WEBVTT 00:00:11.384 --> 00:00:16.091 Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris. 00:00:16.491 --> 00:00:17.251 My name is Wes. 00:00:17.251 --> 00:00:18.321 And my name is Brent. 00:00:19.064 --> 00:00:23.081 Well, hello, gentlemen. Coming up on the show today, your favorite free software 00:00:23.081 --> 00:00:25.350 projects, they're not taking the summer off. 00:00:25.629 --> 00:00:27.571 We're going to round up the releases that you need to know about, 00:00:27.571 --> 00:00:30.552 including some big changes coming to a kernel near you very soon. 00:00:30.987 --> 00:00:33.199 Then we'll round out the show with some great boos, some picks, 00:00:33.535 --> 00:00:37.111 and a lot more. So before we get there, let's say time-appropriate greetings 00:00:37.111 --> 00:00:40.261 to that virtual lug of ours. Hello, mumble room. Hello, hello. Hello. 00:00:40.931 --> 00:00:43.191 Hey Chris, hey Wes, and hello Brent. 00:00:43.191 --> 00:00:43.851 Hi guys. 00:00:43.851 --> 00:00:44.629 Hello. 00:00:44.791 --> 00:00:47.861 Hello everybody up there in the quiet listening too. Nice to have you out there. 00:00:47.861 --> 00:00:50.498 Also happy Father's Day to all the dads out there listening. 00:00:51.601 --> 00:00:52.854 You know, happy Father's Day. 00:00:53.150 --> 00:00:55.792 Happy Father's Day to the dad dogs out there too, and the squirrel dads. 00:00:55.931 --> 00:00:58.741 Oh, thank you. Happy Father's Day to you as well. 00:00:58.741 --> 00:00:59.685 And the cat dads. And the cat dads. 00:01:01.090 --> 00:01:04.765 Cat dads, I think, would count, right? Of course. Dads will come in all shapes and sizes. 00:01:05.467 --> 00:01:08.588 Also, good morning to our friends over at Defined Networking. 00:01:08.588 --> 00:01:10.581 Go to defined.net slash unplug. 00:01:10.930 --> 00:01:14.428 Your network is the most important part of your stack. Let's be real about that. 00:01:14.428 --> 00:01:18.476 It's your core infrastructure. Performance and control really, really matter here. 00:01:19.149 --> 00:01:23.422 That's why we love Nebula. So you can get started with Managed Nebula from Defined Networking. 00:01:23.793 --> 00:01:28.403 You go to defined.net slash unplug, you get set up 100 hosts for free. 00:01:28.733 --> 00:01:29.842 No credit card required. 00:01:30.208 --> 00:01:34.198 It's built on the open-source Nebula platform, which was originally created 00:01:34.198 --> 00:01:38.039 to connect Slack's global infrastructure of Snacks. 00:01:38.778 --> 00:01:41.918 I mean, you can imagine how complex that was from the go. That was day one, 00:01:41.918 --> 00:01:43.583 and they've just built on that since. 00:01:44.250 --> 00:01:47.791 And unlike traditional VPNs, Nebula is totally decentralized. 00:01:48.227 --> 00:01:50.838 Your hosts connect directly to each other, and then you can set up your own 00:01:50.838 --> 00:01:54.589 lighthouses. You can use Manage Nebula to discover them. You control that aspect of it. 00:01:54.943 --> 00:01:57.687 No mysterious control plane between you and your nodes. 00:01:58.314 --> 00:02:02.198 So if it's your home lab or it's a global fleet, you get fast, 00:02:02.198 --> 00:02:06.778 secure networking without giving up ownership of the stack. And that's what I love about Nebula. 00:02:08.096 --> 00:02:12.798 Try it. 100 hosts for free. No credit card required. Go to define.net slash 00:02:12.798 --> 00:02:17.703 unplugged. Redefine your VPN experience. That is define.net slash unplugged. 00:02:25.562 --> 00:02:29.497 Let's do the news, boys, and let's start with one of the stories I am really 00:02:29.497 --> 00:02:32.757 genuinely excited about because a feature I have been waiting for for a very 00:02:32.757 --> 00:02:34.310 long time has landed in Plasma. 00:02:34.879 --> 00:02:41.747 Plasma 6.7 was released this last week, and they finally baked in per-screen virtual desktop. 00:02:42.727 --> 00:02:44.607 I think they did it just for you as well. 00:02:44.607 --> 00:02:46.917 No, that's not true at all, but I am very grateful. 00:02:46.917 --> 00:02:52.120 I think that took 21 years is what I saw. There was a bug open from 2005. 00:02:52.317 --> 00:02:55.727 And I think I've been complaining about that on there for about 21 years. 00:02:57.255 --> 00:02:58.487 So have you tried it? 00:03:00.233 --> 00:03:03.267 No, because the only machine I have right here in the studio only has one screen right now. 00:03:03.267 --> 00:03:03.847 That's true. 00:03:03.847 --> 00:03:07.707 I know, it's horrible. I will, though, because Aurora just updated so fast, 00:03:07.707 --> 00:03:13.597 so I got it right away. But immediately can notice, I hate to use this phrase, but it is snappier. 00:03:13.597 --> 00:03:14.817 It is. I'm running it as well. 00:03:14.817 --> 00:03:15.267 Oh, you are? 00:03:15.267 --> 00:03:15.517 Yeah. 00:03:15.517 --> 00:03:16.067 And you agree. 00:03:16.067 --> 00:03:19.436 Shout out to the next packages in K900 for the great work packaging stuff there. 00:03:19.987 --> 00:03:21.097 Oh, I noticed it immediately. 00:03:21.097 --> 00:03:22.117 Yeah. Me too. 00:03:22.117 --> 00:03:22.827 I didn't expect that. 00:03:23.187 --> 00:03:23.854 I didn't either. 00:03:24.150 --> 00:03:25.897 I mean, because it's not like Plasma's sluggish. 00:03:25.897 --> 00:03:31.257 No. I would not have categorized it as sluggish either. But so this release 00:03:31.257 --> 00:03:35.797 is the all Wayland, you know, release. And Nate Graham confirmed that 79% of 00:03:35.797 --> 00:03:37.807 Plasma 6 users were already on Wayland. 00:03:37.807 --> 00:03:40.077 Wow. That's a lot. 00:03:40.077 --> 00:03:42.677 A lot of Wayland improvements in here. A lot of Wayland improvements. 00:03:42.687 --> 00:03:45.007 So like X11's gone for good in 6.8? 00:03:45.007 --> 00:03:45.187 Yeah. 00:03:45.187 --> 00:03:46.477 Something like early 2027? 00:03:46.487 --> 00:03:51.247 Yeah, it's gone for good in 6.8. But look at this. Weyland only remains planned, 00:03:52.427 --> 00:03:54.886 almost 80% now are on Weyland. Look at that. 00:03:55.037 --> 00:03:55.977 That's pretty wild. 00:03:58.161 --> 00:04:01.487 I mean, it is really a smooth experience now. I mean, I think it is generally pretty good. 00:04:02.131 --> 00:04:06.837 They fixed a bug, too, that was impacting Aurora and Discover, 00:04:06.837 --> 00:04:10.113 which is nice to see. Just one of those little small things that they worked in here. 00:04:11.127 --> 00:04:13.837 The polish overall is something that people often talk about, 00:04:13.837 --> 00:04:18.345 but it does seem like every single release, we just keep seeing an iteration on Breeze. 00:04:18.838 --> 00:04:23.507 It's getting better and better. so what I did was once I had the default or once I had the install, 00:04:24.747 --> 00:04:28.547 I switched everything back to all the default so I could see the new styles 00:04:28.547 --> 00:04:33.757 and everything like that and I love it I love it very happy with this release 00:04:33.757 --> 00:04:37.844 just only been using it for like two days so far but those two days have been really good. 00:04:38.331 --> 00:04:42.247 There's some interesting things in place for like what they're calling the union 00:04:42.247 --> 00:04:45.907 engine they're working on underneath basically they're going to use CSS and 00:04:45.907 --> 00:04:50.417 then a Rust parser from Servo to like parse that and handle it. 00:04:50.754 --> 00:04:55.427 And then it's going to be basically like one style to rule them all eventually. 00:04:55.427 --> 00:04:59.337 It's disabled by default for now. It's a work in progress, but you can see they're 00:04:59.337 --> 00:05:00.528 doing a lot of work under the hood. 00:05:00.888 --> 00:05:03.447 A feature they've added that I think we'll find useful is you can now test your 00:05:03.447 --> 00:05:06.566 microphone volume right there in the little volume pop-up widget. 00:05:07.247 --> 00:05:08.547 Now that is slick. 00:05:09.067 --> 00:05:09.787 It's really nice. 00:05:09.787 --> 00:05:14.757 There's also a nice little touch here. It looks like the oxygen theme is returning 00:05:14.757 --> 00:05:17.727 for KD's 30th anniversary. Can you believe that? 00:05:17.727 --> 00:05:18.646 30 00:05:19.395 --> 00:05:24.407 And apparently it was quote really really in bad shape uh before the restoration 00:05:24.407 --> 00:05:29.037 which was led by philip phila so if you i don't know they were nostalgic for 00:05:29.037 --> 00:05:32.229 the oxygen theme uh well there's a modern version now. 00:05:32.996 --> 00:05:36.887 Yep improvements to the clipboard manager color lots of performance improvements 00:05:36.887 --> 00:05:42.295 um gpu rendering optimizations too so you don't have to have a big old gpu to do all this stuff, 00:05:42.922 --> 00:05:46.847 here's a couple of other just uh quick hits you can now sync your mouse and 00:05:46.847 --> 00:05:49.877 stylus pointers, an option to set and change preferred to the calendar app, 00:05:49.877 --> 00:05:53.077 an option to assign keyboard shortcuts to toggling the global push to talk microphone 00:05:53.077 --> 00:05:54.595 mute throughout the plasma applications. 00:05:55.030 --> 00:05:57.207 There's now an easy way of selecting mixed skin tones and emojis. 00:05:57.207 --> 00:05:59.507 So you can just have, oh, there's the new emoji thing, the system monitor app 00:05:59.507 --> 00:06:02.339 that respects your preferences regarding storage units for gigabytes versus, 00:06:03.935 --> 00:06:08.678 Another gigabyte display. I mean, it's like a long list of things to go through here. 00:06:09.483 --> 00:06:13.313 And, like, a lot of times we joke that nothing really comes out in the summer. 00:06:13.313 --> 00:06:17.107 But, man, this is one of the major, major releases for Plasma in general. 00:06:18.279 --> 00:06:23.852 And here we are in the summer. It's really great. And it's impressive how fast we're getting it. 00:06:24.096 --> 00:06:26.423 You've gotten it already. I mean, thanks to the maintainers. 00:06:26.723 --> 00:06:28.624 But I'm running it already. 00:06:29.088 --> 00:06:31.583 I just seem to recall it used to be months before I get my hands on it. 00:06:31.583 --> 00:06:34.823 Absolutely. We live in a different world now. At least it depends on your distro. 00:06:34.823 --> 00:06:38.063 But you have the options now if you want something that lets you get it real fast. 00:06:38.916 --> 00:06:43.873 And, I mean, I think it's just also, you know, the project ships quite regular 00:06:43.873 --> 00:06:48.910 releases now. It seems like that cadence is very well polished. So it works well. 00:06:49.612 --> 00:06:52.483 Yeah, shout out to them, too, for that microphone volume test. 00:06:52.483 --> 00:06:54.262 I think that's going to be really useful for our guests. 00:06:55.034 --> 00:06:58.763 Brentley, you tagged a story that Firefox shared with us. I guess they're kind 00:06:58.763 --> 00:07:02.933 of putting out there essentially what is a roadmap for the future features that 00:07:02.933 --> 00:07:04.752 they plan to add to Firefox in the near term. 00:07:05.315 --> 00:07:09.133 Yeah, I thought this was interesting because it's been a while since I've dove 00:07:09.133 --> 00:07:11.503 into what they're trying to do over there. 00:07:11.979 --> 00:07:15.043 And there's been a little bit of controversy recently with some of their AI 00:07:15.043 --> 00:07:19.235 features, you know, and they added an on and off button because of all of the pushback. 00:07:19.659 --> 00:07:24.489 So I was just curious about like, well, what are they thinking the future of Firefox will be? 00:07:25.655 --> 00:07:30.453 And it's kind of interesting. Like there's a lot that they're working on in 00:07:30.453 --> 00:07:36.573 PDFs. So right in the browser, you will be able to split, merge, 00:07:36.573 --> 00:07:40.406 and reorder PDFs, which is an interesting thing. 00:07:42.363 --> 00:07:47.963 It makes me believe more that the browser is actually an OS on top of our OS. 00:07:47.963 --> 00:07:52.166 But for most regular users, I imagine that's really interesting. 00:07:52.549 --> 00:07:58.075 Also, this feature has been asked for for basically ever, customizable built-in hotkeys. 00:07:58.656 --> 00:08:03.543 Yeah. I will note, too, what they've done is this website that you shared with 00:08:03.543 --> 00:08:05.877 us, which we'll have a link in the show notes. 00:08:06.527 --> 00:08:09.503 For the features that you can try now, they just have a button here. 00:08:09.503 --> 00:08:11.163 So they have like this roadmap of features. 00:08:11.783 --> 00:08:16.823 And if that feature is available for testing, they have a try now button. That's a neat approach. 00:08:16.823 --> 00:08:21.003 And they have them by category as well, like productivity. There's privacy controls 00:08:21.003 --> 00:08:24.772 that they're working on. AI done differently, they call it. That's the category name. 00:08:25.684 --> 00:08:29.203 Speed and performance, of course, built-in protection, better web stuff, 00:08:29.543 --> 00:08:32.453 which is an interesting category. So if you're interested in where the browser 00:08:32.453 --> 00:08:34.118 is going, it's worth looking at. 00:08:34.350 --> 00:08:40.683 And I just felt good about this. I think having a map like this feels productive. 00:08:40.683 --> 00:08:44.673 I had totally lost sight of what Firefox was even trying to do. 00:08:44.673 --> 00:08:48.235 And some of this makes me feel a little better about the future of Firefox. 00:08:48.687 --> 00:08:51.713 I guess it answers that question. It's like, what is Mozilla even doing for Firefox? 00:08:51.713 --> 00:08:52.745 Exactly. 00:08:53.122 --> 00:08:55.874 Here you go. This is what they're doing. They made a what's next page. 00:08:56.652 --> 00:08:59.566 The thing that grabbed my attention was new web APIs. 00:09:00.849 --> 00:09:06.316 Wasm.js promise integration, deferred module evaluation, web transport, and more. 00:09:06.876 --> 00:09:10.796 Also some CSS improvements to support more stuff there. So all stuff making 00:09:10.796 --> 00:09:13.876 Firefox better at handling whatever random web page you get. 00:09:14.276 --> 00:09:18.038 HDR video on Linux was listed there. 00:09:18.252 --> 00:09:22.147 This one stood out to me as well. Pasci support for Firefox Sync. 00:09:22.618 --> 00:09:25.555 I don't think you would boiler Sync users, but that's interesting. 00:09:25.839 --> 00:09:27.696 If you are, you probably want Pasci support on this one. 00:09:27.696 --> 00:09:31.366 I'm not using Firefox to manage my passwords, but I do use it to sync everything 00:09:31.366 --> 00:09:35.636 else. Huh. Huh. Okay. Yeah. Good find, Brent. Thank you. 00:09:35.656 --> 00:09:38.716 There's some stuff in there. There's something, you know, it's still doing stuff, 00:09:38.716 --> 00:09:40.131 which makes me feel a little better. 00:09:41.112 --> 00:09:46.216 Wes, you were excited about the SystemD 261 release, and I was looking through 00:09:46.216 --> 00:09:47.891 it. There are a couple of interesting goodies in here. 00:09:48.025 --> 00:09:52.926 Yes, indeed. One you'll notice is the new Instance Metadata Service, 00:09:52.926 --> 00:09:57.956 or IMDS. Of course, you get systemd-imdsd as part of that. 00:09:57.956 --> 00:09:58.536 Gotta have that. 00:09:58.896 --> 00:10:02.206 It's a unified way to access virtual machine metadata. 00:10:02.334 --> 00:10:06.386 In particular, you're going to see this a lot because it interworks with major 00:10:06.386 --> 00:10:08.966 cloud platform and their metadata services, right? 00:10:08.966 --> 00:10:15.006 So if you boot up an AWS EC2 instance, then there's an IP you can query and 00:10:15.006 --> 00:10:19.856 get info about all the stuff, plus all the stuff that's exposed by the virtual 00:10:19.856 --> 00:10:23.470 BIOS and whatever is programmed in at the virtual machine layer. 