WEBVTT 00:00:11.278 --> 00:00:15.935 Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris. 00:00:15.935 --> 00:00:16.585 My name is Wes. 00:00:16.585 --> 00:00:17.495 And my name is Brent. 00:00:18.221 --> 00:00:22.515 Hello, gentlemen. Well, coming up on the episode this week, the kernel release 00:00:22.515 --> 00:00:24.630 that deprecates our childhood has landed. 00:00:24.891 --> 00:00:27.825 There's a lot to like. There's one thing that really stings, 00:00:27.825 --> 00:00:31.805 but 7.3 is looking like such a banger. I'm not sure if I'm even going to care. 00:00:31.805 --> 00:00:32.757 We'll tell you more about that. 00:00:33.018 --> 00:00:36.658 Then, how sync things saved the day this week. 00:00:37.029 --> 00:00:39.328 Plus, we have some great boosts and picks that you're going to love. 00:00:39.984 --> 00:00:42.925 and a lot more. So before we get into all of that, let's say time-appropriate 00:00:42.925 --> 00:00:46.056 greetings to that virtual lug. Hello, my room. 00:00:46.747 --> 00:00:49.015 Hey, Chris. Hello, Brent. 00:00:49.015 --> 00:00:51.843 Hello, and hello up there in the quiet listening. Nice to see you up there. 00:00:52.430 --> 00:00:55.885 You guys looking good up there. Handsome as ever. Go say good morning to our 00:00:55.885 --> 00:01:01.461 friends over at Define Networking as well. Define.net slash unplugged. Go meet Manage Nebula. 00:01:02.198 --> 00:01:06.029 Decentralized VPN built on the open-source Nebula platform that we love and use. 00:01:06.435 --> 00:01:11.045 Nebula is so awesome. Oh, I'm so, so grateful for Nebula, especially as I have 00:01:11.045 --> 00:01:16.475 been having issues with other mesh networks. It really is so superior and simple 00:01:16.475 --> 00:01:20.083 in its way and in its design too, yet extremely secure. 00:01:20.460 --> 00:01:26.172 And I love that I have my own lighthouse, so I completely control the infrastructure. I own everything. 00:01:26.556 --> 00:01:29.835 But like friends and family, I can hook them up with Define Networking's managed 00:01:29.835 --> 00:01:33.045 Nebula. They automatically provision a lighthouse now for people, 00:01:33.045 --> 00:01:34.131 but you can also run your own. 00:01:34.741 --> 00:01:36.871 It's really great. It's a great company. It's a great product. 00:01:37.126 --> 00:01:40.275 It's open source, independently audited, and it's built with modern encryption. 00:01:40.275 --> 00:01:43.901 So go get started. Up to 100 hosts, absolutely free. No credit card required. 00:01:44.290 --> 00:01:47.815 Go to defined.net slash unplugged. Redefine your VPN experience. 00:01:47.815 --> 00:01:52.579 See why I love it. Own your infrastructure. Defined.net slash unplugged. 00:01:52.974 --> 00:01:54.826 Thank you to Defined for sponsoring the Unplugged program. 00:01:57.803 --> 00:02:04.061 Well, with the release of Linux 7.2, it is the end of AppleTalk and it is the 00:02:04.061 --> 00:02:06.331 end of our sticker. We'll have to remove it from the store this week. 00:02:06.331 --> 00:02:10.620 So if you want to get your commemorate sticker, now's the time. 00:02:10.864 --> 00:02:12.901 It's like we're making a coin for AppleTalk, but it's a sticker. 00:02:14.111 --> 00:02:16.913 I'll not dwell on this. The audience knows how I feel about this. 00:02:17.284 --> 00:02:18.132 But get them while you can. 00:02:18.689 --> 00:02:23.721 Also, we wanted to make a mention of the passing of a longtime Samba contributor, 00:02:23.721 --> 00:02:27.007 Steve French. He passed away recently after dealing with an illness. 00:02:27.576 --> 00:02:30.816 He's been around forever. He's been contributing since 2002. 00:02:31.321 --> 00:02:34.391 He worked at Microsoft. He worked at IBM. He's worked in Azure. 