00:46:19.436 have because it lets me really experiment but it's not the most advanced model 00:46:19.436 --> 00:46:22.196 it's good minimax it's an open source model. It's very good. 00:46:22.416 --> 00:46:25.816 I would love to run it locally. It's not there yet, though. 00:46:26.754 --> 00:46:32.834 Um, but it's, it's good, but it's not great. And it will often mess up Nix config. 00:46:33.454 --> 00:46:37.394 And the great thing is, is the Nix config has to build and verify. 00:46:37.614 --> 00:46:43.234 And so then the agent sees the build fails, goes and fixes its syntax and runs the build again. 00:46:43.454 --> 00:46:47.074 And I often think if I was on a Debian system or a Red Hat system, 00:46:47.414 --> 00:46:49.874 would it have just injected a bogus config option? 00:46:49.874 --> 00:46:52.914 And then I would have restarted the service and the service would have just failed or whatever, 00:46:53.094 --> 00:46:55.854 or the OS wouldn't have booted and so what wes 00:46:55.854 --> 00:46:58.514 is saying is the reason why it's nice to have it in a declarative environment 00:46:58.514 --> 00:47:01.774 maybe it's even just a container i don't know that you are you know you can 00:47:01.774 --> 00:47:06.114 take image snapshots up or github backups of config files whatever you're doing 00:47:06.114 --> 00:47:09.994 so you can iterate uh is really useful because they're not great yet and of 00:47:09.994 --> 00:47:13.334 course the fancier models are but if you're trying to do this on a budge yeah. 00:47:13.334 --> 00:47:17.774 And having having some kind of feedback mechanism whether it's a linter a format 00:47:17.774 --> 00:47:20.894 or something they can just tell you you know or do a smoke test of any kind 00:47:20.894 --> 00:47:22.694 as a just a fast feedback mechanism. 00:47:22.694 --> 00:47:23.454 Too so. 00:47:23.454 --> 00:47:28.154 You catch mistakes before you're like way down six steps and it's moved on and that helps a lot too. 00:47:28.154 --> 00:47:33.434 We're starting off our boost this week with a space ball boost from kangaroo 00:47:33.434 --> 00:47:37.774 paradox one two three four five six satoshis, 00:47:47.985 --> 00:47:51.825 Thank you, Kangaroo. Here's the message. Fell behind on the shows. 00:47:52.005 --> 00:47:53.365 I'm slowly catching up, though. 00:47:54.125 --> 00:47:59.045 Your pre-show rant on open source projects versus AI really resonated with me, 00:47:59.065 --> 00:48:01.325 so I had to give it some value back. 00:48:01.565 --> 00:48:05.305 I was mostly on board with open source software maintainers and their approach 00:48:05.305 --> 00:48:09.525 to block big tech AI bots, and it seemed reasonable to me. 00:48:10.125 --> 00:48:15.665 But as usual, your words and passion, Chris, made me feel that I should only 00:48:15.665 --> 00:48:20.625 be at most a short-term solution because letting OpenAI and others hammer your 00:48:20.625 --> 00:48:22.985 forge doesn't really seem viable. 00:48:23.245 --> 00:48:26.025 However, on the long term, you're absolutely correct. 00:48:26.345 --> 00:48:29.665 Yeah, that is, it's a tricky thing, right? Because there is a real problem of 00:48:29.665 --> 00:48:33.345 resources and open source projects are limited in resources. 00:48:33.545 --> 00:48:33.825 For sure. 00:48:34.025 --> 00:48:36.525 They don't have time to be chasing server infrastructure problems. 00:48:37.565 --> 00:48:42.985 But again, I don't think you fix that by blocking them because they just go right around you. 00:48:43.325 --> 00:48:45.545 Thank you very much, Kangaroo, for being our baller booster. 00:48:45.905 --> 00:48:47.045 Appreciate you very much. 00:48:48.185 --> 00:48:52.465 Hybrid Sarcasm comes in with a row of McDucks 22,222 sats. 00:48:54.824 --> 00:48:59.044 Having set up my own clanker with OpenClaw, I concur that what I want to find 00:48:59.044 --> 00:49:03.064 is APIs for all the things, starting with LubeLogger. 