00:10:23.958 --> 00:10:28.746 And so now all of that kind of stuff, sort of classifying what are you about, 00:10:28.746 --> 00:10:32.536 like where you are, the VM environment, what it knows about the VM is exposed 00:10:32.536 --> 00:10:33.960 in this metadata service. 00:10:34.772 --> 00:10:37.994 And you have one spot to look if you hook it up with the systemd way. 00:10:38.273 --> 00:10:42.216 I suppose, you know, in my current mindset, where that seems useful would be 00:10:42.216 --> 00:10:45.986 being able to query it at like a scale and then build an inventory and things 00:10:45.986 --> 00:10:49.156 like that, right? It sort of almost becomes like an API of information about your system. 00:10:49.456 --> 00:10:52.986 And you maybe don't have to do the handling of what's the difference between 00:10:52.986 --> 00:10:59.477 how Amazon does it or Azure does it or Oracle Cloud does it or Hetzner's in here too or even Volter. 00:11:00.516 --> 00:11:02.441 Okay, that's not what I thought you were going to start with. 00:11:04.661 --> 00:11:06.691 Oh, you, okay, okay, well, it's finally happening. 00:11:06.691 --> 00:11:07.361 Yeah, here we go. 00:11:07.381 --> 00:11:08.281 It's finally happening. 00:11:08.281 --> 00:11:09.151 Ladies and gentlemen. 00:11:09.151 --> 00:11:14.971 Yes, okay, another notable update here is live update orchestrator and kernel 00:11:14.971 --> 00:11:17.271 handover support in System D. 00:11:17.321 --> 00:11:19.101 I thought that's what you'd start with. 00:11:20.001 --> 00:11:22.151 And this has been building, there's been, like, other parts of this. 00:11:22.151 --> 00:11:25.531 This, in particular, includes support for user units and stuff. 00:11:25.531 --> 00:11:28.775 But the idea is, so there's KExec, right? 00:11:29.048 --> 00:11:33.361 Who wants to reboot? it's so slow let's just jump immediately to a new kernel 00:11:33.361 --> 00:11:34.781 without rebooting the system I'm sorry KExec. 00:11:34.781 --> 00:11:35.422 Is too slow? 00:11:35.718 --> 00:11:39.561 No no KExec's great I'm just trying to say like if you haven't heard about KExec 00:11:39.561 --> 00:11:42.191 for a while that's what it does okay yeah right instead of doing a whole reboot. 00:11:42.191 --> 00:11:43.001 You do a live swap. 00:11:43.001 --> 00:11:46.501 Yeah you just load you like load the new kernel into memory and you jump into it. 00:11:46.821 --> 00:11:50.041 So you're really avoiding the whole post process and like you know the bootloader. 00:11:50.461 --> 00:11:50.663 Mhm, 00:11:51.278 --> 00:11:55.935 But what's getting really tricky here is the idea, because you're not blanking 00:11:55.935 --> 00:12:00.005 RAM, right? You're not cycling through things. You can hold stuff in memory, in theory. 00:12:00.405 --> 00:12:03.415 Like, we don't now. When you reboot the kernel, it's as if you rebooted. 00:12:03.415 --> 00:12:04.385 You get a whole fresh thing. 00:12:04.385 --> 00:12:05.235 Right. We'd played with, like, 00:12:05.235 --> 00:12:08.655 persistent RAM disks. We'd tried to kind of solve this with a hacky way. 00:12:08.655 --> 00:12:11.955 Yes, right. You can, like, tell the kernel to mark memory as sort of, 00:12:11.955 --> 00:12:14.205 like, don't touch this. It's supposed to be a fake, sort of, 00:12:14.525 --> 00:12:15.931 as if you had a persistent RAM. 00:12:16.233 --> 00:12:20.775 But at a more fine-grained level, what we now have with the live update orchestration 00:12:20.775 --> 00:12:25.795 and the kernel handover stuff is the idea of you can basically declare memfds, 00:12:25.795 --> 00:12:29.535 which are like memory file descriptors, which is basically just claiming a chunk of memory, right? 00:12:29.535 --> 00:12:31.773 But it's interfaced as a file descriptor. 00:12:32.441 --> 00:12:37.105 And so if an application gets one of those and then serializes whatever state 00:12:37.105 --> 00:12:39.295 it wants, right? So whatever stuff is happening for the application, 00:12:39.365 --> 00:12:41.147 it's running all of its runtime state. 00:12:41.339 --> 00:12:46.115 If it dumps that in this memfd or a few or whatever, it can pass that off to 00:12:46.115 --> 00:12:51.364 systemd and systemd will use the kernel's new orchestration and update features 00:12:51.706 --> 00:12:53.445 to save that across a case. 00:12:53.445 --> 00:12:54.325 Like a handoff thing. 00:12:54.325 --> 00:12:57.140 It's a state handoff so it stores it in memory and then it tells the kernel 00:12:57.372 --> 00:13:01.505 hey when we reboot you need to read like keep the that these bits of memory 00:13:01.505 --> 00:13:03.565 untouched and then reclaim them and restore them, 00:13:04.045 --> 00:13:09.835 and then systemd picks up after that and then as systemd goes it does the same 00:13:10.143 --> 00:13:13.916 and then it's able to hand the memfds back to the service when it boots, 00:13:14.549 --> 00:13:18.926 and then the service is able to if it has support for this load that back in, 00:13:19.448 --> 00:13:21.895 and then just restore itself and it would be like you didn't reboot at all. 00:13:21.895 --> 00:13:24.425 That is so sexy that is neat. 00:13:24.785 --> 00:13:27.365 Now it's not doesn't support everything like you can't preserve like a network 00:13:27.365 --> 00:13:30.375 socket or stuff like that okay but there's a bunch of other caveats so network. 00:13:30.375 --> 00:13:31.415 Traffic would get dropped. 00:13:31.415 --> 00:13:37.295 Yeah for a bit but but the primitives that this is could enable is pretty exciting yeah everything. 00:13:37.295 --> 00:13:40.665 From like emergency infrastructure scaling oh man there's so many things you 00:13:40.665 --> 00:13:42.928 could do that live swapping to a new system. 00:13:44.026 --> 00:13:48.643 And continuing your work? Ho, ho, ho! Okay, do we want to talk at all about 00:13:48.643 --> 00:13:53.685 StorageCTL, a tool that sounds really cool, provides a command line interface 00:13:54.068 --> 00:13:56.268 for, I would imagine, storage resources. 00:13:56.657 --> 00:14:00.753 And this is sort of the SystemD mantra, right, of unifying a bunch of things under a common control? 00:14:00.753 --> 00:14:04.283 Yeah, there's more and more capability built into SystemD to talk about what 00:14:04.283 --> 00:14:07.073 storage is on the machine, expose it to you, have APIs around it, 00:14:07.073 --> 00:14:09.306 and I think, like, format and work with the disk. 00:14:09.666 --> 00:14:14.583 Because it kind of comes also with SystemD sysinstall, which is a simple, 00:14:14.583 --> 00:14:17.173 modern, text-based operating system installer. 00:14:17.173 --> 00:14:18.723 Oh my god, are you serious? 00:14:18.723 --> 00:14:22.263 Yep, it builds on Systemd's existing partitioning, credential management, 00:14:22.263 --> 00:14:23.906 and system management features. 00:14:24.208 --> 00:14:27.807 It can copy an OS from temporary boot media like a USB drive. 00:14:28.312 --> 00:14:32.893 So now Systemd is able to sort of go from, you know, nothing is on your system 00:14:32.893 --> 00:14:36.463 to it's formatted and partitioned things and also installed. 00:14:36.463 --> 00:14:41.070 Again, clearly useful in cloud deployments. Clearly useful in cloud deployments. 00:14:41.830 --> 00:14:45.563 i have to wonder though if you wanted to roll your own distro say like a hypervibe 00:14:45.563 --> 00:14:50.693 kind of thing could you could you use something like this um. 00:14:50.693 --> 00:14:54.103 Yeah i mean more and more there's just systemd primitives you can build on instead 00:14:54.103 --> 00:14:55.228 of having to roll it yourself. 00:14:55.339 --> 00:14:58.572 Wow so systemd sys install, 00:14:59.646 --> 00:15:02.891 A modern text-based operating system installer. I kind of want to play with that. 00:15:02.891 --> 00:15:04.111 Yeah, me too. I think we should. 00:15:04.111 --> 00:15:07.221 So you have to have systemd 261. I mean, this is a big release. 00:15:07.221 --> 00:15:12.649 There's a lot in here. There's also some hardening stuff. Something called restrict file system access. 00:15:13.125 --> 00:15:17.951 It uses a BPF program to restrict execution to binaries stored on signed and 00:15:17.951 --> 00:15:21.111 verified DM Verity protected file systems. 00:15:21.411 --> 00:15:25.211 so if you really want to be super strict that like only the binaries I've explicitly 00:15:25.211 --> 00:15:30.267 signed and told can run on this system another way to make that happen in a secure way with systemd. 00:15:30.650 --> 00:15:34.281 So maybe a way for a really lean system or a container that doesn't need the 00:15:34.281 --> 00:15:39.711 overhead of sc linux or app armor but to still have like just only essentially an allow list. 00:15:41.242 --> 00:15:45.191 And you've seen this right like a lot of the we've seen things like the boot 00:15:45.191 --> 00:15:51.171 c systems more and more are you trying to use systems that do have signed file 00:15:51.171 --> 00:15:53.131 systems do use DM Verity under the hood. 00:15:53.131 --> 00:15:56.370 So you could combine that with some of those technologies and have a very secure 00:15:56.684 --> 00:16:02.105 sort of setup that has fast delivery but can play nicely in a modern enterprise. 00:16:04.451 --> 00:16:07.742 Very impressive. And I suppose also noteworthy boys, right? 00:16:07.992 --> 00:16:12.351 Because we now know that Lenart Pottering, you know, original creator SystemD, 00:16:12.822 --> 00:16:16.717 he was working at Microsoft for a bit. He left in January of this year. 00:16:17.431 --> 00:16:20.591 And he co-founded a Linux security startup and this is the first major system 00:16:20.591 --> 00:16:23.386 D release since he's made that transition and, 00:16:25.000 --> 00:16:27.960 Do you think you see a reflection in some of these releases? 00:16:27.960 --> 00:16:30.305 I mean, I think there might be a bit of a theme. 00:16:30.479 --> 00:16:35.660 Yeah. Yeah. I mean, a pretty solid release for cloud deployments that are just 00:16:35.660 --> 00:16:38.200 trying to spin up Linux systems as secure and fast as possible. 00:16:38.200 --> 00:16:40.710 But I still think there's a lot of primitives that people that are building 00:16:40.710 --> 00:16:42.472 distributions out there could use as well. 00:16:43.870 --> 00:16:47.030 It's always interesting, too, how they have thought of a new thing to take over 00:16:47.030 --> 00:16:52.195 that I have never even considered. I didn't think about building a Linux installer into SystemD. 00:16:52.840 --> 00:16:57.050 It's inspired. I wonder if they, I wonder if the chat GPT come up with that 00:16:57.050 --> 00:16:59.660 idea, like, Hey, chat GPT, I want some new features. And that's like, 00:16:59.660 --> 00:17:00.950 chat GPT is like, this is a great idea. 00:17:01.630 --> 00:17:05.270 We should build a Samba server in too. But you let me know if they ever build a Samba server in. 00:17:05.590 --> 00:17:06.790 I will, I will keep an eye out. 00:17:07.370 --> 00:17:11.520 That gets me thinking about some of the criticisms people have with system D 00:17:11.520 --> 00:17:14.770 is like it just does too much and it's too complex, etc. 00:17:15.290 --> 00:17:19.680 Do you think this is just heading more and more down that path? 00:17:19.680 --> 00:17:23.792 Because it seems soon we're almost going to have just system D be the OS. 00:17:25.115 --> 00:17:26.654 What do you think? What do you think there, Wes? 00:17:29.230 --> 00:17:34.240 Um well it kind of depends on how well you think it uh does these things and how much it's, 00:17:35.406 --> 00:17:40.310 open protocol something you could have other things implement and play with 00:17:40.630 --> 00:17:45.330 um there's a lot of varlink push in this release for instance so whatever you 00:17:45.330 --> 00:17:47.138 think of varlink might play into that, 00:17:47.730 --> 00:17:51.430 you know like is it a kernel feature that systemd is just using that other systems 00:17:51.430 --> 00:17:53.286 could take advantage of or plug into? 00:17:54.203 --> 00:17:55.538 Those would be the questions I'd ask. 00:17:56.252 --> 00:17:59.770 Yeah, I think the other question is if you were to build an operating system 00:17:59.770 --> 00:18:01.250 today, would it have this feature? 00:18:02.667 --> 00:18:04.292 Hmm. Hmm. That's a good way to look at it. 00:18:05.656 --> 00:18:11.212 And if you were, I think, honestly, a lot of these features come because these, 00:18:11.978 --> 00:18:15.652 are getting deployed in these massive cloud infrastructures and the cloud providers 00:18:16.041 --> 00:18:18.852 at talks and through emails and through private connections. 00:18:18.852 --> 00:18:21.892 They just complain about the things that take a long time to do and spin up. 00:18:21.892 --> 00:18:28.792 Like, my mind was blown when we went to Anisha's talk from Anthropic about spending 00:18:28.792 --> 00:18:33.582 up 40,000 containers at once and how, you know, something that takes 35 milliseconds 00:18:33.582 --> 00:18:34.502 ends up adding up big time. 00:18:34.522 --> 00:18:37.162 Like, just these crazy optimizations that they have to do. 00:18:37.642 --> 00:18:41.472 There's also just, you could phrase it, like, how much, oh, there was one thing 00:18:41.472 --> 00:18:43.805 I should mention, too, let's say. 00:18:44.090 --> 00:18:46.292 To speak to that, there's something called condition fraction. 00:18:46.292 --> 00:18:46.792 Oh. 00:18:46.792 --> 00:18:50.353 Which lets you do canary deployments at the service manager level. 00:18:50.632 --> 00:18:53.782 you can roll out to a percentage of machines by hashing machine id. 00:18:54.582 --> 00:18:58.842 Oh this is great for your remember remember forever ago you're asking how to 00:18:58.842 --> 00:19:03.132 do staged rollouts brent at uh your brother's store there you go wow. 00:19:03.132 --> 00:19:03.442 All right. 00:19:03.442 --> 00:19:04.053 Solved it for you. 00:19:04.187 --> 00:19:06.985 But the other thing do you remember the whole xz uh vulnerability. 00:19:07.112 --> 00:19:08.592 Nope forgotten all yeah never. 00:19:08.592 --> 00:19:13.722 Heard right yeah so so that involved instead of having something like plugged 00:19:13.722 --> 00:19:17.102 into the elf metadata saying like oh system d needs uh, 00:19:17.945 --> 00:19:24.665 xz it was done as a DL open which is like a dynamic way to open a shared library right and, 00:19:25.252 --> 00:19:30.262 as a somewhat indirect result of that system D wanted to have a system to be 00:19:30.262 --> 00:19:33.882 able to expose that as metadata so like as part of the build process it would 00:19:33.882 --> 00:19:36.171 tell you in the metadata and it's a way you could read, 00:19:36.641 --> 00:19:40.472 like what do I do dynamically what shared libraries do I depend on with DL open 00:19:40.472 --> 00:19:45.462 that you can't see from the regular elf metadata and that is now shipping as well well, 00:19:46.748 --> 00:19:49.322 So it's a nice little bow on sort of responding to that, 00:19:50.695 --> 00:19:52.651 terrible little supply chain scare. 00:19:52.872 --> 00:19:58.302 Well, we have been waiting to see what the first quote unquote AI features are 00:19:58.302 --> 00:19:59.606 that come to the Ubuntu desktop. 00:20:00.576 --> 00:20:03.920 And we recently got word that it's not going to be quite what you expect. 00:20:03.920 --> 00:20:07.147 It's going to be more of an accessibility thing, and it's going to be text-to-speech. 00:20:07.705 --> 00:20:12.099 And this week, we have a much clearer picture of that. They've launched a new project. 00:20:12.645 --> 00:20:16.435 They say the project will be desktop-focused, local speech-to-text dictation, 00:20:17.045 --> 00:20:19.878 with the first target release being Ubuntu 26.10. 00:20:20.429 --> 00:20:23.428 The initial interface is intentionally simple. 00:20:23.922 --> 00:20:28.010 You press a keyboard shortcut, you speak, and then the text gets inserted into 00:20:28.010 --> 00:20:32.072 the currently focused application with visual feedback while dictation is active. 