00:02:35.001 --> 00:02:37.561 He worked on my favorite project, KSMBD. 00:02:39.105 --> 00:02:43.256 And so he had to step down for health reasons this week, and then he did pass away. 00:02:44.260 --> 00:02:45.141 Our condolences. 00:02:45.141 --> 00:02:51.400 Yeah, really. A longtime contributor. So our thoughts are with his family. 00:02:52.282 --> 00:02:55.972 And they do have a new maintainer stepping up for CIFS. 00:02:56.140 --> 00:03:00.861 So there will be. And I guess this individual has a history with NTFS, the rewrite. 00:03:00.861 --> 00:03:06.531 Yeah, we have Paulo Alcantara and someone we've mentioned multiple times on the show, Namjai Jian. 00:03:07.731 --> 00:03:13.911 Those folks are now co-maintaining CIFS, CIFS, which is Namjai's third file 00:03:13.911 --> 00:03:16.725 system, the NTFS rewrite, ex-fat, and now this. 00:03:17.207 --> 00:03:18.631 So it sounds like it's in pretty good hands. 00:03:19.151 --> 00:03:19.732 It does indeed. 00:03:20.254 --> 00:03:24.151 But we did want to note, because, you know, it's part of the community. 00:03:24.151 --> 00:03:26.936 And I think it's worth observing, and our thoughts are with the family. 00:03:27.522 --> 00:03:30.669 out there, and it is good to see the project continue to get support as well. 00:03:30.773 --> 00:03:32.761 This is how open source gets made. 00:03:32.761 --> 00:03:36.581 Mm-hmm. Mm-hmm. All right. Well, then, let's get into Linux, 00:03:36.581 --> 00:03:38.174 speaking of the kernel, Linux 7.2. 00:03:40.770 --> 00:03:45.307 So just as we finished up the live stream, I think Linus probably got done listening 00:03:45.307 --> 00:03:47.607 to the show, clearly. And then decided to release. 00:03:47.607 --> 00:03:48.597 Oh, better get some work done. 00:03:48.597 --> 00:03:51.777 Yeah. Pushed out Linux 7.2. I don't think he wants to push it out before the 00:03:51.777 --> 00:03:54.463 show's over because he doesn't want to be distracted. I understand that. 00:03:54.660 --> 00:03:55.827 It might distract us, too. 00:03:56.448 --> 00:04:00.857 And this was classified as a very busy kernel cycle release due to increased 00:04:00.857 --> 00:04:03.182 patch and reporting churn due to AI-assisted development. 00:04:04.117 --> 00:04:05.597 We do have a quote from Linus. You want to take it? 00:04:06.143 --> 00:04:10.317 Well, this last week of the release was, once again, bigger than I would have wished for. 00:04:10.657 --> 00:04:14.977 But hey, with the whole quote-unquote new normal thing, if I delayed releases 00:04:14.977 --> 00:04:17.957 for that reason, we'd probably never have a release at all. 00:04:18.477 --> 00:04:22.207 There are a number of fairly late reverts. The DRM scheduling reverts stand 00:04:22.207 --> 00:04:24.341 out, but there's a few other ones in here too. 00:04:24.695 --> 00:04:27.441 It may not be pretty, but it's the correct way to deal with, 00:04:27.836 --> 00:04:32.648 oh, that code wasn't ready and caused problems. People will be trying that again later. 00:04:33.095 --> 00:04:37.106 But while the DRM reverts stand out and are some of the bigger patches in here, 00:04:37.321 --> 00:04:40.457 there's a lot of small fixes all over, mostly in drivers, but networking remained 00:04:40.457 --> 00:04:42.076 fairly active this last week, too. 00:04:42.226 --> 00:04:45.466 And then there are a few random patches to architecture files and the perf core. 00:04:46.052 --> 00:04:49.512 With this, the merge window for 7.3 obviously opens tomorrow, 00:04:49.761 --> 00:04:52.111 and I already have 40 pending pull requests. 00:04:52.326 --> 00:04:55.083 I appreciate all the timely people. You know who you are. 00:04:55.310 --> 00:04:59.507 Yeah, we'll get into 7.3 here in just a moment. You know, when Linus calls out 00:04:59.507 --> 00:05:03.317 a subsystem, it's probably worth paying attention. When he calls out networking 00:05:03.317 --> 00:05:05.317 specifically, it's probably worth paying attention. 