00:49:03.264 --> 00:49:05.264 Did you see that also Tunar has an API? 00:49:06.444 --> 00:49:09.264 LubeLogger is the one I need to set up. Hey, add the oil change. 00:49:09.624 --> 00:49:13.084 That's another really nice workflow of the thing in your chat app is, 00:49:13.084 --> 00:49:16.404 you know, adding to the grocery cart, adding to something, tracking this and 00:49:16.404 --> 00:49:17.684 writing it up nicely for me, whatever. 00:49:17.944 --> 00:49:21.804 Unbelievably, both Wes and I have grocery stores that have APIs. and 00:49:21.804 --> 00:49:24.584 so we both have wired up our clause to talk 00:49:24.584 --> 00:49:27.744 to the api and i find this very useful actually the 00:49:27.744 --> 00:49:30.624 most when i come home from the grocery store i come home and i'm unloading the 00:49:30.624 --> 00:49:34.324 groceries and i go oh crap i forgot to get and then i open up telegram and i 00:49:34.324 --> 00:49:38.004 go add this add this add this add this add this to my shopping cart and then 00:49:38.004 --> 00:49:40.744 the next time i go it's just there in my shopping ready to go or like dylan 00:49:40.744 --> 00:49:43.564 comes in and he's like dad i i'd like to get some more bottled water i'm like 00:49:43.564 --> 00:49:46.844 bottled water dylan i don't really want some bottled all right let me all right 00:49:46.844 --> 00:49:49.864 add bottled water the shopping cart, fine. And now it's in the shopping cart, right? 00:49:50.544 --> 00:49:52.464 So because if it has an API, that's what I'm doing. 00:49:52.484 --> 00:49:53.444 But don't get the expensive stuff. 00:49:53.504 --> 00:49:56.684 Yeah, right, yeah. Well, and so what I set up is a preferences file. 00:49:57.004 --> 00:50:01.164 So the first time I choose a brand, it remembers that that's my preferred preference. 00:50:01.444 --> 00:50:03.584 And then, you know, size and flavors and stuff like that too. 00:50:04.284 --> 00:50:07.724 But you're absolutely right. Tunar does have an API and it's glorious hybrid 00:50:07.724 --> 00:50:09.384 because it'll let me fix a problem this weekend. 00:50:09.604 --> 00:50:12.684 He also sent us a row of duckies to say happy Easter. 00:50:13.624 --> 00:50:17.584 Happy Easter to you. The show is here on Easter Sunday and we appreciate the value. 00:50:18.624 --> 00:50:20.644 Nyquist comes in with 5,000 cents. 00:50:23.289 --> 00:50:24.709 No message, just value. 00:50:24.949 --> 00:50:25.269 Thank you, sir. 00:50:25.449 --> 00:50:30.009 Appreciate that. And then emasy01 comes in with 4,096 cents. 00:50:31.869 --> 00:50:35.969 For a few years, I've been using Secure Boot with my own keys on my laptop with 00:50:35.969 --> 00:50:38.069 Arch Linux, Windows, and Mac OS. 00:50:38.209 --> 00:50:39.089 Ah, excellent. Okay. 00:50:39.289 --> 00:50:44.909 I generated my init RAMFS with DrawCut, then generated a UKI and signed it with SBCTL. 00:50:45.069 --> 00:50:49.489 I used OpenCore as my boot manager, and I signed that and the Windows boot manager 00:50:49.489 --> 00:50:53.609 as well. After each kernel update, I would regenerate and re-sign the UKI, 00:50:53.709 --> 00:50:56.169 and after every Windows update, I would re-sign the Windows Boot Manager. 00:50:56.509 --> 00:50:59.889 I also encrypted all three operating systems with BcacheFS, encryption, 00:51:00.089 --> 00:51:01.849 BitLocker, and then it kind of gets cut off. 00:51:02.189 --> 00:51:02.589 Interesting. 00:51:02.849 --> 00:51:05.289 This is incredible. Thank you for the experience report. 