00:20:32.583 --> 00:20:35.357 The first implementation targets the Ubuntu desktop on Wayland, 00:20:35.729 --> 00:20:37.796 with GNOME as the primary validated environment. 00:20:38.231 --> 00:20:41.420 Voice assistance, voice commands, desktop control, translation, 00:20:41.960 --> 00:20:45.040 and automatic language detection are outside the initial scope. 00:20:46.712 --> 00:20:51.193 So they, I think, are intentionally couching this as limited in scope. 00:20:51.675 --> 00:20:56.250 They have provided diagrams that show the architecture overview about how it works locally. 00:20:56.946 --> 00:21:00.040 um how are we feeling about this first go at it. 00:21:00.429 --> 00:21:04.760 Well my initial response is this is the way you ship an ai feature you'd be 00:21:04.760 --> 00:21:08.510 totally open about what it's doing what's under the hood you know firefox kind of, 00:21:09.380 --> 00:21:13.650 botched that a little while ago with their first features but uh they they do 00:21:13.650 --> 00:21:17.750 describe you know under the hood it uses speech recognition models running locally 00:21:17.750 --> 00:21:19.200 on your machine so i think that 00:21:19.200 --> 00:21:23.723 will make people happy or at least less hesitant maybe of these features, 00:21:25.192 --> 00:21:25.743 So 00:21:27.083 --> 00:21:30.479 communicating lots, I think is great. That's what we're seeing here. 00:21:30.810 --> 00:21:36.722 And also starting slow with some features that are pretty obvious if they work 00:21:36.722 --> 00:21:40.092 or not and are pretty simple. That's a good way to gain some trust. 00:21:40.412 --> 00:21:42.532 So I'm liking this measured approach. 00:21:43.581 --> 00:21:46.422 Right, Wes, this has to be probably the least controversial implementation of 00:21:46.422 --> 00:21:47.652 the technology possible, right? 00:21:47.652 --> 00:21:50.952 Indeed. And I would say so far, it seems like the reaction from what a little 00:21:50.952 --> 00:21:54.646 I've seen anyway, but it seems overall positive, I think. 00:21:55.192 --> 00:21:58.602 I'm taking a look here. Here they have a sort of a test branch where you can 00:21:58.602 --> 00:21:59.922 see a little of the development. Yeah. 00:22:00.253 --> 00:22:04.662 And just to be clear, it is named after the minor bird because it's renowned 00:22:04.662 --> 00:22:09.372 for its ability to listen to, mimic, and reproduce human speech with astonishing clarity. 00:22:09.812 --> 00:22:13.852 Just like its avian counterpart, this application is designed to master voice 00:22:13.852 --> 00:22:18.196 audio listening intently to your spoken words, instantly translating them to accurate, clean text. 00:22:18.892 --> 00:22:20.832 It's a Python app. It uses UV. 00:22:21.545 --> 00:22:26.171 They use hypothesis for generative testing, which is great. Love to see that. It looks very nice. 00:22:26.832 --> 00:22:31.172 Right now they have Whisper with Faster Whisper as one option, 00:22:31.172 --> 00:22:33.631 and then also Nemetron, which can do CUDA. 00:22:33.944 --> 00:22:37.746 So it sounds like they kind of got stuff for CPU and GPU going on here. 00:22:38.536 --> 00:22:43.272 Not a ton of stuff yet, but you've got like a desktop interface, 00:22:43.272 --> 00:22:47.592 it looks like. There's obviously Snap support. They're pulling in Whisper and Nemetron from Snaps. 00:22:47.888 --> 00:22:52.782 There's also a back-end server, standalone Ubu STT server. That's the main thing 00:22:52.782 --> 00:22:55.347 to Snap. This standalone Snap is going to ship, I guess. 00:22:56.369 --> 00:22:58.792 So this will be interesting from a couple of different vectors, 00:22:58.792 --> 00:23:02.112 I think. Number one, from like a user experience, what kind of load does this 00:23:02.112 --> 00:23:04.832 put on your desktop? What kind of memory usage are we looking at? 00:23:05.349 --> 00:23:08.432 And what does it set sort of the minimum requirements to have a decent experience 00:23:08.432 --> 00:23:10.783 here? So that's, I think, one vector to consider this under. 00:23:11.526 --> 00:23:17.782 The other is just, I think, what a hole there is in the accessibility space 00:23:17.782 --> 00:23:19.001 for the Linux desktop right now. 00:23:19.529 --> 00:23:25.952 And how this is table stakes on Android, iOS, I assume Windows, 00:23:25.952 --> 00:23:27.652 I don't actually know, and macOS. 00:23:29.363 --> 00:23:29.996 Table stakes. 00:23:31.280 --> 00:23:36.168 And it's like minimum bar accessibility. And we had this email from listener 00:23:36.168 --> 00:23:39.808 Matthew, and he said, Hey, Linux Unplugged hosts, I'm a blind Arch Linux user, 00:23:39.808 --> 00:23:41.154 and I really enjoy your content. 00:23:41.676 --> 00:23:44.318 I really feel like Linux accessibility is understated. 00:23:45.061 --> 00:23:49.578 Many desktop environments have little to no support for usable screen or usable 00:23:49.578 --> 00:23:52.288 screen reader experience, especially Wayland-based sessions. 00:23:52.549 --> 00:23:55.794 I'm comfortable with Monte, though I believe more people need to know about this. 00:23:56.079 --> 00:23:58.558 It's true that blind people themselves have made their own projects, 00:23:58.558 --> 00:24:01.688 thanks to open source, specifically for these individuals, but it shouldn't 00:24:01.688 --> 00:24:03.364 be like that. Thanks for reading my post. 00:24:04.148 --> 00:24:07.428 And we're thinking about, you followed up with Matthew a bit, 00:24:07.428 --> 00:24:12.478 but we're thinking like, this is probably the best implementation that Canonical 00:24:12.478 --> 00:24:16.318 could do that actually adds value to people in this space that could really 00:24:16.318 --> 00:24:18.124 use accessibility tools like this on Linux. 00:24:18.698 --> 00:24:22.768 And to Matthew's point, we don't really have much in GNOME and Wayland right now. 00:24:23.731 --> 00:24:25.968 And this is the exact area that Canonical is going to go after. 00:24:25.968 --> 00:24:30.420 Now, it's not a screen reader, but it's getting us to part of the way there you know and. 00:24:30.988 --> 00:24:35.848 It's yeah it's it's great to see i'm excited to see what happens uh it seems 00:24:35.848 --> 00:24:37.468 you know they're making some good progress and 00:24:37.828 --> 00:24:40.608 it seems like especially as john mentioned he's been on the show that you know 00:24:40.608 --> 00:24:43.336 they're giving pretty regular updates with the work that they're doing so we'll 00:24:43.336 --> 00:24:47.097 there'll probably be more to come i do also see it's gpl3 so that's pretty cool. 00:24:48.089 --> 00:24:51.236 I have a little bit of a pondering here. And I think, Chris, 00:24:51.236 --> 00:24:54.551 this is in your expertise. 00:24:55.496 --> 00:24:59.216 Why? Like, they built their own thing here. And it sounds like Wes, 00:24:59.216 --> 00:25:01.307 you know, gives it the little approval. 00:25:02.178 --> 00:25:06.246 But why not use something like what Home Assistant has been working on with 00:25:06.246 --> 00:25:12.274 their voice stuff and voice input? Do you think that that's a missed opportunity? 00:25:13.836 --> 00:25:17.333 Whisper is being used by both projects. So some of the primitives are the same. 00:25:18.175 --> 00:25:21.926 Uh, so that's good. I, I think, you know, I could see using the bits that do 00:25:21.926 --> 00:25:24.630 make sense and then building a custom layer for the other parts, 00:25:25.309 --> 00:25:27.045 um, the interface with the desktop. 00:25:27.887 --> 00:25:32.086 I think there is like a lot of people that don't know that you have whisper 00:25:32.086 --> 00:25:35.306 and Piper happening over there at the home assistant community. 00:25:35.306 --> 00:25:39.067 That is really powerful for text to speech and speech to text and all of that. 00:25:39.613 --> 00:25:43.450 But basing it on whisper is going to be, I think they're going to have decent results. 00:25:43.984 --> 00:25:46.686 Fast whisper usually does decent results if you have clean audio, 00:25:46.686 --> 00:25:50.366 but if people are in a noisy work environment, it's going to be pretty hit and 00:25:50.366 --> 00:25:51.978 miss. So it's going to really depend on... 00:25:52.721 --> 00:25:55.746 And I think that open... I haven't dug through the code base, 00:25:55.746 --> 00:25:58.496 but I imagine there's an interface that they have so that they're intending... 00:25:58.496 --> 00:26:02.256 Like, they already support too, so I imagine, like, as other options are available 00:26:02.256 --> 00:26:04.818 or whatever, like, you'll be able to extend with different ASRs. 00:26:05.616 --> 00:26:09.626 Or, I mean, is this some sort of thing that app developers could build on top 00:26:09.626 --> 00:26:13.996 of and maybe create a screen reader using the backend that Canonical is providing here? 00:26:13.996 --> 00:26:14.556 That's a good question. 00:26:15.330 --> 00:26:18.486 I think we'll have to just find out if people experiment and build something 00:26:18.486 --> 00:26:20.556 with it, right? That'll be the question. 00:26:20.556 --> 00:26:21.516 That's the open source way. 00:26:22.236 --> 00:26:23.991 I guess we'll check it out, right? Because we're supposed to get it, 00:26:26.496 --> 00:26:28.652 in the next release, you know? Isn't that what they're targeting? 00:26:28.902 --> 00:26:29.111 Yep. 00:26:29.349 --> 00:26:31.806 I know I wrote it down, but I'm an old man and I can't remember. 00:26:31.806 --> 00:26:34.933 But yeah, the next release. Yeah, 2610. We don't have to wait that long to try it out. 00:26:38.805 --> 00:26:44.900 Webroot.com slash unplugged. Webroot is a cloud-based antivirus engineered to stay out of your way. 00:26:45.162 --> 00:26:48.401 It takes up to 33 times less space than the bulky competitors, 00:26:48.621 --> 00:26:50.508 and it scans six times faster. 00:26:50.885 --> 00:26:54.914 So instead of waiting around, you're getting instant powerful protection without the leg. 00:26:55.344 --> 00:26:59.083 And I know more of us than ever, we're trying to make our PCs last longer. 00:26:59.083 --> 00:27:00.771 You've probably got family members as well. 00:27:01.224 --> 00:27:04.196 They have a perfectly good PC. They want to use it a little bit longer. 00:27:04.481 --> 00:27:07.383 Maybe it's feeling a little slow. Maybe it's feeling a little sluggish. 00:27:07.383 --> 00:27:09.583 And they need good online protection. 00:27:10.273 --> 00:27:14.830 Webroot can keep their PC running fast and safe because the intelligence, it lives in the cloud. 00:27:15.445 --> 00:27:19.093 See, most people are blaming their device for the slowdown. But the real slowdown 00:27:19.093 --> 00:27:21.651 is coming from that old antivirus software. 00:27:22.069 --> 00:27:24.473 Those traditional programs, they just, they hog up the storage. 00:27:24.473 --> 00:27:27.563 They soak up the disk IO and they eat a lot of RAM. And they do, 00:27:27.563 --> 00:27:29.702 they make a machine feel old and slow. 00:27:30.648 --> 00:27:34.613 And it's not even a feeling. You can see it in the numbers. WebRoot Essential 00:27:34.613 --> 00:27:36.117 versus, say, Norton Antivirus? 00:27:36.894 --> 00:27:42.014 Well, the numbers say it all. WebRoot scans 3.7 times faster. 00:27:42.792 --> 00:27:45.968 It installs 35 times faster. But here's the big one. 00:27:46.548 --> 00:27:51.244 This is the one. It uses five times less RAM when idle. 00:27:51.564 --> 00:27:55.534 That's going to make a real-world difference on their PC. It's going to make a difference. 00:27:56.179 --> 00:27:59.667 You know, the other thing I'm thinking about is LLM-assisted phishing attacks. 00:28:00.480 --> 00:28:04.333 They're just getting better and better. Some of us can spot them, 00:28:04.333 --> 00:28:07.411 but not all of us, and especially some of those family members. 00:28:07.992 --> 00:28:10.063 This is why I'm recommending WebRoot Total Protection. 00:28:10.661 --> 00:28:14.103 Unlike the free tools, or really the older traditional tools. 00:28:15.153 --> 00:28:19.202 Bygone traditional tools. You know what I'm talking about. They weren't built 00:28:19.202 --> 00:28:23.344 for this, but WebRoot is, is designed to counter modern AI-driven attacks. 00:28:24.029 --> 00:28:28.336 It's fast, it's lightweight, and it's designed to spot these threats before they reach you. 00:28:28.842 --> 00:28:31.492 It literally doesn't get better than that, whether you're working, 00:28:31.492 --> 00:28:36.184 browsing, or streaming. WebRoot is the antivirus that works for you, not against you. 00:28:36.579 --> 00:28:41.252 Make the switch and feel the difference of truly fast modern antivirus protection. 00:28:41.502 --> 00:28:46.902 And for a limited time, you can save 60% on WebRoot when you go to webroot.com 00:28:46.902 --> 00:28:50.842 slash unplugged. Yeah, I said 60%, but you got to go to that URL. 00:28:50.842 --> 00:28:54.319 It's the only way to get the deal. You go to webroot.com slash unplugged. 00:28:54.708 --> 00:28:58.766 Great way to support the show too. That's webroot.com slash unplugged. 00:29:01.517 --> 00:29:10.037 As happens to us last week, as we are wrapping up the show, Linus announced a new kernel, Linux 7.1. 00:29:10.882 --> 00:29:11.532 Huzzah. 00:29:11.532 --> 00:29:15.562 Yeah, it's pretty exciting. Linus wrote, normally I try to front load the merge 00:29:15.562 --> 00:29:19.912 window and do as much possible the first few days, which we're going to come 00:29:19.912 --> 00:29:21.914 back to the 72. Oh my goodness. 00:29:22.506 --> 00:29:26.712 Oh, this is so funny too, because spoiler alert, uh, it didn't go this way. 00:29:26.712 --> 00:29:28.996 He says this time, I'm not sure it's going to work out with my laptop. 00:29:29.658 --> 00:29:32.582 He didn't think, wow, a couple of long flights without internet, 00:29:32.582 --> 00:29:35.222 but I've made sure that I have fetched the early pull request. 00:29:35.222 --> 00:29:36.252 Thank you. You know who you are. 00:29:36.502 --> 00:29:40.772 So I'll be able to do some of it offline. Well, I guess he worked the entire flight. 00:29:40.772 --> 00:29:41.642 Oh my goodness. 00:29:42.382 --> 00:29:47.020 Anyway, possible slight hiccups in the merge window aside, the news today is 7.1. 00:29:47.380 --> 00:29:50.562 Below is the short log for the last week. Nothing particularly interesting or 00:29:50.562 --> 00:29:52.866 scary stands out, which is as it should be. 00:29:53.232 --> 00:29:56.572 It's mostly various smaller driver updates, GPU networking, sound miscellaneous 00:29:56.572 --> 00:30:01.039 with some networking and trace tooling fixes and random minor changes elsewhere. 00:30:02.531 --> 00:30:10.308 Well, one of those random minor changes is this major rework to the NTFS driver 00:30:10.308 --> 00:30:12.738 implementation, which now has full write support. 00:30:13.898 --> 00:30:19.658 That is correct. And Linus merged it with the commit message NTFS resurrection. 00:30:19.658 --> 00:30:19.978 Yeah. 00:30:19.978 --> 00:30:20.078 Wow. 00:30:20.398 --> 00:30:24.752 And that's because it's – so, okay, let's put the – This is a long saga. It is. 00:30:24.891 --> 00:30:26.828 We should try to maybe do a short version. 00:30:26.828 --> 00:30:29.818 Yeah, okay, so there's a lot of different ways to talk NTFS, 00:30:29.818 --> 00:30:32.519 right, which is the sort of Windows file system. 00:30:33.860 --> 00:30:39.363 We've had NTFS 3G, which is like a fuse option, which works well but slow. 00:30:39.699 --> 00:30:40.688 And is in user space. 00:30:40.688 --> 00:30:43.688 Yeah, there has been the traditional NTFS driver in the kernel, 00:30:43.688 --> 00:30:45.638 which has been there a long time, but it's been read-only. 00:30:46.678 --> 00:30:51.898 Then in 2021, we got NTFS 3, which was from Paragon, which was previously out 00:30:51.898 --> 00:30:53.618 of tree, but then they got it merged. 00:30:53.618 --> 00:30:54.058 But then... 00:30:54.498 --> 00:30:57.288 It kind of has just been rotting in the tree, unfortunately. 