00:51:05.389 --> 00:51:08.849 Yeah, that's exactly what we wanted. I'm wondering why. I'm wondering what the 00:51:08.849 --> 00:51:11.889 motivation was. Is there a corporate requirement? Because that's a work. 00:51:12.049 --> 00:51:12.809 It is a lot of work. 00:51:12.849 --> 00:51:16.629 Every kernel update? That guy's not messing around. Emacy, let us know. 00:51:17.829 --> 00:51:20.889 Why? I mean, other than just because it's cool to be secure, 00:51:21.009 --> 00:51:23.569 which I agree with. And then come over and set up all our bootloaders for us. 00:51:23.769 --> 00:51:24.149 There we go. 00:51:24.729 --> 00:51:26.709 Thank you, sir, for the boost. Appreciate it. 00:51:27.309 --> 00:51:30.649 Well, our dear Gene being boosted in a series of boosts here. 00:51:30.769 --> 00:51:35.609 There's a couple rows of ducks, some elite boosts for a total of 9,340 sats. 00:51:44.610 --> 00:51:51.210 A little comment on Linux Unplugged 658 saying, I too am thankful for your scale coverage. 00:51:51.390 --> 00:51:52.450 Oh, thanks. We missed you, Gene. 00:51:52.790 --> 00:51:53.570 We sure did. 00:51:53.770 --> 00:51:58.430 At home, I'm using SystemdBoot most places because it's the default in Nix OS. 00:51:58.890 --> 00:52:00.070 Yeah, yeah, yeah. 00:52:00.150 --> 00:52:04.210 And the more I think about it, the more I feel that I'm very against any age 00:52:04.210 --> 00:52:06.730 verification that is done off device. 00:52:07.190 --> 00:52:11.170 I'm not a fan of the idea in general, but can live with something that isn't 00:52:11.170 --> 00:52:13.670 sending me to a third party for that verification. 00:52:13.670 --> 00:52:14.510 These. 00:52:14.510 --> 00:52:18.510 Inherently won't work as well but it's okay considering the alternative is to 00:52:18.510 --> 00:52:23.250 give up any semblance of privacy if i have to prove myself to my computer i've 00:52:23.250 --> 00:52:27.030 just effectively registered it and everything it does with the governments. 00:52:27.030 --> 00:52:32.070 Yeah i guess if you're thinking it's gonna happen one way or the other and if 00:52:32.070 --> 00:52:35.190 it doesn't happen the way it's been talked about now it's probably going to 00:52:35.190 --> 00:52:38.050 happen through some sort of third-party verification that does seem kind of 00:52:38.050 --> 00:52:41.150 a bad direction good point gene i. 00:52:41.150 --> 00:52:46.850 Read a really neat uh contrast to this saying hey the content should tell us 00:52:46.850 --> 00:52:52.490 what age it's appropriate for not us telling them we are and what our age is. 00:52:52.490 --> 00:52:57.950 That's what's always made me think it's really just about waving the hands on 00:52:57.950 --> 00:53:01.450 like the advertisers side well we we checked right we did something if they 00:53:01.450 --> 00:53:04.250 really were trying to prevent it they probably would do it at the content side 00:53:04.250 --> 00:53:06.570 but maybe that gets more legally murky i don't know. 00:53:06.570 --> 00:53:10.870 Gene continues here with Elite Boost. Do you know of any podcast clients or 00:53:10.870 --> 00:53:15.610 other clients that will show the video version that you mentioned being in your feed? 00:53:17.192 --> 00:53:21.432 Podverse and fountain yeah fountain um i know there's a couple of others but 00:53:21.432 --> 00:53:24.232 i don't have direct experience with them recently but they will let you see 00:53:24.232 --> 00:53:27.032 the video version of the show gene thank you for asking gene. 00:53:27.032 --> 00:53:31.552 Also wants to make sure we saw an article here about the euro office launching 00:53:31.552 --> 00:53:37.772 uh europe's open source office rival and links to uh a nice little source here. 00:53:37.772 --> 00:53:43.812 Yeah there's also been quite a nasty breakup between collabra and libre office it's not good it's. 00:53:43.812 --> 00:53:45.652 Never been a better time to be a markdown user. 00:53:45.652 --> 00:53:46.112 And not. 00:53:46.112 --> 00:53:46.732 Need an office. 00:53:46.732 --> 00:53:51.812 Which is a privileged position agree completely that is true thank you gene 00:53:51.812 --> 00:53:56.292 appreciate you very much it's always good to hear from you theo mal comes in with 6 000 sets, 00:53:58.546 --> 00:54:01.226 Oh, he's a long-time listener. First time, Booster. 00:54:01.466 --> 00:54:02.106 Hey-o. 00:54:04.646 --> 00:54:09.326 Love the show, as so does my son, 15. That's great. Fully running Linux and loving every minute. 