00:30:57.288 --> 00:30:57.991 Yeah, they kind of left. 00:30:58.698 --> 00:31:05.716 And so now this is like a four-year rewrite from the ex-fat maintainer, 00:31:06.285 --> 00:31:11.185 Namjai Jian, and he took the bones of the original read-only NTFS driver, 00:31:11.457 --> 00:31:14.828 added full write support using modern kernel APIs like IOMAP, 00:31:14.828 --> 00:31:19.198 delayed allocation, folios, and now we're getting it, yeah. 00:31:19.198 --> 00:31:25.458 A real, true, modern NTFS implementation on Linux after this multi-year saga. 00:31:25.838 --> 00:31:29.008 There was a little hiccup. There was a kind of got pulled and then un-pulled 00:31:29.008 --> 00:31:31.428 because there was some Git structure issues with the commits, 00:31:31.428 --> 00:31:33.258 but that all got cleared up and obviously it's here now. 00:31:33.978 --> 00:31:37.620 I'm going to bet that next week Windows comes out with a new file system. 00:31:38.613 --> 00:31:42.328 I don't think so. I don't think so. It's been a while too. 00:31:42.665 --> 00:31:47.158 Also good to see the Apple Silicon MacBooks getting SMC power driver support 00:31:47.158 --> 00:31:50.008 for some of the systems, giving better battery and power metrics. 00:31:50.008 --> 00:31:54.525 Of course, some of that's changing with Mac OS 27 Golden Gate, but it's there. 00:31:55.379 --> 00:32:01.873 Also, did you see this? In 7.1, 7.1, they mainline support for 12 new SOCs and 00:32:01.873 --> 00:32:07.023 hardware platforms across ARM and RISC-V and added real-time kernel build support 00:32:07.023 --> 00:32:09.523 for 32-bit ARM systems out there in the field. 00:32:09.523 --> 00:32:10.283 Fascinating. 00:32:11.003 --> 00:32:14.853 We not only have real-time kernel support these days, but we have it on 32-bit ARM. 00:32:15.022 --> 00:32:16.473 Not 32x86, though. 00:32:16.473 --> 00:32:21.513 No. Well, who wants that? There's this small little lining here about Intel 00:32:21.513 --> 00:32:25.363 FRED being enabled by default for supported systems on Panther Lake, 00:32:25.363 --> 00:32:27.084 Diamond Rapids, and AMD. 00:32:29.163 --> 00:32:30.097 It's interesting... 00:32:30.974 --> 00:32:34.561 So this goes, this is like a 30-year upgrade in a way. I mean, 00:32:34.561 --> 00:32:38.114 this is already, this is text not new, but just sort of making it the default here is. 00:32:38.636 --> 00:32:43.001 So basically any time, since basically the 386, any time a user space program 00:32:43.001 --> 00:32:46.481 needs to talk to the kernel with like a system call or an interrupt or an exception 00:32:46.481 --> 00:32:50.235 or something like that, it goes through what's called the interrupt descriptor table. 00:32:50.682 --> 00:32:53.828 The CPU saves state, switches to ring zero, jumps to the handler. 00:32:54.240 --> 00:32:56.243 When done, it returns back to user space. 00:32:56.771 --> 00:33:01.628 This process involves multiple microcode steps, memory reads and serialization operations. 00:33:01.889 --> 00:33:07.421 It's been the same since like 1985. Wow. But with Fred, we're replacing that 00:33:07.421 --> 00:33:13.621 IDT delivery mechanism with new instructions that do the same sort of ring transitions, 00:33:13.621 --> 00:33:16.373 but they do them atomically. They do them with fewer steps. 00:33:16.883 --> 00:33:21.031 The CPU handles the context switch more directly, so you get fewer memory accesses, 00:33:21.031 --> 00:33:24.035 fewer serialization operations, and lower latency. 00:33:24.511 --> 00:33:28.581 So it's a kind of cool new thing that x86, like it's doing it both on Intel. 00:33:28.702 --> 00:33:33.347 It's called Intel Fred, but AMD's doing it as well. And we're finally starting to get access to it. 00:33:33.625 --> 00:33:37.701 I like the name. Larable and Pharonix notes, quote, great news for Panther Lake 00:33:37.701 --> 00:33:42.071 on Linux and pairs nicely with Panther Lake Intel Idle Driver C-State Editions 00:33:42.071 --> 00:33:46.661 for Linux 7.1 as well as ongoing Z-Kernel Linux driver development improvements. 00:33:47.626 --> 00:33:52.061 I feel like the Panther Lake generation of Intel chips might pull me back to 00:33:52.061 --> 00:33:54.772 the Intel side if I were in the market to get a laptop. 00:33:54.947 --> 00:33:56.181 We should at least get to play with one. 00:33:56.941 --> 00:34:00.448 The good news for Panther Lake on Linux just keeps stacking up. 00:34:00.710 --> 00:34:01.493 It's really impressive. 00:34:01.836 --> 00:34:06.221 Fred's also expected to be used by the upcoming AMD Zen 6 processor and Intel 00:34:06.221 --> 00:34:08.921 Xeon Diamond Rapid Server processor. 00:34:08.921 --> 00:34:09.341 Oh, boy. 00:34:10.201 --> 00:34:14.381 I like those names. I want one of those. Nice to see the Steam Deck OLED audio 00:34:14.381 --> 00:34:18.861 fixes getting now upstream, reducing the need for downstream patches in SteamOS, 00:34:18.861 --> 00:34:21.196 which is good to see. Oh, no. 00:34:22.600 --> 00:34:27.395 Oh, no. You should have warned me about this. PCM ICA support has been removed 00:34:27.395 --> 00:34:30.845 as well as more 486 era x86 support paths. 00:34:30.845 --> 00:34:31.125 Yeah. 00:34:31.865 --> 00:34:35.551 Well, I still have PCM ICA cards. PCM ICA cards. 00:34:35.894 --> 00:34:41.965 They were the best Wi-Fi. Look, look, when 802.11b came out, 00:34:41.965 --> 00:34:43.539 no laptop had that built in. 00:34:44.038 --> 00:34:48.136 But a guy could get himself a PCM CIA card and he could get himself Wi-Fi. 00:34:48.427 --> 00:34:51.125 And then if you needed to switch over, you could get double-wide ones with an 00:34:51.125 --> 00:34:53.047 Ethernet adapter, you could get ones with storage. 00:34:54.046 --> 00:34:57.135 I feel like it didn't get a chance to live up to its potential, 00:34:57.135 --> 00:34:58.091 and we shouldn't remove it yet. 00:34:58.259 --> 00:35:06.525 Okay, but the upside is, as part of that, we're dumping 140,000 lines of legacy code in the kernel. 00:35:08.348 --> 00:35:09.765 Okay, that sounds like a good thing. 00:35:10.165 --> 00:35:12.965 You can have this, but you can't take anything else for the rest of the episode from the kernel. 00:35:12.965 --> 00:35:18.215 All right, well, we do have some good news from 7.1. It included nearly 13,000 00:35:18.215 --> 00:35:23.775 non-merge change sets, contributions from more than 2,000 developers with hundreds 00:35:23.775 --> 00:35:24.965 of first-time contributions. 00:35:24.965 --> 00:35:26.575 That's cool. That's really cool. 00:35:26.575 --> 00:35:27.667 I think it was like 300 and some. 00:35:28.021 --> 00:35:31.845 Okay, so we opened up with teasing Linus about low-keying what the next merge 00:35:31.845 --> 00:35:35.235 window is going to look like for Linux 7.2, right? Because now 7.1's out the 00:35:35.235 --> 00:35:38.835 door, so work never stops for the kernel team. They immediately open up the 00:35:38.835 --> 00:35:40.003 merge window for the next kernel. 00:35:40.310 --> 00:35:43.219 And Linus is like, I'm going to be on a flight, going to be working on my battery. 00:35:43.381 --> 00:35:44.606 Maybe he has a panther, like, I don't know. 00:35:45.024 --> 00:35:49.883 because he sure did get the work done. So we have now seen the first merge window, 00:35:50.115 --> 00:35:55.015 the first half of it for Linux 7.2, and more than 7,000 non-merge change sets 00:35:55.015 --> 00:35:56.621 have already been pulled into mainline. 00:35:57.225 --> 00:36:01.495 We are seeing some code removal around I-486 again, but let's talk about some 00:36:01.495 --> 00:36:03.970 of the stuff that's landing that looks really interesting, 00:36:04.725 --> 00:36:09.815 including a new feature in Linux 7.2 that adds the ability to limit programs 00:36:09.815 --> 00:36:15.546 to only open, quote, regular files and avoid being tricked or doing, quote, silly things. 00:36:16.145 --> 00:36:22.471 So this is a brilliant feature that basically limits the type of file. Okay, all right. 00:36:23.719 --> 00:36:28.409 So this landing in Linux 7.2 is this open NAT2 underscore regular flag. 00:36:28.607 --> 00:36:32.654 Yeah, so there's already been the existing open, which is a system call for 00:36:32.654 --> 00:36:36.287 opening files. Open NAT2 is this new version of it. 00:36:36.426 --> 00:36:37.599 It doesn't allow silly files? 00:36:39.284 --> 00:36:40.244 I hate those silly files. 00:36:40.244 --> 00:36:42.884 It just gets everywhere. It's a sort of anti-silly system call, if you will. 00:36:42.944 --> 00:36:43.794 Okay, good, good, good. 00:36:43.794 --> 00:36:47.594 Well, okay, so open NAT2 is the modern version, sorry. Now we're getting open 00:36:47.594 --> 00:36:49.094 NAT2 underscore regular. 00:36:49.094 --> 00:36:50.393 Regular, right, regular. 00:36:50.613 --> 00:36:54.364 But it's part of open NAT2. It's basically because OpenI2 supports passing a 00:36:54.364 --> 00:36:57.944 struct of various things that it's this open how parameter. And so now we have 00:36:57.944 --> 00:37:01.845 more ways to tell it, you know, be regular. 00:37:02.652 --> 00:37:05.984 Yeah, the flag indicates the path should be open if it's a regular file. 00:37:05.984 --> 00:37:09.114 It's useful to write secure programs that want to avoid being tricked in opening 00:37:09.114 --> 00:37:14.314 device nodes with special semantics while thinking they operate on regular files. 00:37:14.314 --> 00:37:17.404 Exactly. So like there might be times, right, that you need to open a device 00:37:17.404 --> 00:37:20.862 node or a special file or something under proc or something like that, right? 00:37:21.007 --> 00:37:26.084 or a pipe, but that's usually a specific point in your program and it's not 00:37:26.084 --> 00:37:30.284 from generic user input for uploading a file into the browser or something like that. 00:37:30.551 --> 00:37:33.624 And so this enables you to opt into that behavior of saying, 00:37:33.624 --> 00:37:37.864 hey, I really want a regular file, which sounds funny, but it filters things, 00:37:38.684 --> 00:37:45.564 like a device node, so dev zero or dev random or a FIFO or a pipe or symlink 00:37:45.944 --> 00:37:50.664 substitution attacks here can kind of be detected as well. 00:37:50.664 --> 00:37:53.224 And this will play into container sandbox hardening. 00:37:54.169 --> 00:37:58.070 This is kind of a way to, you know, you could add this in as part of that. 00:37:59.080 --> 00:38:00.754 So it's a nice new, yeah, it's a nice new feature. 00:38:00.754 --> 00:38:04.439 I think it's mostly important as a flex. Like you can walk up to a Mac user and be like... 00:38:05.896 --> 00:38:10.262 My kernel only opens regular files. Does your kernel open up strange, 00:38:10.262 --> 00:38:13.802 silly files? Your kernel might open silly files. You should be concerned about that. 00:38:14.462 --> 00:38:16.930 You know, like that kind of thing. It's a flex now. I like that. 00:38:17.169 --> 00:38:21.743 I also like seeing that BTRFS, our buddy ButterFS, is now enabling large folios. 00:38:21.854 --> 00:38:24.142 Because I think, like, Brent has a ton of those. 00:38:24.142 --> 00:38:24.342 Oh, yeah. 00:38:24.362 --> 00:38:29.817 By default, huge folios are coming with Linux 7.2. Huge folios. 00:38:30.922 --> 00:38:35.272 A folio is a kernel data structure that represents one or more physically contiguous pages. 00:38:35.272 --> 00:38:38.502 And a huge one is two megabytes, apparently. 00:38:38.822 --> 00:38:43.762 Yeah, because the small ones were four kilobytes. So it's kind of huge, apparently. 00:38:43.762 --> 00:38:44.836 It's relative. It's relative. 00:38:46.118 --> 00:38:50.602 Okay, Butterfest also gains a new IOCTL for returning raw checksums to user 00:38:50.602 --> 00:38:54.992 space, a stable UUID for overlay OS-style use cases, and a lot of performance 00:38:54.992 --> 00:38:59.382 work, including reported sequential write and direct I.O. throughput improvements. 00:38:59.822 --> 00:39:00.472 Nice to see. 00:39:00.472 --> 00:39:03.492 And our classic buddy XFS has also seen some work. 00:39:03.492 --> 00:39:05.502 Yeah, it doesn't just stop at Butterfest. Heck no. 00:39:05.902 --> 00:39:10.822 No, no. New zone allocator is now out of experimental status in Linux 7.2. 00:39:10.822 --> 00:39:14.340 Reminder, you're not going to be getting 7.2 for a little bit, but that's good to see. 00:39:14.491 --> 00:39:19.842 Yeah, it's used for zoned storage devices, including those SMR hard drives and ZNS SSDs. 00:39:19.862 --> 00:39:20.192 Okay. 00:39:20.900 --> 00:39:26.562 Most of the rest is kind of just fixes and minor changes, but EXT4 got some cool stuff too. 00:39:27.702 --> 00:39:27.982 Yeah. 00:39:27.982 --> 00:39:31.132 It reworked fast commit handling to reduce locking contention, 00:39:31.132 --> 00:39:34.242 which is always good, and avoid deadlock scenarios, which is even better. 00:39:34.242 --> 00:39:37.092 Yeah, it's nice to see that. Because you really don't want your file system driver to deadlock. 00:39:37.092 --> 00:39:40.289 I mean, you can't claim that extended four is abandoned. It's nice to see that. 00:39:40.416 --> 00:39:43.992 No, there's like hybrid journaling path that's being used now for recording 00:39:43.992 --> 00:39:47.102 metadata deltas instead of always having to write full metadata every time. 00:39:47.102 --> 00:39:49.622 Should make things a little bit faster there when updating metadata updates. 00:39:49.622 --> 00:39:50.792 Just doing a little boom-boom. 00:39:50.792 --> 00:39:55.792 Yeah, and directory hash computation is optimized by processing input in four-byte 00:39:55.792 --> 00:40:01.022 chunks and removing function pointers, which is like a 2x improvement for at least longer inputs. 00:40:01.022 --> 00:40:03.182 Ugh, that'll finally get Brent to stop complaining. 00:40:04.035 --> 00:40:04.980 Big EXT4 guy. 00:40:04.980 --> 00:40:06.790 Yeah, he's been bitching about that for years. 00:40:06.790 --> 00:40:10.370 Yeah, yeah. Okay, well, how about this? We were talking about you don't want 00:40:10.370 --> 00:40:13.560 to open pipes when you don't think you need to be. I do not. 00:40:14.100 --> 00:40:19.120 But I think many of us run a lot of shell pipelines or use pipes as part of 00:40:19.540 --> 00:40:21.220 operating on a Unix-like operating system. 00:40:21.320 --> 00:40:22.050 Big pipe guy. 00:40:22.050 --> 00:40:27.310 Yeah, right? So Linux 7.2 is improving a non-pipe write, the kernel path used 00:40:27.310 --> 00:40:31.335 for anonymous pipes such as shell pipelines and standard streams. 00:40:31.962 --> 00:40:38.400 So this change avoids allocating pages, reducing contention between concurrent readers and writers. 00:40:38.400 --> 00:40:40.681 So if you're passing a whole bunch of data over that pipeline, 00:40:41.088 --> 00:40:45.012 this could get somewhere between 6% to 28% better throughput, 00:40:45.568 --> 00:40:49.590 and somewhere between 5% and 22% lower average write latency. 00:40:49.590 --> 00:40:50.210 That is significant. 00:40:50.620 --> 00:40:52.730 And it gets better when you're under memory pressure. 00:40:52.730 --> 00:40:56.069 That's the thing right there. That's really cool. Because that's always when I'm piping. 00:40:56.551 --> 00:40:59.610 I'm always piping under memory pressure because that's the secret, 00:40:59.610 --> 00:41:01.671 boys. I'm always under memory pressure. 00:41:02.507 --> 00:41:05.270 We also have some changes in the crypto driver code. I just want to mention 00:41:05.270 --> 00:41:09.368 that they have been deprecating for a period of time bits of this interface, 00:41:09.815 --> 00:41:14.367 that allowed user space applications to call into the crypto API directly in the kernel. 00:41:15.227 --> 00:41:20.146 And not crypto in the currency, but crypto in the sense that it used to mean, as in cryptography. 00:41:20.156 --> 00:41:20.766 Yeah, that's right. 00:41:20.766 --> 00:41:24.296 Remember the good old days? But I guess they've been deprecating this. 00:41:24.296 --> 00:41:26.056 It must be because it's insecure, I would assume. 00:41:26.056 --> 00:41:27.417 Yeah, so you remember copy fail? 00:41:28.476 --> 00:41:30.926 No. No, I've completely forgotten that. 00:41:30.926 --> 00:41:34.366 Yeah, well, that reason, somewhat reason. 00:41:34.366 --> 00:41:35.806 So this is some follow-up from that, huh? 00:41:35.806 --> 00:41:39.176 Uh-huh, vulnerability. Yeah. Yeah, this was used in part of that attack. 00:41:39.816 --> 00:41:43.286 because it has optimizations that kind of let you overwrite memory in a way 00:41:43.286 --> 00:41:47.096 that one area was treating it as read-only, but now suddenly it's getting used 00:41:47.096 --> 00:41:49.516 letters right, and you've been able to change it. 00:41:52.716 --> 00:41:56.366 Essentially, no one is maintaining it, and these optimizations just aren't really 00:41:56.366 --> 00:42:00.646 worth it anymore, and especially now that we're in the era of a whole bunch 00:42:00.646 --> 00:42:03.116 of sustained LLM attacks. 00:42:03.416 --> 00:42:07.516 Yep, exactly. Okay, all right, that makes sense. 00:42:07.516 --> 00:42:10.066 Really, there's kind of a lot of that. Like if you look under the hood, 00:42:10.066 --> 00:42:16.896 a lot of the things being pulled or kind of reconsidered what the value is has stemmed from that. 00:42:16.896 --> 00:42:19.036 Yeah, it's interesting to see the fallout just kind of continue. 00:42:19.036 --> 00:42:22.156 I said you could not take anything else away from the kernel in this episode, 00:42:22.156 --> 00:42:23.301 but that one I'm okay with. 00:42:23.875 --> 00:42:25.930 That one I feel like makes sense. We need to be safe. 00:42:26.853 --> 00:42:29.196 All right, though, let's just see what we have next here on the list here. 00:42:29.636 --> 00:42:32.310 The next feature in Linux 7.2 is they are... 00:42:34.022 --> 00:42:36.806 Sorry. Let me finish that. 00:42:36.806 --> 00:42:38.596 This one was I really didn't want to think about. 00:42:39.241 --> 00:42:45.119 It reads here, removing Apple talk protocol support from the mainline kernel. Sorry, Chris. 00:42:45.526 --> 00:42:48.545 Why? Why do we have to remove? What? 00:42:50.854 --> 00:42:54.604 Why? Why? It doesn't hurt anybody. It's a great little protocol for those of 00:42:54.604 --> 00:42:57.187 us who are enthusiasts that like to have classic machines. 00:42:58.624 --> 00:42:58.744 Why? 00:42:59.124 --> 00:43:02.704 Why? What is this hurting? It's 4,000 lines of code. It's nothing. 00:43:02.704 --> 00:43:06.274 It's nothing in the Linux kernel. I just can't stand for this. 00:43:06.274 --> 00:43:10.167 Do you, are you going to put some effort into building yourself a fork that maintains this? 00:43:11.049 --> 00:43:14.994 No, but what I have thought about doing is, you know, creating a consistent 00:43:14.994 --> 00:43:16.364 complaining campaign about it, 00:43:16.364 --> 00:43:19.654 you know, just consistently getting on a soapbox. you think that's a brand. 00:43:19.654 --> 00:43:20.494 New segment on the show. 00:43:20.494 --> 00:43:24.224 Huh yeah yeah do you think that might help if i just complain a lot i i get 00:43:24.224 --> 00:43:31.134 it but it it is it is sad to see the removal was triggered they say partially by a burst of, 00:43:33.160 --> 00:43:38.354 ai generated fixes that was rarely used code that lacked active reviewers so 00:43:38.354 --> 00:43:44.884 you're telling me it's a victim of slop trimming yep so this is the problem people I. 00:43:44.884 --> 00:43:45.984 Had forgotten about it and the. 00:43:47.024 --> 00:43:51.439 Slop brought it up. The AI can't consider the art and history, 00:43:52.356 --> 00:43:56.182 the full context around Apple talk, right? This isn't IPX. 00:43:56.722 --> 00:44:01.354 Okay. But this isn't like some random Bluetooth 2.0 protocol we don't use anymore 00:44:01.354 --> 00:44:03.049 for like low bit rate headpieces. 00:44:03.171 --> 00:44:08.070 But Chris, the kernel is not a museum. Like we don't need these things in there if nobody's using it. 00:44:08.697 --> 00:44:14.114 First of all, shut your mouth. Second of all, listen, man, I got old Macs somewhere 00:44:14.114 --> 00:44:18.394 around here. i'd love to fire up one day i don't know like if i can't use linux 00:44:18.394 --> 00:44:21.805 and i can't use maples what am i gonna do. 00:44:22.711 --> 00:44:29.314 Oh it does look like they will be maintaining it still in this uh linux net 00:44:29.314 --> 00:44:35.470 dev out mod orphan module really yeah that's what i see here in the patch for removal, 00:44:37.026 --> 00:44:40.122 There's also some AX25 and ham radio that are over there. 00:44:40.122 --> 00:44:42.191 Oh, the ham guys are getting it too? 00:44:42.673 --> 00:44:46.533 Well, they might have already been there. I don't know much about the ham radio driver. 00:44:47.832 --> 00:44:53.812 It does also say that Apple Talk was removed in OSX in 10-6 Snow Leopard in 2009. 00:44:53.812 --> 00:44:58.232 Yeah, I know. It sucked then, doesn't it? And just because they do something wrong. 00:44:58.232 --> 00:44:59.212 You've only gotten over that one. 00:44:59.232 --> 00:45:02.392 Two wrongs don't make a right, Wes. You know that. You know that. 00:45:03.132 --> 00:45:07.712 So we can collectively cope. Before the show, I created Apple Talk Rest in Peace 00:45:07.712 --> 00:45:11.222 sticker, the memorial sticker where we as a community, for those of us that 00:45:11.222 --> 00:45:15.117 remember, can put it on our machines 1985 to 2026. 00:45:15.831 --> 00:45:19.598 It just talked different. Stickers available right now on jupitergarage.com. 00:45:20.144 --> 00:45:24.412 And while you're at it, why not pick up Sorry, I Only Open Regular Files so 00:45:24.412 --> 00:45:31.208 you can flex and celebrate Linux 7.2, the can't-do attitude of saying no to special files. 00:45:31.731 --> 00:45:36.061 Sorry, I only open. Real stickers available now, jupitergarage.com. Go check them out. 00:45:36.414 --> 00:45:40.692 Sorry, I only open regular files and the Apple Talk Memorial sticker. 00:45:40.692 --> 00:45:42.382 I am putting them on my rigs. 00:45:43.746 --> 00:45:47.242 I'll tell you what. It's too soon. It's too soon for both of those things. 00:45:47.242 --> 00:45:49.156 Too soon. I'll tell you what. 00:45:50.092 --> 00:45:54.402 If I could have my way, we would have a little network of Apple Talk going, 00:45:54.402 --> 00:45:58.496 and you would just use it for, like, device discovery and file transfer. 00:45:59.059 --> 00:46:03.142 You don't need TCP IP for that. Like, everybody's doing all these mesh networks 00:46:03.142 --> 00:46:07.802 and overlay networks. Well, if you're just on a LAN, why not use TCP IP for 00:46:07.802 --> 00:46:12.632 everything you use TCP IP for, but then use something like AppleTalk for your local file sharing? 00:46:14.772 --> 00:46:16.512 Guys, you're laughing. There's a real use case here. 00:46:16.512 --> 00:46:17.402 You want to set this up? 00:46:17.402 --> 00:46:22.202 First of all, I know obscurity isn't a thing, but it is a thing with this. 00:46:22.202 --> 00:46:25.732 Okay, this would be a thing. Second of all, it's so easy. Third of all, 00:46:25.732 --> 00:46:29.309 it does name resolution, so you don't have to set up home DNS. Yeah? 00:46:30.876 --> 00:46:34.226 And then if you really want to get classic, you get classic Mac OS, 00:46:34.226 --> 00:46:37.386 then you go into the Apple menu, you go into chooser. Did you ever have to use chooser? 00:46:37.386 --> 00:46:38.346 Yeah. Yeah. 00:46:38.346 --> 00:46:42.995 It's a lot of fun. I'm just thinking. I mean, what else are you going to use? IPX, NetBui? 00:46:43.895 --> 00:46:49.356 Why not use AppleTalk? And then you want a private file sharing network that 00:46:49.356 --> 00:46:52.206 nobody's ever going to discover and nobody's going to come looking for? 00:46:52.206 --> 00:46:53.986 It'd be your local little AppleTalk network. 00:46:54.046 --> 00:46:55.526 You want to try it out here at the studio? 00:46:55.526 --> 00:46:59.316 I mean, I kind of do before they take it away from us. I think it makes sense. 00:46:59.316 --> 00:47:00.938 We got one kernel release left. 00:47:01.478 --> 00:47:04.648 I feel like this deserves a moment of silence. Do you think it does? 00:47:05.306 --> 00:47:06.546 I mean, I'm feeling sad, boys. 00:47:06.546 --> 00:47:10.348 We've got a cone for this, and you can just, you know, grieve in the cone if you'd like. 00:47:16.583 --> 00:47:20.086 Okay, we should totally do this. In the studio. In the studio. 00:47:20.086 --> 00:47:24.126 We should totally. Come on. Have the whole studio just be Apple talk. Wouldn't that work? 00:47:24.946 --> 00:47:25.911 All right, you get over here. 00:47:26.097 --> 00:47:26.726 Oh, yeah. 00:47:26.726 --> 00:47:28.486 I'm for it. I'm officially approving it. 00:47:28.486 --> 00:47:33.758 Do you think if we had like a layer two VPN, we could sort of extend it across the network? 00:47:34.292 --> 00:47:37.676 That's what I was wondering is how you would route it. Because obviously you'd 00:47:37.676 --> 00:47:41.090 want to do that. Could you somehow run it over TCP IP? 00:47:41.822 --> 00:47:43.006 Like an Apple talk mesh? 00:47:43.766 --> 00:47:48.776 Encapsulate it, exactly. Could you encapsulate it? And then you would run it 00:47:48.776 --> 00:47:50.401 over your mesh network on top of IP? 00:47:51.411 --> 00:47:53.553 I don't see why. All right, moving on. 00:47:59.068 --> 00:48:02.175 Well, we may have something for you in the future. We'll have links to all the 00:48:02.175 --> 00:48:03.445 new, there's a lot more stuff. 00:48:03.445 --> 00:48:06.215 We just wanted to kind of cover the highlights and we'll put a link to the new 00:48:06.215 --> 00:48:11.605 stickers over in Jupyter Garage too, if you want to commiserate with me or flex 00:48:11.605 --> 00:48:14.126 that your kernel doesn't mess around with silly files. 00:48:16.750 --> 00:48:19.885 I want to take a moment and thank our members for making this episode possible. 00:48:19.885 --> 00:48:22.723 Members, you now get a free web boost to get your message in. 00:48:23.043 --> 00:48:28.354 You can send one free boost per episode. Also, you get access to the bootleg, a lot more content or 00:48:28.737 --> 00:48:33.115 the lean and mean ad-free version that editor drew puts together also available 00:48:33.115 --> 00:48:36.895 to our members the core contributor applies to this show and the jupiter.party 00:48:36.895 --> 00:48:39.150 is all the shows and all the special features, 00:48:39.661 --> 00:48:43.829 and you members you make it possible thank you very very much we appreciate you. 00:48:46.250 --> 00:48:55.021 We've got a special baller boost here this week from hybrid sarcasm the amount is 61,923 satoshis, 00:48:59.265 --> 00:49:03.885 all right hybrid says i boosted this amount to coder radio when my daughter 00:49:03.885 --> 00:49:08.915 was born three years ago since it's father's day today i figured i would do 00:49:08.915 --> 00:49:10.933 that again so happy father's day. 00:49:11.542 --> 00:49:15.485 Ah that's great oh that brings back the memories thank you hybrid thank you. 00:49:15.485 --> 00:49:19.791 Indeed there's a bit of metadata here too it says uh zap right you guys know anything about that. 00:49:21.045 --> 00:49:22.575 Oh but zap right is web boost. 00:49:22.575 --> 00:49:26.855 Yeah so that's our first baller web boost i would say oh Well. 00:49:26.855 --> 00:49:31.975 It might be. It might be. It might be. It might be. It might be not, though, either. 00:49:32.271 --> 00:49:35.434 Saltros comes in with 24,000 sats. 00:49:39.922 --> 00:49:44.502 Hey, guys, no time, long time, no boost. Soltros, we appreciate you. 00:49:44.728 --> 00:49:47.532 Soltros here. I retired Soltros OS, but I learned a lot. 00:49:47.880 --> 00:49:52.501 Right now, I'm working on Cabinet, a file locker, because I really don't like using NextCloud. 00:49:52.751 --> 00:49:56.401 Anyways, I'm very grateful for all you do. You do incredible work with the podcast, 00:49:56.401 --> 00:49:57.557 and thanks for being you. 00:49:58.155 --> 00:50:01.545 File Locker is a solid, or Cabinet, is a file locker. 00:50:01.911 --> 00:50:05.571 He writes, it's a lean, high-performance file locker designed for self-hosting. 00:50:05.571 --> 00:50:09.393 It prioritizes a fast web experience, a mobile-first design, 00:50:09.544 --> 00:50:12.807 and a simple file management without the bloat or complex syncing. 00:50:13.184 --> 00:50:15.558 That does sound maybe like a very useful thing. 00:50:15.947 --> 00:50:16.851 Yeah, that does look nice. 00:50:16.851 --> 00:50:18.252 Have something that runs somewhere. 00:50:18.641 --> 00:50:21.171 You're sitting on top of like a file share or something like that? 00:50:21.171 --> 00:50:21.613 Uh-huh. 00:50:22.385 --> 00:50:26.451 Optimized for mobile screens, he says. It can be added to your home screen as a progressive web app. 00:50:26.451 --> 00:50:27.690 That's a nice touch. 00:50:28.142 --> 00:50:29.832 Well done, Soltros. Well done. 00:50:30.795 --> 00:50:34.609 Daja Boo Sin with 9,494 sats. 00:50:36.619 --> 00:50:39.687 All right, question for you guys and anyone in the community wants to boost 00:50:39.687 --> 00:50:42.687 in. Do you have any preferred methods of dealing with burnout? 00:50:43.712 --> 00:50:47.337 I used to get rejuvenated by doing something weird with my home lab or joining 00:50:47.337 --> 00:50:51.590 in on a JB challenge, going for a hike, or letting my ADHD pick a random new hobby for me. 00:50:52.030 --> 00:50:56.500 But nowadays, everything seems to just make it worse. Love to hear strategies you guys use. 00:50:57.747 --> 00:51:02.117 Such a good question. It feels like you're always having to evolve the strategies. 00:51:02.117 --> 00:51:03.127 It does feel like it. 00:51:03.127 --> 00:51:06.847 Yeah, I had some answers to this one, but then he mentioned them all as not 00:51:06.847 --> 00:51:08.207 working anymore. So I don't know. 00:51:09.947 --> 00:51:11.367 I'd say get a squirrel. That helps. 00:51:12.040 --> 00:51:15.117 Are you sure about the home lab thing? I mean, sometimes stacking a dub in the 00:51:15.117 --> 00:51:19.547 home lab can help. But I think, you know, what I've been trying to do is get 00:51:19.547 --> 00:51:21.887 outside and do a couple of small things outside. Like I'm trying to, 00:51:22.247 --> 00:51:26.133 I'm currently fighting the biggest fight of my life with poison hemlock. 00:51:26.522 --> 00:51:29.567 So every day I try to get out there, do a little trimming and spraying, 00:51:29.567 --> 00:51:31.657 you know, just stack a little dubs. It isn't related to work. 00:51:31.657 --> 00:51:33.227 I still think there's something to that. 00:51:33.227 --> 00:51:34.407 Yeah, I do agree with that. 00:51:34.407 --> 00:51:37.567 I would wonder, though, if there isn't probably some bigger pressures that are 00:51:37.567 --> 00:51:40.397 just making all this other stuff not as enjoyable. And a lot of times for me 00:51:40.397 --> 00:51:44.097 when that happens, it's like bad sleep, maybe, you know, not eating well. 00:51:44.097 --> 00:51:47.407 So like a couple of issues are compounding, usually like three or more. 00:51:47.407 --> 00:51:49.324 And that's usually when I start to just grind down. 00:51:50.427 --> 00:51:52.099 And you're going to have to kind of address those. 00:51:52.505 --> 00:51:53.607 Yeah, it can be difficult, too, 00:51:53.607 --> 00:51:57.573 if it's, you know, there's pressures and maybe standards or, you know. 00:51:58.722 --> 00:52:05.079 It can be useful maybe to try and unpack why the things that no longer give 00:52:05.079 --> 00:52:08.549 you that, what about them has changed. 00:52:09.351 --> 00:52:10.749 Yeah, or if it's not them, it's changed. 00:52:11.209 --> 00:52:14.879 You may also consider getting a van. It's a really busy hobby that gives you 00:52:14.879 --> 00:52:18.449 all sorts of adventures. So that would be my top advice. 