00:54:09.586 --> 00:54:13.146 I'm wanting to migrate away from Google Workspace and was leaning towards Proton. 00:54:13.266 --> 00:54:15.486 Now I'm thinking maybe Nextcloud might be a good way to go. 00:54:15.626 --> 00:54:18.226 Hosted on a local server. What are your thoughts? 00:54:19.526 --> 00:54:23.166 Well, I think if it's only a handful of users, Nextcloud, well, hmm. 00:54:24.426 --> 00:54:28.146 Depends on what kind of work you want to do on the long term, I would say. 00:54:28.546 --> 00:54:33.606 Yeah. Yeah. I do like this idea. I was going to maybe suggest, 00:54:33.846 --> 00:54:36.426 what's the one that you like? Zoho. 00:54:37.026 --> 00:54:42.126 Zoho. You know, I know it's not self-hosted, but it's a nice alternative to Google Workspace. 00:54:42.926 --> 00:54:47.906 I think you should try it, to be honest with you, Theo, because it's good for 00:54:47.906 --> 00:54:51.486 most people, but it really comes down to users and how they interact with web 00:54:51.486 --> 00:54:53.546 apps and how they take to the performance of NextCloud. 00:54:53.806 --> 00:54:56.246 True. Well, exactly what you're doing with it. Yeah. 00:54:56.246 --> 00:55:00.046 I wish it could be a solid yes. I really want to be able to say that, 00:55:00.166 --> 00:55:02.746 but I just don't think it is if I'm being honest with you, but I think it's 00:55:02.746 --> 00:55:04.646 a worth trying. Is that fair? 00:55:04.886 --> 00:55:07.266 Yeah. I mean, different people have different standards, different needs, 00:55:07.946 --> 00:55:09.966 different performance characteristics that they're okay with. 00:55:11.006 --> 00:55:13.306 And honestly, we'd love to hear feedback from people that have it out there 00:55:13.306 --> 00:55:15.886 working successfully as a Google workspace alternative. 00:55:16.046 --> 00:55:20.686 No, not just talking file sync or your darn photos, talking full Google workspace 00:55:20.686 --> 00:55:22.726 sync alternative or workspace alternative. 00:55:22.946 --> 00:55:25.306 It's also hard. Cause I don't know. Maybe there are, there probably are some, 00:55:25.306 --> 00:55:29.626 but like i don't love even google workspace or the microsoft offering so i don't 00:55:29.626 --> 00:55:36.046 know what the best version even is uh forward humor boosts in with 4444 sets, 00:55:38.537 --> 00:55:43.517 Hey guys, I'm enjoying hearing the compliance conversation with Determinant Systems in episode 657. 00:55:43.757 --> 00:55:47.797 Have you heard of anyone running NixOS in a CMMC or ITAR environment? 00:55:48.057 --> 00:55:52.617 I'm not sure if it's even possible to meet FIPS requirements on NixOS and would love any input. 00:55:53.037 --> 00:55:56.057 I imagine these are the exact kind of problems Determinant Systems and Phlox 00:55:56.057 --> 00:55:57.277 are trying to solve, right? This 00:55:57.277 --> 00:56:00.177 is the value add that they can bring to enterprises that are using Nix. 00:56:00.397 --> 00:56:04.557 Yeah, I would probably go ask around maybe on the NixOS discourse could be one 00:56:04.557 --> 00:56:08.217 spot, maybe also see, and you may have done these things already, so, you know. 00:56:08.537 --> 00:56:12.757 Maybe go troll some recent NICS conferences. There might be folks talking about 00:56:12.757 --> 00:56:15.357 that kind of thing, because there definitely are people exploring this space. 00:56:15.517 --> 00:56:18.857 It's exactly where the progress is and the exact things I don't necessarily know. 00:56:19.337 --> 00:56:22.877 And then, yeah, third, maybe go reach out to folks like Detsys or Phlox or various 00:56:22.877 --> 00:56:27.037 folks who are more interfaced with people who might be in those environments. 00:56:27.197 --> 00:56:29.757 And you might be able to find some people who would be willing to have a talk. 00:56:30.217 --> 00:56:34.657 Yeah, it is an area that Red Hat has focused on and SUSE has focused on for 00:56:34.657 --> 00:56:38.217 a very long time, just trying to get each one of those checked off over the years. 