00:52:18.449 --> 00:52:21.693 And it won't financially destroy you quite as much as a boat. 00:52:22.076 --> 00:52:22.569 No, no, no. 00:52:23.150 --> 00:52:24.048 Not as much as a boat. 00:52:24.478 --> 00:52:28.909 Okay, well, we got a second Daja boost here. I'm going to some feedback on, 00:52:30.074 --> 00:52:33.699 local models versus cloud models and that kind of thing I'm in a similar boat 00:52:33.699 --> 00:52:37.959 to PJ as a general rule I don't fully integrate anything into my core day-to-day 00:52:37.959 --> 00:52:42.345 production until I can own it and run it on my own hardware or a VPSI control yeah 00:52:42.583 --> 00:52:45.529 probably I've been listening to Chris's to find networking ad reads for too long, 00:52:46.049 --> 00:52:51.129 but that being said one thing that is handy is that the OpenAI API spec is quickly 00:52:51.129 --> 00:52:54.969 turning into S3 of the AI world so in the interim even with the cloud models 00:52:54.969 --> 00:52:58.193 I can run them through a proxy Yeah. Which is, I think, a good tip. 00:52:58.448 --> 00:53:02.799 I think, too, that we're probably not injecting into the conversation very much 00:53:02.799 --> 00:53:06.499 is the potential and possibility of people running their own injector that sits 00:53:06.499 --> 00:53:07.709 in front of these cloud models. 00:53:08.387 --> 00:53:13.269 Which could be your way to make sure certain secrets or certain things you don't 00:53:13.269 --> 00:53:16.949 want shared, a way to route between the free available models. There's a lot. 00:53:17.429 --> 00:53:19.469 I mean, we really could almost do a whole stream just on the, 00:53:19.469 --> 00:53:21.219 we should do a stream on the injector stuff. 00:53:21.219 --> 00:53:22.029 That's a good idea. 00:53:22.029 --> 00:53:24.949 Yeah, because there's a lot there that I think would address some of the concerns 00:53:24.949 --> 00:53:28.981 of our audience. Thank you, Daja. Great boost. Good to hear from you. 00:53:29.382 --> 00:53:36.549 There's a boost here from Not Sure. It's a row of ducks. Web boost changes everything. Great work. 00:53:37.474 --> 00:53:44.309 Great. Glad you like it, Not Sure. Good to hear from you. Gene Bean's here with 7,777 sets. 00:53:46.651 --> 00:53:50.898 I am super excited about the U-clock, Chris. I wanted a nice-looking Home Assistant 00:53:50.898 --> 00:53:52.717 controllable alarm clock for my son. 00:53:53.861 --> 00:53:55.950 I already have him waking up with lighting in Home Assistant. 00:53:56.363 --> 00:53:59.938 Until then, alarms have been an Echo Dot thing. I just ordered the clock, 00:53:59.938 --> 00:54:00.868 and I can't wait to play with it. 00:54:00.868 --> 00:54:01.658 That is great. 00:54:01.658 --> 00:54:05.938 You know what, Gene? That's such a great—it's so much more useful than the Echo 00:54:05.938 --> 00:54:07.886 Dot. It's so much more useful. 00:54:08.257 --> 00:54:12.948 I, again, saved my bacon. This time on the—I think it was the fridge last time. 00:54:12.948 --> 00:54:14.898 It was the freezer this time, or vice versa. 00:54:15.664 --> 00:54:19.938 I saved our bacon and the other thing I added was now a three minute warning 00:54:19.938 --> 00:54:23.578 when the freezer door has been open for more than that's nice yeah just a little 00:54:23.578 --> 00:54:24.658 warning there on the thing. 00:54:24.658 --> 00:54:25.698 I kind of want that. 00:54:25.698 --> 00:54:29.278 It's if you clock is great if you don't know what we're talking about I think 00:54:29.278 --> 00:54:32.718 I buried the lead a lot by putting what is probably one of my favorite gadgets 00:54:32.718 --> 00:54:36.694 of the year so far casually as a pick last week but sometimes the picks hit, 00:54:37.136 --> 00:54:38.192 sometimes those picks hit. 00:54:39.078 --> 00:54:42.488 Yeah Nathan bussin with 2223 sets. 00:54:44.032 --> 00:54:44.318 Love it. 00:54:44.758 --> 00:54:51.168 Looking forward to trying windows mcb anything to spend less time remoting in fixing it's s. 00:54:51.168 --> 00:54:56.018 You get it nathan you get it that's it oh god when you it was in the clip the 00:54:56.018 --> 00:54:58.308 moment i realized oh i'm not gonna 00:54:58.308 --> 00:55:02.080 have to use windows hardly at all ever anymore like it's already rare. 00:55:02.938 --> 00:55:03.815 Now it's even more rare. 00:55:04.837 --> 00:55:08.778 And also like if a family member really needs help i got an option that's really, 00:55:09.438 --> 00:55:11.548 that's the one that gets me it's huge. 00:55:12.517 --> 00:55:17.458 Well aaron boosted in with uh one two three four five sets, 00:55:22.349 --> 00:55:26.129 I know it's only ever played in the bootleg of the show, but my favorite AI 00:55:26.129 --> 00:55:32.379 song you have made is the time-killing song you play sometimes during the pre-show break. 00:55:33.139 --> 00:55:38.609 Can you share that with the audience somehow? I think it would be fun to use 00:55:38.609 --> 00:55:42.459 the college acapella group profile on. 00:55:44.719 --> 00:55:45.389 I like this. 00:55:45.389 --> 00:55:46.980 We do have we have yet to do this. 00:55:58.405 --> 00:55:59.206 Yeah, it's a fun song. 00:55:59.363 --> 00:56:00.762 I've never heard this one before. 00:56:01.197 --> 00:56:05.233 It's because you're always out having a break. We play it when you're away from the song. 00:56:05.233 --> 00:56:07.573 Is that right? What else do you play when I'm gone? 00:56:07.573 --> 00:56:10.845 The song may be all about you for all you know. I don't know. 00:56:11.687 --> 00:56:15.103 One day we'll figure out a way to post those AA rounds. We do need to figure 00:56:15.103 --> 00:56:17.513 that out. Just, there's a lot to do. 00:56:20.323 --> 00:56:23.092 All right, Budai comes in with a row of duck-a-ducks. 00:56:23.765 --> 00:56:24.733 Oh, with a flex. 00:56:25.033 --> 00:56:28.781 Okay, what do we got here? Here we have LS piped into a word count, 00:56:29.753 --> 00:56:34.653 eight, which I treat my download. Oh, talking about he's got eight files in his download folder. 00:56:35.793 --> 00:56:38.550 Finally, I wanted more people to give us this. I want to know. 00:56:38.829 --> 00:56:40.582 I'm not going to reveal mine until we get a few of these. 00:56:40.948 --> 00:56:45.153 I treat my downloads folder like slash temp. It even gets cleared on reboot. 00:56:45.153 --> 00:56:47.693 Oh, that's wild. I love that. 00:56:47.693 --> 00:56:50.085 So eight actually feels pretty high in this case. 00:56:50.677 --> 00:56:53.733 I'm going to definitely admit that I do not clear it out on reboot, 00:56:53.733 --> 00:56:59.454 boys. Let's see if I can just tell you the oldest file in my downloads folder. 00:57:02.653 --> 00:57:05.953 Bad to Hi goes on to say, if there was something on there that I needed and 00:57:05.953 --> 00:57:08.609 got deleted, then I didn't need it that much. 00:57:08.997 --> 00:57:09.410 Good point. 00:57:09.979 --> 00:57:12.153 Okay. All right. That's not as bad as... 00:57:12.213 --> 00:57:12.913 There's a certain zen to it. 00:57:12.913 --> 00:57:17.477 Oh, I cleaned up. It's not as bad as I expected. It's going into last year. 00:57:17.756 --> 00:57:20.368 I'll just leave it at that. If we get more, I might reveal. 00:57:20.479 --> 00:57:23.583 I wonder what the mechanism is here. I'd be curious about that. 00:57:23.583 --> 00:57:25.383 Mechanism for cleaning? 00:57:25.383 --> 00:57:30.023 A little boot script? Yeah, we'd love to know. How are you actually cleaning that? 00:57:30.023 --> 00:57:34.561 Just a little RMRF as a boot script. That's not dangerous at all. 00:57:34.759 --> 00:57:36.587 Tomato comes in with 5,000 SATs. 00:57:39.309 --> 00:57:43.737 It's a boost. The one reason I'm sometimes I have to use Windows is when the 00:57:43.737 --> 00:57:47.867 latest DDRM tools, ah, for the Kindle, stop working with Linux. 00:57:47.867 --> 00:57:48.187 Oh. 00:57:48.187 --> 00:57:51.546 This Windows MCP thing sounds like it might just save my Unix bacon here. 00:57:51.703 --> 00:57:55.077 Nice. I haven't even thought of that. You should let us know how that goes. 00:57:55.077 --> 00:57:59.301 I'd be very curious. That is a great use case, Tomato. Please keep us boosted on that one. 00:57:59.621 --> 00:58:03.783 Distro Stew boosting in with 12,345 sets. 00:58:06.308 --> 00:58:07.823 Okay. 00:58:08.705 --> 00:58:10.437 I came to test out the web. 00:58:10.437 --> 00:58:14.683 Boost Looking good Nice I don't know why you boosted with Windows XP though That's weird. 00:58:15.032 --> 00:58:19.737 And And Here's my Nix router config. 00:58:19.737 --> 00:58:23.507 First web boost with a link First web boost with a link boys That's cool We got. 00:58:23.507 --> 00:58:24.377 A flake here Ooh. 00:58:24.377 --> 00:58:28.087 Router.nix Wonder what's in here Let's go check out that router.nix Oh he's 00:58:28.087 --> 00:58:34.447 using ZFS On his router I love it That's great Couple interfaces here. 00:58:34.447 --> 00:58:40.202 We got DNS mask Yep Oh, uh, Kia for DHCP. Nice. 00:58:41.229 --> 00:58:44.267 I could look at these configs all day. I really could. I also like, 00:58:44.267 --> 00:58:46.587 I'd like to start at the flake and you gotta, you gotta, you gotta go through 00:58:46.587 --> 00:58:48.997 the main config and then get to the, you gotta, you're jumping ahead. 00:58:48.997 --> 00:58:50.755 You're going to the end of the book. I gotta say. 00:58:51.377 --> 00:58:54.012 Well, I wanted to get to the good stuff. We only have so much time. 00:58:54.564 --> 00:58:57.785 Whomever whiz comes in with 9,009 sats. 00:59:00.316 --> 00:59:03.885 I really wanted to clip some audio for you from the Star Trek The Next Generation, 00:59:03.885 --> 00:59:06.465 but I didn't even know what episode it was in. I realized, however, 00:59:06.465 --> 00:59:09.615 the subtitles exist. So instead of just clipping the audio, I made a tool to 00:59:09.615 --> 00:59:12.765 easily clip dialogue from any video in your library. Check it out. 00:59:13.305 --> 00:59:15.727 Wecker. W-E-C-K-E-R. 00:59:16.314 --> 00:59:19.825 Also, you could have just told us the line and we would have told you the episode it's from. 00:59:20.345 --> 00:59:21.985 But I love the idea of this tool. 00:59:21.985 --> 00:59:24.855 I don't think the repository is public yet, but it's a great idea. 00:59:24.855 --> 00:59:26.791 Check out. Oh, Quip Clipper. 00:59:27.372 --> 00:59:31.405 We need this in our lives so bad. we're constantly trying to figure out how 00:59:31.405 --> 00:59:32.585 we could take little sound bits. 00:59:33.333 --> 00:59:35.035 So we'd love to keep this thing fresh all the time. 00:59:36.125 --> 00:59:38.982 And just as an aside, if anybody ever wanted to help us, 00:59:39.718 --> 00:59:45.756 if you have like, say, a backup MKV file of an episode or a movie that's in 5.1 stereo, 00:59:46.244 --> 00:59:50.065 if you were to open that MKV file in an application like Reaper or anything 00:59:50.065 --> 00:59:53.195 that supports multi-track editing, you might be surprised to learn that often 00:59:53.195 --> 00:59:56.395 the dialogue is isolated on its own track and you can extract it from all the 00:59:56.395 --> 00:59:58.040 other background noises to a certain degree. 00:59:58.811 --> 01:00:02.005 Just if you didn't know that. Whomever, Wiz, thank you very much for the boost. 01:00:02.005 --> 01:00:03.728 We really do appreciate you on that. 01:00:04.460 --> 01:00:11.209 Well, I'm going to sneaky pull one up here. Todd from Northern VA with 1,200 sats. 01:00:12.475 --> 01:00:17.255 He says, long time no boost, but this is a test boost after resurrecting my 01:00:17.255 --> 01:00:19.955 node after one plus year of downtime. 01:00:20.295 --> 01:00:22.735 Nicely done. Test boost received. 01:00:22.735 --> 01:00:25.774 Todd, we know exactly how that feels. We have that problem too. 01:00:27.068 --> 01:00:31.225 That's great. Greg the lawyer sent in 50 fiat dollars. 01:00:34.593 --> 01:00:37.812 Using the web boost to say, keep up the great work. Thank you, Greg. 01:00:37.812 --> 01:00:38.332 Thank you. 01:00:38.332 --> 01:00:42.282 Nice to hear you. What kind of lawyering do you do, Greg? Always love to know 01:00:42.282 --> 01:00:45.222 that kind of stuff. That's awesome. Thank you very much. Appreciate it. 01:00:45.500 --> 01:00:49.047 Carl N. comes in with $25 via Zapparat Fiat. 01:00:50.922 --> 01:00:51.161 Woo! 01:00:53.535 --> 01:00:54.051 Bah, bah. 01:00:54.899 --> 01:00:58.262 Easy in my guild for writing on the self-hosted membership deal for so long. 01:00:58.582 --> 01:01:00.792 Thanks for clinker therapy. Really enjoyed it. 01:01:00.792 --> 01:01:04.211 I'm glad you liked it. Thank you very much. Appreciate it. 01:01:04.796 --> 01:01:08.905 Well, Brad sent in 20 fiat dollars. 01:01:10.066 --> 01:01:13.422 I enjoyed the Clanker Therapy recording and all that Hermes info. 01:01:13.802 --> 01:01:17.832 More links and examples would be appreciated. Love the easier to contribute 01:01:17.832 --> 01:01:20.112 web boost option with this fiat. 01:01:20.112 --> 01:01:22.432 Awesome. Thank you. Sire came 01:01:22.432 --> 01:01:26.302 in with $5. Finally, a great way to spend my hard-earned fiat currency. 01:01:29.211 --> 01:01:31.608 Yeah, get rid of that dirty fiat. Throw it at us. That's fine. 01:01:32.482 --> 01:01:34.232 We had a couple of member boosts come in too. 01:01:34.232 --> 01:01:35.022 More than a couple. 01:01:35.662 --> 01:01:36.391 Yeah, it's new. 01:01:36.589 --> 01:01:37.062 It's new. 01:01:37.062 --> 01:01:38.512 It's new. So we should expect that. 01:01:38.512 --> 01:01:39.636 Uh-huh. It's exciting. 01:01:40.302 --> 01:01:43.851 Now, some of you decide to go anonymous. You can put your name in there in the field. 01:01:44.228 --> 01:01:48.077 It is all right. We would still like it. Anonymous boosts in with two boosts here. 01:01:48.442 --> 01:01:52.992 IDK about folder numbers, but I do know my Bitcoin numbers. And with Bitcoin 01:01:52.992 --> 01:01:55.501 under 100K, USD member boosts are a steal. 01:01:56.227 --> 01:01:59.797 But will you increase membership prices when the price goes above 100,000? 01:02:00.215 --> 01:02:02.512 We haven't really ever adjusted. I mean, if you think about it, 01:02:02.512 --> 01:02:05.842 I think at one point it was around 16 or so when we started. 01:02:05.842 --> 01:02:06.413 Yeah, true. 01:02:06.605 --> 01:02:09.352 And we've always just kind of kept it what it is because we don't really denominate 01:02:09.352 --> 01:02:14.082 in dollars. We denominate in sats. So it's just a sats a sat really is kind 01:02:14.082 --> 01:02:17.812 of how we think about it. But I suppose anything's always up for reevaluation 01:02:17.812 --> 01:02:18.720 if you have a better idea. 01:02:19.051 --> 01:02:21.292 Somebody sends us a really great idea. You know, why not? 01:02:21.588 --> 01:02:21.936 Mm hmm. 01:02:23.387 --> 01:02:26.676 Ah, TimeKillerTK comes in with a member boost. 01:02:26.676 --> 01:02:27.845 Ooh, I like that name. 01:02:28.031 --> 01:02:30.766 Long-time listener since the days of TechSnap with Alan. 01:02:30.766 --> 01:02:31.286 Right on. 01:02:31.286 --> 01:02:35.166 Loved Clanker Therapy? Please do more. Zip code boost? 01:02:35.166 --> 01:02:39.293 Oh, great. All right, so we got a zip code boost and a plus one for Clanker Therapy. 01:02:39.444 --> 01:02:39.836 Yeah. 01:02:39.836 --> 01:02:45.936 Okay, so here's the numbers, Wes. Three, three, three, L-E. 01:02:46.336 --> 01:02:46.606 Oh. 01:02:46.606 --> 01:02:47.716 No, no, you missed a three. 01:02:48.871 --> 01:02:49.506 Oh, did I? 01:02:49.506 --> 01:02:49.916 Yeah. 01:02:50.038 --> 01:02:52.575 How many three? I need to know the right number of threes here. 01:02:53.022 --> 01:02:55.266 Oh, the blurry vision. It looks like three threes to me. 01:02:55.266 --> 01:02:56.376 Brent, how many threes am I working on? 01:02:56.576 --> 01:02:58.046 There's four threes. There are four threes. 01:02:58.046 --> 01:03:02.056 One, two, yeah. It says three, three, three, three, L-E. 01:03:02.376 --> 01:03:06.529 There are four lights. There are four lights. It is true. 01:03:06.877 --> 01:03:08.356 Okay. Well, my man. 01:03:08.356 --> 01:03:09.926 Are you playing on your map? I don't quite. 01:03:09.926 --> 01:03:10.226 Yeah. 01:03:10.226 --> 01:03:12.067 Yeah, I don't hear it, Wes. Oh. 01:03:12.171 --> 01:03:12.496 Oh. 01:03:12.496 --> 01:03:12.936 There it is. 01:03:12.936 --> 01:03:14.896 For a minute there, I thought he had a silent map. 01:03:15.608 --> 01:03:16.827 Yeah, well, he did for a while. 01:03:19.056 --> 01:03:19.386 Okay. 01:03:19.386 --> 01:03:20.016 What you got there, Wes? 01:03:21.196 --> 01:03:27.328 If I've picked the right continent, you never know. It is in the town of Zwingsdrecht, 01:03:28.076 --> 01:03:29.716 in South Holland, Netherlands. 01:03:29.716 --> 01:03:31.293 Sounds probably just like you. Nice. 01:03:31.432 --> 01:03:34.660 I probably butchered that, so tell us, boost back in. Thank you, Time Killer. 01:03:35.008 --> 01:03:36.556 Thank you, Time Killer. Appreciate you. 01:03:36.556 --> 01:03:37.196 That's great. 01:03:37.196 --> 01:03:37.725 All right, Bradley. 01:03:37.980 --> 01:03:40.981 Gong Fuja boosted in a member's boost. 01:03:41.167 --> 01:03:41.596 Hmm. 01:03:43.842 --> 01:03:49.221 Loving the content and extras for members. I actually agree with producer Jeff 01:03:49.221 --> 01:03:51.418 here on the idea of using AI. 01:03:51.830 --> 01:03:55.781 I agree it is helpful and open source will catch up in the end, 01:03:55.781 --> 01:04:01.292 but I don't like short cutting knowledge and not really learning it as a result, 01:04:01.443 --> 01:04:03.340 especially when things go wrong. 01:04:03.572 --> 01:04:07.241 I think we had an interesting follow-up chat today on that. PJ has really been 01:04:07.241 --> 01:04:09.104 diving in deep on the local model stuff. 01:04:09.406 --> 01:04:13.383 We've got a good local model combo in the bootleg today. So I'd like to know what you think, Gong. 01:04:13.921 --> 01:04:17.761 And I think part of, you know, playing with the tools is learning how to set 01:04:17.761 --> 01:04:19.281 them up so that it does work for you, right? 01:04:19.281 --> 01:04:23.041 So you do find the right, like, where do you offload and where do you inspect 01:04:23.041 --> 01:04:26.421 or follow up or gatekeep or, you know, like, where do you set the gates in your 01:04:26.421 --> 01:04:29.271 workflow and which pieces do you focus on? 01:04:29.904 --> 01:04:32.161 And that'll probably be different for a lot of different people. 01:04:32.161 --> 01:04:36.231 What are you really trying to get done? That really is the defining thing which 01:04:36.231 --> 01:04:37.665 you'd expect from a local model right now. 01:04:38.170 --> 01:04:41.841 Well, and like you can have different, you know, some folks might vibe stuff 01:04:41.841 --> 01:04:43.029 up and throw it up and test it. 01:04:43.330 --> 01:04:46.721 it's kind of smoke test it live but you you could you know have a proposing 01:04:46.721 --> 01:04:49.781 architecture and then you walk through every piece with it and then it sets 01:04:49.781 --> 01:04:52.879 it up you know you could do take totally different approaches in terms of like 01:04:53.211 --> 01:04:54.644 how involved you were at each step. 01:04:55.991 --> 01:05:01.611 Sensor smile comes in with a member boost long time listener since 2013, 01:05:04.691 --> 01:05:08.211 first time booster call me a curmudgeon but i'm not going to try crypto in this 01:05:08.211 --> 01:05:12.034 life after hearing all these boosts all these years i can finally send in a Message as a member. 01:05:12.411 --> 01:05:14.941 I'll be sure to send some USD when I can in the future. Thanks and keep them 01:05:14.941 --> 01:05:17.011 coming. Glad to make it happen for you. 01:05:17.011 --> 01:05:17.371 Amazing. 01:05:17.371 --> 01:05:19.563 Sensor smile. Welcome to the Boost Club. 01:05:21.641 --> 01:05:26.839 Uh, whomever wheeves boosts in here, um, maybe it's a follow-up boost from before 01:05:26.839 --> 01:05:32.399 because one cool thing I did with Quip Clipper was batch export all the lines from TNG. 01:05:32.399 --> 01:05:37.169 Resistance is... No way! You could batch line all the Resistance is Feudal lines? 01:05:37.169 --> 01:05:37.609 Oh my goodness. 01:05:37.609 --> 01:05:40.839 It works perfectly with only a few clicks, and I also made it so you can export 01:05:40.839 --> 01:05:42.109 just the center channel. 01:05:42.109 --> 01:05:43.079 That is legit. 01:05:43.079 --> 01:05:43.929 Oh, wow. 01:05:43.929 --> 01:05:45.329 What? I was just saying that! 01:05:45.329 --> 01:05:46.049 You sure were. 01:05:46.049 --> 01:05:49.319 That is so good. Yeah, look at this. He sent us some copies of it, 01:05:49.319 --> 01:05:52.309 too. That is good work. Well done, sir. 01:05:56.249 --> 01:05:59.866 So he's got all the resistances futile now? I'm jelly. 01:06:01.555 --> 01:06:05.219 Aw, Hugh. Aw, that's nice. That one's a good one. 01:06:05.219 --> 01:06:08.129 Yeah, you should. You got to send us an updated link because we got, 01:06:08.129 --> 01:06:11.669 I think, maybe just your GitHub profile, but not, I want to try this. 01:06:11.669 --> 01:06:16.393 Yeah, that is wicked, dude. That's great. Who's next? 01:06:16.782 --> 01:06:17.489 Marcel. 01:06:17.489 --> 01:06:21.124 All right. Marcel comes in with a free member boost. 01:06:21.576 --> 01:06:26.089 You can read higher voltages with it. Oh, good. Yes, with using an ESP resistor 01:06:26.089 --> 01:06:30.859 divider. It basically maps 0 to 12 volts to 0.3 volts. Yes, I do know about 01:06:30.859 --> 01:06:34.545 this. It's not my preferred way to measure this voltage. 01:06:35.044 --> 01:06:37.912 Brent, we've talked about this before. It's not quite ideal. 01:06:38.086 --> 01:06:43.719 Yeah, we did play with this when we were doing that diesel heater reverse engineering 01:06:43.719 --> 01:06:47.009 because we had several strange issues. 01:06:47.009 --> 01:06:51.149 And voltage was one thing that helped with diagnosing a bunch of stuff. 01:06:51.149 --> 01:06:54.671 So we did use a resistor to do exactly this. 01:06:55.321 --> 01:07:00.509 But you kind of end up, I think, in the car with diagnostics and stuff, 01:07:00.509 --> 01:07:03.939 you end up with like, remember, Chris, I had like wires everywhere and like 01:07:03.939 --> 01:07:07.365 little tiny components and resistors and like transistors and stuff. 01:07:11.114 --> 01:07:14.394 Yeah, a dedicated device would be really nice. Something also that, 01:07:14.394 --> 01:07:17.154 you know, if you drop it, it doesn't fall into a million pieces. 01:07:17.994 --> 01:07:19.294 So mostly a packaging issue. 01:07:19.875 --> 01:07:22.614 I also wonder, could there be a slight resolution loss? 01:07:22.614 --> 01:07:23.654 Yeah, that's a good point. 01:07:23.654 --> 01:07:26.454 Because we're looking at really micro measurements here. 01:07:26.514 --> 01:07:29.634 Mm-hmm, mm-hmm. So that's something to consider. Depends what you're doing. 01:07:29.634 --> 01:07:34.204 But Marcel, I think it's in the category of worth me playing around with because 01:07:34.204 --> 01:07:37.414 I think I have most of the components I need to actually try it. so. 01:07:37.954 --> 01:07:41.854 I like that marcel's clearly paying attention because there's this line there's 01:07:41.854 --> 01:07:44.624 lots of online calculators or i'm sure data could help you. 01:07:44.624 --> 01:07:48.778 With it he was paying attention but i like this last line. 01:07:49.805 --> 01:07:54.658 Uh just be careful because you could easily fry the esp with that 12 volts he got over there. 01:07:55.261 --> 01:07:56.354 Yeah that is true. 01:07:56.354 --> 01:07:59.364 Have done which is another thing to consider perhaps in that you have to just 01:07:59.364 --> 01:08:00.816 navigate that with part of the setup. 01:08:00.938 --> 01:08:05.384 I i think you know my route's probably going to be a meter that has a usb port 01:08:05.384 --> 01:08:07.173 on it that can talk to a linux box 01:08:07.596 --> 01:08:12.914 it's just there's the perfect devices available if you're in europe that are 01:08:13.254 --> 01:08:18.283 just exactly what i'm looking for but with with taxes and shipping to get two of them, 01:08:18.857 --> 01:08:24.726 is almost 300 and they're like 40 to 60 devices if you do the conversion. 01:08:25.150 --> 01:08:26.688 Should you just toss a link in here i'm curious. 01:08:26.985 --> 01:08:30.944 Yeah i need to find them um they're It just, you know, like when you have like 01:08:30.944 --> 01:08:34.174 a dream device, it's funny that it's like a $50 device or something like that. 01:08:34.174 --> 01:08:36.018 And that's, we'd love to have two of those. 01:08:36.771 --> 01:08:39.756 It's just the way the world works right now. Tomatoes here with a member boost. 01:08:40.086 --> 01:08:43.704 I'm trying out the free web boost, but I have to say boo for the cloud flare 01:08:43.704 --> 01:08:48.074 nonsense, boosting two shows in a row. And I had to prove I'm not some bad actor. Heck no. 01:08:49.274 --> 01:08:54.574 So this was we were experimenting with serverless to essentially avoid spinning 01:08:54.574 --> 01:08:59.974 up even more VPS servers with NGINX and stuff that we're running hundreds of. 01:09:01.534 --> 01:09:06.475 So we'll see. You know, I'm always a little skeptical on the Cloudflare side myself. 01:09:07.868 --> 01:09:12.431 Greg, the lawyer member boosts in, boosting in to say, keep up the good work. 01:09:13.151 --> 01:09:15.450 Thank you, Greg. Hello. And again, thank you. 01:09:15.844 --> 01:09:19.392 We also have a leaky canoe member with a members boost. 01:09:21.098 --> 01:09:26.533 Plus one for topical live streams off the main feed. I followed Chris, 01:09:26.533 --> 01:09:30.283 Chris's lead after the clanker therapy stream and set up a multi-agent setup 01:09:30.283 --> 01:09:35.465 and deployed wrappers for the critical functions. I want my agents to accomplish. 01:09:36.186 --> 01:09:42.746 However, I'd like to ask, how do you handle when your agents are, quote, moving too fast? 01:09:43.180 --> 01:09:46.246 It's usually a gap in the skill. 01:09:46.634 --> 01:09:50.547 And so they're just kind of going off on their own and trying to figure it out, 01:09:50.674 --> 01:09:52.502 just using tool calls and whatnot. 01:09:53.313 --> 01:09:57.293 So you might see if maybe there's a little more implicit instructions you can 01:09:57.293 --> 01:09:59.343 give in the skill. Any thoughts on that, Wes? 01:09:59.343 --> 01:10:03.873 Yeah, maybe more some sort of gate or interaction point. or. 01:10:04.924 --> 01:10:07.863 Yeah approval gates are good too if they're having a problem you could always 01:10:07.863 --> 01:10:11.658 say okay stop here and you know put approval gate in that does work pretty well 01:10:11.942 --> 01:10:13.881 and that's sometimes the way you can work through the problem. 01:10:13.986 --> 01:10:17.873 I also like if you you know if you make them work in super small increments 01:10:17.873 --> 01:10:22.293 and maybe you could have checkpoints or have them at regular intervals use various 01:10:22.293 --> 01:10:28.243 you know other agents or sub agents for review as a way to sort of quote unquote slow things down yeah. 01:10:29.966 --> 01:10:33.734 Let us know how it goes, Leaky. Stuart comes in with a member boost, 01:10:33.734 --> 01:10:35.719 saying excited to see a web boost option. 01:10:36.299 --> 01:10:39.014 Getting AlbiHub set up has always been on my backlog. I'm sure I'm going to 01:10:39.014 --> 01:10:41.471 get to it eventually, but this is nice in the meantime. 01:10:41.855 --> 01:10:44.684 I found self-host about six months before it ended, and I decided to come over 01:10:44.684 --> 01:10:47.754 and give LUP a chance. Been a fan since. Well, I'm glad you did, Stuart. 01:10:47.754 --> 01:10:48.426 Nice. 01:10:48.757 --> 01:10:49.523 Nice to hear from you. 01:10:51.118 --> 01:10:56.476 Well, this one's from Pavel, who's a friend from Berlin that I met. 01:10:57.701 --> 01:11:01.265 testing this good old member boost functionality. 01:11:01.660 --> 01:11:02.304 It worked. 01:11:02.420 --> 01:11:03.251 Good to see you. 01:11:04.830 --> 01:11:09.868 Bobby member boost in woohoo less friction more inclination to boost thoughts 01:11:10.211 --> 01:11:13.914 to that end i'm beginning to get it ai i mean been playing with some of the 01:11:13.914 --> 01:11:17.734 stuff you talk about burning question though are you running llms. 01:11:17.734 --> 01:11:19.724 Locally isn't that always the burning question if so. 01:11:19.724 --> 01:11:23.254 What is that magic hardware want to self-host where possible. 01:11:24.014 --> 01:11:28.134 Don't we all that's been one of our kind of two week theme conversations in 01:11:28.134 --> 01:11:31.494 the bootleg is like what can you actually do locally and the answer is not a 01:11:31.494 --> 01:11:34.004 lot right now getting to be better. 01:11:34.004 --> 01:11:36.114 But we have a producer actively researching. 01:11:36.114 --> 01:11:40.624 Yeah pj's on the case and uh he's reporting back good question also shout out 01:11:40.624 --> 01:11:46.782 to zippy frog did a test boost and space warlock came in with a member boost his first boost, 01:11:49.446 --> 01:11:52.391 This is great. I haven't found an iOS Matrix client I like, and I'm not usually 01:11:52.391 --> 01:11:55.171 at my desk during the live streams. I've been hoping for this ever since you 01:11:55.171 --> 01:11:56.633 did Texas Linux Fest last year. 01:11:56.964 --> 01:11:59.791 Thanks for getting this figured out. Thank you, Warlock. 01:11:59.791 --> 01:12:00.271 Thank you. 01:12:00.271 --> 01:12:01.221 Nice to hear from you. 01:12:01.221 --> 01:12:04.631 Hey, props to Chris, who got the whole front end done. 01:12:04.951 --> 01:12:09.439 Oh, please. It wouldn't be anything without the fancy back end and reporting, Westpain. 01:12:10.379 --> 01:12:13.911 Scuffed member boosts in. Free member boosting check-in. 01:12:13.911 --> 01:12:14.381 Yeah. 01:12:14.381 --> 01:12:19.281 Nice. And then we got Jimmy member boosting in. Okay, this is one I've been looking forward to. 01:12:19.281 --> 01:12:19.562 Okay. 01:12:20.631 --> 01:12:24.721 It's dressed to me, so this is appropriate, at Wes. Do you remember how Chris 01:12:24.721 --> 01:12:26.311 fell in love with another man's van? 01:12:27.251 --> 01:12:30.051 Are we really sure it is another man's van? 01:12:31.111 --> 01:12:36.791 I'm a bit fussy on details, you see, but I'm pretty sure it's not actually another man's van. 01:12:36.811 --> 01:12:39.931 How did it cross the border? Who's it registered to? 01:12:40.471 --> 01:12:43.281 I suspect it's actually Chris's bank bus. 01:12:43.281 --> 01:12:49.431 Well, I mean, it depends on the technicalities. you know jimmy just technicalities 01:12:49.431 --> 01:12:51.308 there technicalities uh, 01:12:52.231 --> 01:12:56.151 it's right now it's the internet's bang bus let's be honest it's the internet, 01:12:57.043 --> 01:12:59.377 it's everyone now if it ever makes it back i have no idea. 01:12:59.905 --> 01:13:01.838 Well you you got to make that tempting you know. 01:13:02.100 --> 01:13:03.069 Put on the sugar uh-huh. 01:13:06.181 --> 01:13:10.407 Well chupacabra came in with a free members boost. 01:13:10.668 --> 01:13:11.684 Hey. 01:13:13.964 --> 01:13:16.844 Clanker therapy was fire. 01:13:17.053 --> 01:13:17.500 Thank you. 01:13:17.668 --> 01:13:23.410 While the stream was loaded with information, I'm experiencing the blank canvas syndrome. 01:13:24.424 --> 01:13:29.894 What would be a good next step on a fresh Nix OS system? I was thinking of unleashing 01:13:29.894 --> 01:13:36.001 clod code over SSH and let it loose with, I want to deploy Hermes agent. Let's cook. 01:13:38.694 --> 01:13:43.784 I mean, why not? But, you know, I've said this before, but what has been supremely 01:13:43.784 --> 01:13:45.874 useful, and whenever people say, I don't know what to do with the agent, 01:13:45.874 --> 01:13:47.481 I think you don't have anything you self-host. 01:13:48.167 --> 01:13:51.393 You don't have anything annoying that you have to, like, move information from one spot to another. 