00:56:38.537 --> 00:56:40.957 And I think you're seeing the same process start with Nix, but I'm not sure. 00:56:41.097 --> 00:56:44.337 It might also depend too, right? Like, are you trying to run something that 00:56:44.337 --> 00:56:47.497 ultimately builds a container that has an S-bomb that runs on whatever? 00:56:47.497 --> 00:56:50.317 Or are you trying to run like full fat Nix OS? 00:56:50.877 --> 00:56:54.157 Yeah, yeah, yeah, yeah, exactly. All right. Well, thank you everybody who supported 00:56:54.157 --> 00:56:57.697 the show with a boost. We really do very, very much appreciate it. 00:56:57.977 --> 00:57:00.537 Thank you everybody who also streamed. Sats 18 of you did that. 00:57:00.657 --> 00:57:03.657 And collectively, you stacked 33,366 Sats. 00:57:04.157 --> 00:57:10.817 Not too bad. Not too bad. You combine that with our boosters, 208,174 sats. 00:57:11.317 --> 00:57:15.937 Thank you very much for supporting episode 661 of your unplugged program. 00:57:16.117 --> 00:57:19.517 If you'd like to send us a boost, I think Fountain makes it probably the easiest. 00:57:19.657 --> 00:57:24.017 There is a whole self-hosted route you can go with AlbiHub and a bunch of apps, which is a lot of fun. 00:57:24.097 --> 00:57:26.217 We talked some of that in a recent episode of This Week in Bitcoin, 00:57:26.277 --> 00:57:29.037 if you want to check that out. And thank you also to our members. 00:57:30.925 --> 00:57:35.245 One pick this week. That's a rarity. Which is technically the rule of the pick 00:57:35.245 --> 00:57:36.745 segment. It's only supposed to be one. 00:57:36.745 --> 00:57:37.545 Yeah, well, we... 00:57:37.545 --> 00:57:42.025 It's been almost a year, I think. But we wanted to talk about single-file CLI 00:57:42.025 --> 00:57:46.605 because it solves a pretty, pretty handy problem. Or whatever. 00:57:47.125 --> 00:57:51.845 It's a CLI tool that solves a problem that I've had probably, 00:57:51.865 --> 00:57:55.645 I don't know, forever. Because I used to solve it with something built into Netscape, to Firefox. 00:57:56.577 --> 00:58:01.737 Is a complete copy of a web page in a single HTML file based on single file. 00:58:01.857 --> 00:58:03.597 And this is single file CLI. 00:58:03.777 --> 00:58:07.997 Yeah, that's right. A CLI tool for saving a faithful copy of a complete web page. 00:58:08.117 --> 00:58:10.637 And crucially, right, like you can do that in a variety of ways. 00:58:10.757 --> 00:58:12.437 And there are probably better or different tools. 00:58:12.677 --> 00:58:16.357 So, you know, boost in, write in if you have a preferred version of getting 00:58:16.357 --> 00:58:18.457 this task accomplished or archiving websites. 00:58:18.797 --> 00:58:22.737 But I liked sort of the idea that maybe for your own archive, 00:58:22.957 --> 00:58:27.797 for processing somehow, whatever you're trying to do of just like a single html 00:58:27.797 --> 00:58:32.477 file per per site instead of having stuff that's like vendored a bunch of images 00:58:32.477 --> 00:58:37.357 into a folder which is better for some use cases for sure but not for this not for simplicity. 00:58:37.357 --> 00:58:40.377 And this could be good or bad depending on your opinion but i like that 00:58:40.377 --> 00:58:43.777 it uses chrome or chromium and then it uses dino 00:58:43.777 --> 00:58:46.557 as a standalone script injected into the web page using the 00:58:46.557 --> 00:58:50.397 chrome dev tools protocol to actually render it through chrome so if it's a 00:58:50.397 --> 00:58:53.677 website chrome can render you're going to be able to capture it with this which 00:58:53.677 --> 00:58:58.017 means everything right so i use firefox as my daily driver but that's absolutely 00:58:58.017 --> 00:59:02.017 for me valuable you just have to have chrome or chromium installed or a chromium 00:59:02.017 --> 00:59:05.517 based browser and then uh you know be able to support that remote. 00:59:05.517 --> 00:59:07.117 Yeah the dev tools. 00:59:07.117 --> 00:59:08.317 Protocol which is pretty. 00:59:08.317 --> 00:59:08.877 Easy these days. 00:59:08.877 --> 00:59:13.477 It is it's very easy these days with the current version so single file cli 00:59:13.477 --> 00:59:21.937 and it is agpl 3.0 thank you wes agpl 3.0 So you could always use your suggestions on some picks, too. 00:59:22.017 --> 00:59:25.317 If there's something that you find very handy that you run on your Linux box 00:59:25.317 --> 00:59:29.097 that we haven't covered or makes your server more useful, send it in to us. 00:59:29.397 --> 00:59:31.937 We'd love to cover it because we're always looking for great, 00:59:32.037 --> 00:59:34.657 useful tools. And I feel a little bad that we only had one pick for you, 00:59:34.737 --> 00:59:35.277 even though that's technically a rule. 00:59:35.277 --> 00:59:36.137 Yeah, here I got a bonus pick. 00:59:36.217 --> 00:59:37.357 What? No way. 00:59:37.577 --> 00:59:39.357 Okay. It's called HTML2Markdown. 00:59:39.857 --> 00:59:40.177 Okay. 00:59:40.397 --> 00:59:43.977 And it's just a single, it's a Go project. This was one of the inspirations 00:59:43.977 --> 00:59:46.657 for me making my own tool, which you can use that, too, if you want. 00:59:46.777 --> 00:59:49.197 I'll throw a link in there. but it's really more meant I just wanted it more 00:59:49.197 --> 00:59:52.877 as a library so this is a sneaky double pick yeah it is and this was one of 00:59:52.877 --> 00:59:57.157 the inspirations along with the Mozilla readability like for their reader mode stuff you. 00:59:57.157 --> 00:59:59.497 Know what Wes you make me want to be a better man, 01:00:01.302 --> 01:00:01.842 Impressive. 01:00:02.142 --> 01:00:06.462 A robust HTML to Markdown converter that transforms HTML, even entire websites 01:00:06.462 --> 01:00:07.802 into clean, readable Markdown. 01:00:07.962 --> 01:00:11.822 Supports complex formatting, customizable options, and plugins for full control. 01:00:11.982 --> 01:00:17.982 But it can handle, you know, tables and complicated nesting and a lot of nice stuff. 01:00:18.082 --> 01:00:21.582 So you're saying I can take like those Libre documents and I can rage quit Libre 01:00:21.582 --> 01:00:25.562 Office because they got beef with Calabra and I like Calabra a lot. 01:00:25.702 --> 01:00:28.662 And now I can just convert them all to Markdown and have beautiful Markdown-rated 01:00:28.662 --> 01:00:30.762 versions of documents even if they have tables in them? 01:00:31.482 --> 01:00:34.762 You know, I don't know about it. It depends on how dynamic. If it's all JavaScript 01:00:34.762 --> 01:00:35.622 rendered, then maybe not. 01:00:35.662 --> 01:00:39.062 I think you just should have said no. Wait a minute. Hold on. 01:00:39.282 --> 01:00:41.202 There's JavaScript and LibreOffice documents? 01:00:42.022 --> 01:00:43.162 Well, if you're talking about a website. 01:00:43.702 --> 01:00:45.562 Oh, no, I'm talking LibreOffice. Okay. 01:00:45.842 --> 01:00:47.362 Yeah, you're in Office world. 01:00:47.482 --> 01:00:49.842 HTML? Because I thought they could save them as HTMLs. I don't know. 01:00:50.022 --> 01:00:52.622 Oh, probably. Yeah, well, then that should work. Okay. I thought you were thinking 01:00:52.622 --> 01:00:55.222 like some sort of online interface. I don't know. 01:00:55.482 --> 01:00:59.822 I'm trying to rage quit LibreOffice because I got beef with my boys at Calabra, 01:00:59.822 --> 01:01:03.422 and I thought you maybe bring me a tool to make that easier and you're just shutting me down. 01:01:03.422 --> 01:01:06.822 No, I'm trying to help your claw read stuff on the internet. 01:01:07.062 --> 01:01:09.602 That's useful and relevant and on theme for the show. Or make your archive. 01:01:10.002 --> 01:01:12.962 You know, instead of some stuff you want full HTML, some stuff you just want 01:01:12.962 --> 01:01:15.682 markdown because do you need the full HTML of the Pharonix article? 