01:13:52.154 --> 01:13:56.584 It's super handy to put it in front of your self-hosted services. 01:13:56.584 --> 01:14:01.014 And I think you might find the family approval factor of using whatever their 01:14:01.014 --> 01:14:04.164 favorite chat app that they've already been using for years to interact with 01:14:04.164 --> 01:14:07.614 whatever you've set up as a smart home thing increases the approval factor by 01:14:07.614 --> 01:14:09.220 a substantial amount that makes it worth it. 01:14:09.696 --> 01:14:12.285 So I would say if you have anything you're already self-hosting, 01:14:12.616 --> 01:14:16.574 you know, get a Hermes thing going, get it cooking, and then stand it up and 01:14:16.574 --> 01:14:21.563 have it start interfacing with the API that that home, that self-hosted service almost no doubt has. 01:14:22.160 --> 01:14:26.094 Such a good way to go. like you got jellyfin or plex or you got things that 01:14:26.094 --> 01:14:30.494 are managing you know media for those instances like why are you going to a 01:14:30.494 --> 01:14:34.204 web ui and searching for a file when you can just tell the bot you know go get 01:14:34.204 --> 01:14:36.904 me this make it easy for them i. 01:14:36.904 --> 01:14:42.004 Also have a lot of fun playing with um having it use like a text-to-speech thing 01:14:42.004 --> 01:14:45.574 so like something like pocket ttsa there's a bunch of options so that could be another thing fun. 01:14:45.574 --> 01:14:47.869 Thing news briefs and stuff like that can be a lot of fun. 01:14:49.174 --> 01:14:53.454 I would also argue that if you have a project that you've been meaning to do 01:14:53.454 --> 01:14:58.174 for like a year, maybe two, maybe five, have at her, right? 01:14:58.594 --> 01:15:03.224 Let Hermes just kind of suggest what to do with it or tackle it completely for 01:15:03.224 --> 01:15:04.245 you and get that thing done. 01:15:04.593 --> 01:15:08.444 Start chewing away at it. Start it. Start it. Let us know how it goes. 01:15:08.444 --> 01:15:10.322 You know what I'm saying? Let's do it. 01:15:11.501 --> 01:15:14.214 All right. Anton comes in with a member boost. All right. 01:15:16.053 --> 01:15:20.604 Long-time listener member since around 2012-13. Hey-o! Nice. 01:15:21.219 --> 01:15:24.390 First-time booster. Hey-o! Not a fan of cryptocurrency, so the web boost is 01:15:24.390 --> 01:15:27.611 awesome. Thanks for all the years of great shows. Hope there are many more to come. 01:15:27.936 --> 01:15:30.391 Any chance the message limit could be longer than 300 characters? 01:15:31.512 --> 01:15:35.200 I don't know about that, man. It makes these segments pretty, pretty long. 01:15:35.520 --> 01:15:36.280 It is a balance. 01:15:36.280 --> 01:15:39.320 It'd be interesting, like, occasional member extra long. Oh, 01:15:39.680 --> 01:15:44.270 yeah. Thank you. Appreciate it. Nice to hear from you. Thank you for the long-time support. 01:15:46.945 --> 01:15:52.803 And our final member boost today, which is our last boost, from Hi5Connoisseur. 01:15:52.813 --> 01:15:53.603 Hey, Hi5. 01:15:54.526 --> 01:15:58.293 So excited about the web boost and the member boost, I fell off of AlbiHub due 01:15:58.293 --> 01:16:03.973 to the cost of opening the channel and, well, my addiction to PocketCasts. Great work, as always. 01:16:05.093 --> 01:16:07.723 You know, I used to be a big PocketCast user. I can't go back. 01:16:07.723 --> 01:16:10.233 I thought about it recently. I'm like, no, I'm very happy. 01:16:10.233 --> 01:16:13.233 Yeah, same here, actually. It's not that it's bad. 01:16:13.233 --> 01:16:17.233 No, yeah, it's good. Thank you, everybody, members, boosters, 01:16:17.233 --> 01:16:20.293 and streamers. We had 20 of you stream sats. Collectively, you sat streamer 01:16:20.293 --> 01:16:25.165 stacked 25,922 satoshis. Thank you very much. 01:16:25.560 --> 01:16:28.243 Now, since this is all new, what we think we're going to do, 01:16:28.243 --> 01:16:31.473 just to make it one final line number, is we're just going to convert the Fiat 01:16:31.473 --> 01:16:35.283 Boost to sats for the total at the current price values. 01:16:35.283 --> 01:16:37.199 We don't really know the best way to, since it's two different, 01:16:37.535 --> 01:16:39.853 but this is what we're experimenting with right now. We always appreciate your 01:16:39.853 --> 01:16:41.319 feedback because the goal is transparency. 01:16:41.812 --> 01:16:49.324 The total boosted for this episode is 332,982 Satoshis. 01:16:50.984 --> 01:16:55.193 That includes the member boosts and the fiat boosts. 01:16:55.441 --> 01:16:58.333 Thank you, everybody. And, of course, the Satoshi boosts. That's all in one 01:16:58.333 --> 01:16:59.663 final line item right there. 01:16:59.802 --> 01:17:02.923 Yeah, that combined to make a total of 39 boosts. 01:17:02.923 --> 01:17:03.413 Woo! 01:17:04.473 --> 01:17:05.473 And some streams, of course. 01:17:05.473 --> 01:17:09.053 That's awesome, of course. Yeah, 20 of you. Thank you, everybody, 01:17:09.053 --> 01:17:11.903 who supported the show with a boost or a membership. It's kind of neat to bring 01:17:11.903 --> 01:17:15.503 it all together in this segment. Finally, it's feeling like it's really happening. 01:17:15.503 --> 01:17:17.103 Thank you for trying it out. 01:17:17.103 --> 01:17:20.563 Yeah. I mean, this year of shows since the beginning of the year would not be 01:17:20.563 --> 01:17:22.302 possible without your support. We continue. 01:17:22.685 --> 01:17:27.153 And even if at this point the show were sold out going forward, 01:17:27.153 --> 01:17:30.983 the deficit that we have dug ourselves into, we would still absolutely need 01:17:30.983 --> 01:17:32.281 your support to survive. It is. 01:17:32.652 --> 01:17:35.665 And it only compounds as the year goes on. So thank you so much. 01:17:36.014 --> 01:17:38.343 It really does make a difference. And we enjoy hearing from you. 01:17:38.343 --> 01:17:40.540 It's our favorite segment of the show. 01:17:42.781 --> 01:17:46.990 A couple of picks, a couple of picks. I think maybe, Wes, you found Fluxcast 01:17:46.990 --> 01:17:52.060 this week, which lets you stream your Linux desktop to a smart TV that supports 01:17:52.060 --> 01:17:57.660 Miracast, DLNA, WFD, or Chromecast. 01:17:57.660 --> 01:18:01.540 Yeah, that's right. You know, Miracast especially has been one of those things 01:18:01.540 --> 01:18:05.470 that just has floated around forever, but I've never actually been able to take 01:18:05.470 --> 01:18:08.320 advantage of or try. Obviously, I Chromecast all kinds of stuff. 01:18:08.320 --> 01:18:10.733 I have used DLNA on and off over the years. 01:18:11.490 --> 01:18:14.770 And of course, there's various ways to do that. But I thought like a nice little 01:18:14.770 --> 01:18:18.860 focused app that could just let you stream your desktop would be pretty handy. 01:18:18.860 --> 01:18:26.150 It does look really nice. Pretty straightforward. Supports KDE Organome and Wayland. 01:18:26.630 --> 01:18:28.510 So if you're on the Wayland desktops. 01:18:31.030 --> 01:18:33.640 You can also get it. You can just since it's a Python app, so you can set it 01:18:33.640 --> 01:18:35.508 up that way. Or there's an app image. 01:18:35.884 --> 01:18:41.940 Ah, and it's GPL 3.0 as well. fluxcast streams your linux desktop to a tv they 01:18:41.940 --> 01:18:44.098 have a great video demo of it and, 01:18:44.771 --> 01:18:48.940 it's it's almost anticlimactic it's so straightforward and simple right because 01:18:48.940 --> 01:18:52.580 especially if you're using the chromecast combo it just works the tv detects 01:18:52.580 --> 01:18:55.270 it if you're on the same land and you're just streaming your desktop to the 01:18:55.270 --> 01:18:58.877 tv like no big deal uh and they have a video embed up on. 01:18:59.115 --> 01:19:02.420 There are some issues like some samsung tv cast stuff has issues so it. 01:19:02.420 --> 01:19:03.260 May not be perfect for. 01:19:03.260 --> 01:19:06.226 Your exact devices you have to give it a go but hopefully it's easy to try. 01:19:06.377 --> 01:19:08.720 Something tells me the chromecast probably going to work best, 01:19:08.720 --> 01:19:12.860 followed by DLNA, then Miracast, and then it's going to drop off from there 01:19:12.860 --> 01:19:14.060 pretty quick. But that's pretty 01:19:14.060 --> 01:19:16.373 neat to see, and it's nice that it works great with modern desktops. 01:19:17.041 --> 01:19:22.390 So I doubled down again this week. I'd kind of gotten out of the habit of backing 01:19:22.390 --> 01:19:28.230 up my legal Audible purchases. I've had an Audible membership since Audible was a company. And... 01:19:29.934 --> 01:19:33.308 Just kind of always every month buy a couple of books. I'm a pretty big audiobook 01:19:33.308 --> 01:19:35.558 listener, especially on long drives or in the evening when I'm trying to go 01:19:35.558 --> 01:19:36.394 to bed, things like that. 01:19:36.974 --> 01:19:40.828 And it's pretty easy to just fall into the habit of browsing for audiobooks 01:19:40.828 --> 01:19:41.838 with the app. You're like, yeah, 01:19:41.838 --> 01:19:43.812 I'll get that. I'll get that because you want to use up those credits. 01:19:44.393 --> 01:19:47.458 And then just kind of forgetting about them and letting them just sit on the service. 01:19:47.859 --> 01:19:51.382 And I've talked about various ways to extract them off of Audible before. 01:19:51.667 --> 01:19:55.928 But this time around, when I was revisiting the solution, I decided to implement 01:19:55.928 --> 01:20:01.618 Audible CLI, which is a command line interface to manage multiple Audible accounts, 01:20:01.618 --> 01:20:04.488 which is nice because we have multiple in my family, browse and export your 01:20:04.488 --> 01:20:05.749 library and your wish list. 01:20:06.022 --> 01:20:11.258 It'll download the audiobooks in the AAXE or AAX formats and then help convert 01:20:11.258 --> 01:20:12.768 them to non-DRM formats. 01:20:13.122 --> 01:20:16.888 It supports a faster method of browsing and retrieving the inventory than some 01:20:16.888 --> 01:20:18.869 of the other picks that we've talked about in the past. 01:20:19.356 --> 01:20:21.887 It's available for Linux, Mac OS, and Windows. 01:20:22.741 --> 01:20:26.988 And it is GPL3 as well. Audible CLI, boys. 01:20:26.988 --> 01:20:28.058 Yeah, this looks nice. 01:20:28.518 --> 01:20:32.738 I like it a lot. It's clean. It's easy to use. And if you're so inclined, 01:20:32.738 --> 01:20:36.138 it's very straightforward to have an agent sit in front of. 01:20:36.138 --> 01:20:39.133 Yeah, it seems like a well-maintained, up-to-date Python app. 01:20:39.980 --> 01:20:41.138 And using UV again. 01:20:41.718 --> 01:20:42.703 And it's already in Nix. 01:20:42.877 --> 01:20:46.894 Yes, it is. Yes, it is. Which made it very easy for me to get up and going with it. 01:20:47.463 --> 01:20:51.196 And then you could combine it with something like import to audiobookshelf. 01:20:51.637 --> 01:20:55.318 Our buddy Steve Ovens from the Ask Noah podcast has a utility that takes books 01:20:55.318 --> 01:21:00.431 which you legally purchase from Audible and then moves them over to your audiobookshelf instance. 01:21:00.722 --> 01:21:03.538 We've talked about audiobookshelves in the past. Fantastic app. 01:21:04.245 --> 01:21:07.252 And Steve's made a tool to manage that. There's a couple other things out there. 01:21:08.198 --> 01:21:14.688 For me now, what I have is essentially a systemd timer that once a week sweeps Audible. 01:21:15.564 --> 01:21:20.081 looks for books that I've picked up, and if they're missing from my audiobookshelf 01:21:20.081 --> 01:21:24.456 library, it automatically downloads and imports them for me to my library. 01:21:24.858 --> 01:21:26.802 And I can have it do a sweep for my wife as well. 01:21:27.225 --> 01:21:31.621 Well, I've got a sort of just sneaky final pick here. 01:21:31.621 --> 01:21:32.621 Oh, I love it. Sneaking in. 01:21:32.621 --> 01:21:39.734 Yeah, because whomever whiz popped into the Matrix chat to give us the correct link to Quid Clipper. 01:21:39.995 --> 01:21:40.645 Oh, good. 01:21:40.796 --> 01:21:45.051 So I can actually now see Quid Clipper is MIT licensed, Find and cut lossless 01:21:45.051 --> 01:21:47.924 clips from movies and TV by searching the subtitle dialogue. 01:21:48.831 --> 01:21:53.801 There's a command line tool. But then also, like, it looks like there's a whole 01:21:53.801 --> 01:21:55.511 web interface here or an app interface. 01:21:55.651 --> 01:21:56.151 What? 01:21:56.151 --> 01:21:56.748 Wow. 01:21:57.607 --> 01:21:58.919 Somebody's been busy. 01:22:00.144 --> 01:22:05.357 Boy, talk about the pivot. This looks awesome. I'm looking at the interface pictures right now. 01:22:06.541 --> 01:22:10.512 Library browser, dialogue search, folder dialogue search, scrolling script view, 01:22:10.512 --> 01:22:16.582 stream selector, lossless clipping, bookmarks, clips as first class items. 01:22:16.582 --> 01:22:18.772 Batch export. There's NixOS support down here. 01:22:18.772 --> 01:22:20.778 Oh, this is looking real good. 01:22:21.304 --> 01:22:25.831 As well as Docker. Yeah, it really is. And the CLI part as well. 01:22:26.818 --> 01:22:28.472 I think we'll be giving this a try. 01:22:29.032 --> 01:22:33.568 Chris, you've been asking for this once a week for about a year now, I think. 01:22:33.957 --> 01:22:38.072 I think he was reading our gosh darn minds. All right, we'll put a link to that 01:22:38.072 --> 01:22:42.752 in the show notes as well, and you'll find that over at linuxunplugged.com slash 01:22:42.752 --> 01:22:45.176 672. What else should we tell people before we go? 01:22:45.548 --> 01:22:49.913 Maybe some extra details around the show, metadata types, structured information about the episode? 01:22:50.083 --> 01:22:53.852 Oh, yeah, well, like, maybe if you want to try to use this tool on something 01:22:54.192 --> 01:22:56.352 like a podcast, like our podcast, I don't know if it will work, 01:22:56.352 --> 01:23:01.426 but we do provide SRT and VTT files right there in our XML RSS feed, 01:23:01.827 --> 01:23:04.602 as well as our Cloud Chapters JSON, of course. 01:23:04.602 --> 01:23:05.772 That's true. Another pro tip. 01:23:06.092 --> 01:23:10.562 Oh, and there's an MP4 file out there lurking, hidden in the feed. 01:23:10.562 --> 01:23:12.392 We should probably tell people that at the beginning of the episode. 01:23:12.492 --> 01:23:15.792 Yeah, we should. Yes, but there is indeed an MP4 file out there. 01:23:16.187 --> 01:23:18.206 And another pro tip, shows live. 01:23:21.872 --> 01:23:22.888 And that's also in the feed. 01:23:22.996 --> 01:23:26.256 Can you believe it? It is. That's right. We do it Sundays, 10 a.m. 01:23:26.256 --> 01:23:29.316 Pacific, 1 p.m. Eastern. You can get converted to your local time zone over 01:23:29.316 --> 01:23:31.448 at jupyterbroadcasting.com slash calendar. 01:23:31.901 --> 01:23:34.816 We also have the lug going. It's our mumble room. Low latency, 01:23:34.816 --> 01:23:38.083 opus audio. Goes all week long, but popping during the show for sure. 01:23:38.402 --> 01:23:42.477 And then our matrix chats, all of that and more at linuxunplugged.com. 01:23:42.727 --> 01:23:44.457 It's our website with our information on it. 01:23:44.805 --> 01:23:47.882 And then there's a bunch of great shows over at jupyterbroadcasting.com. 01:23:48.172 --> 01:23:50.896 Okay, that's it from us. We'd love to know what you think and any feedback. 01:23:50.896 --> 01:23:54.470 Please do boost it in. Don't forget boost.jupiterbroadcasting.com. 01:23:54.958 --> 01:23:57.913 Thanks so much for joining us on this week's episode of Your Unplugged Program. 01:23:58.174 --> 01:24:01.065 And we'll see you all right back here next week.