01:01:15.762 --> 01:01:19.102 He's doing it again, Brent, where he's bringing a show relevant thing. It's on theme. 01:01:20.002 --> 01:01:21.322 I'm kind of with Wes on this one. 01:01:22.522 --> 01:01:25.982 You probably should be. All right, links to that are in the show notes over 01:01:25.982 --> 01:01:31.542 at linuxunplugged.com slash 661. You can find all of that and our contact information, 01:01:31.662 --> 01:01:33.822 RSS feed, the mumble room. But you know what? 01:01:34.282 --> 01:01:38.862 If their claw does want more information, more metadata, or maybe they're just their own... 01:01:39.838 --> 01:01:42.098 Hooks we got some for them don't we wes. 01:01:42.098 --> 01:01:47.738 Oh we got a data structure data rich rss feed it's an xml file don't you know. 01:01:47.738 --> 01:01:50.318 I like xml well i don't but the machines do well. 01:01:50.318 --> 01:01:53.158 And you can have namespaces which is pretty cool because then you can put the 01:01:53.158 --> 01:01:56.038 podcast namespace in there and that's got all kinds of fancy goodies. 01:01:56.038 --> 01:01:59.138 You could just ask your claw to expose the mp4 that snuck in there that your 01:01:59.138 --> 01:02:03.898 podcast client doesn't show you yeah that's right what about like uh information 01:02:03.898 --> 01:02:06.438 that was in the for the content of the show wes there's got to be something 01:02:06.438 --> 01:02:10.718 i can like the description tag No, I mean, yeah, I guess that's a starter. 01:02:10.818 --> 01:02:12.138 Or like the iTunes keywords? 01:02:12.258 --> 01:02:16.878 No, that seems old. No, I want to know all the brilliant things Brent said. 01:02:17.318 --> 01:02:19.498 Oh, for that, you want cloud chapters? 01:02:19.818 --> 01:02:21.298 Well, I mean, that would get me close. 01:02:21.358 --> 01:02:22.918 That gives you the Brent section of the show. 01:02:23.038 --> 01:02:26.478 Okay, okay, okay. I thought maybe you'd have a transcript for me or something. 01:02:26.598 --> 01:02:27.118 Oh, yeah. 01:02:27.238 --> 01:02:27.978 Oh, yeah, yeah, yeah. 01:02:28.058 --> 01:02:29.458 Yeah, no, we do, actually. 01:02:29.678 --> 01:02:30.198 Oh, we do. 01:02:30.378 --> 01:02:32.038 Yeah, VTT and SRT. 01:02:32.158 --> 01:02:35.918 You might even say we've had it for a couple of years now, but it is handy more so than ever. 01:02:36.058 --> 01:02:38.778 All right, and then last but not least, a little bit of metadata for you. 01:02:38.858 --> 01:02:42.358 We are live every single Sunday over at jblive.tv. 01:02:45.898 --> 01:02:50.098 Yeah, we do it at 10 a.m. Pacific, 1 p.m. Eastern, and of course we gots it over at, 01:02:52.678 --> 01:02:56.038 jupiterbroadcasting.com slash calendar. Say it like that, jupiter. Jup. 01:02:56.558 --> 01:03:00.558 And if you go to jupiterbroadcasting.com slash calendar, the script will just 01:03:00.558 --> 01:03:02.698 convert it to your local time zone so you don't even have to do the math. 01:03:02.698 --> 01:03:04.838 You can come hang out in our mumble room, our chat room. 01:03:05.138 --> 01:03:08.978 Tell us we're a bunch of goofballs we love it and help title the show as well, 01:03:09.638 --> 01:03:12.258 shout out to our members for supporting this episode and for everybody tuning 01:03:12.258 --> 01:03:15.258 in that shares it we always appreciate that as well word of mouth is the number 01:03:15.258 --> 01:03:17.198 one way to promote a podcast also. 01:03:17.198 --> 01:03:19.678 Brent better leave soon because Linux Fest Northwest is coming up. 01:03:19.678 --> 01:03:20.478 And the schedule is. 01:03:20.478 --> 01:03:21.238 Actually live now. 01:03:21.238 --> 01:03:24.298 Linux Fest Northwest go check it out schedules live and we'll see you there 01:03:24.298 --> 01:03:27.858 it's going to be great we're going to do a live show thanks so much for joining us see you next Sunday.