00:05:06.257 --> 00:05:09.777 So I think we'll get to that. But the thing that you and I have been watching, And, 00:05:11.313 --> 00:05:13.791 Something that I think we talked about before on the show, so we'll just give 00:05:13.791 --> 00:05:18.807 a brief mention, is there's been some real improvements to, quote-unquote, schedulers. 00:05:19.881 --> 00:05:21.661 CPU schedulers, GPU schedulers. 00:05:21.661 --> 00:05:23.641 The scheduling system, all kinds of stuff. 00:05:23.641 --> 00:05:30.777 Scheduling got a big work, a lot of work in 7.2. So we have a now more cache-aware CPU scheduler. 00:05:31.166 --> 00:05:34.545 Yeah. This is a pretty big deal. 00:05:34.812 --> 00:05:37.731 And it's almost surprising in a way that we haven't had it already, 00:05:37.731 --> 00:05:43.113 right? Like, caches are a huge part of CPU and computer performance, of course. 00:05:43.502 --> 00:05:47.091 And these days, especially on systems with, like, multiple dies, 00:05:47.091 --> 00:05:51.301 right, it's not just one giant CPU, right? It's, like, multiple CPUs, 00:05:51.301 --> 00:05:54.379 each with their own last-level cache right next to them. 00:05:54.658 --> 00:05:58.391 And so understanding, like, where is a process or a thread running, 00:05:58.391 --> 00:06:02.851 and then what caches is it near, and then what data is in those caches can be a big deal. 00:06:02.851 --> 00:06:05.151 So if you think about something like a database, right, maybe you're doing a 00:06:05.151 --> 00:06:08.741 query where you have a lot of working memory that is, the whole point is you 00:06:08.741 --> 00:06:11.691 want a query optimizer that can parallelize things, right? So you're not waiting 00:06:11.691 --> 00:06:13.315 on trying to seriously do stuff. 00:06:13.593 --> 00:06:15.654 So maybe you have a list of things you're trying to look up. 00:06:15.805 --> 00:06:19.561 Well, both subqueries are going to need to be accessing that data, right? 00:06:19.822 --> 00:06:23.299 So if one of the subquery threads is running on a different CPU core, 00:06:23.595 --> 00:06:27.891 suddenly it's got to do the inter-core communication to go get the data from the other cache. 00:06:28.239 --> 00:06:31.391 So you don't really want to do that. So now the goal is upgrade the scheduling 00:06:31.391 --> 00:06:35.701 system so that it can figure out, like it starts tracking stuff in a multiple 00:06:35.701 --> 00:06:38.404 way. So one is just obviously sort of threads of the same thing, right? 00:06:38.607 --> 00:06:42.391 But it can figure out when things that are running need to share the same cache 00:06:42.391 --> 00:06:45.041 and then try to schedule them on the same cores appropriately. 00:06:45.041 --> 00:06:45.474 Woo! 00:06:46.032 --> 00:06:47.396 That's a nice one. 00:06:48.096 --> 00:06:52.026 There's multiple steps to it. They're learning a lot, or they're taking advantage 00:06:52.026 --> 00:06:54.066 of a lot of neat stuff that happens. 00:06:54.066 --> 00:06:57.536 Like the hardware actually has cache monitoring stuff built in. 00:06:57.536 --> 00:07:01.846 So then the first level is basic structural stuff. It understands what the system 00:07:01.846 --> 00:07:05.581 has and the basic schedule threads together. 00:07:05.946 --> 00:07:10.046 But then it can actually go check to see what the layout the hardware tells 00:07:10.046 --> 00:07:13.506 it is doing in real time. And it can even go sample the misses to figure out 00:07:13.506 --> 00:07:17.476 like, oh, we tried to hit something in the cache, but that's actually on the other system. 00:07:17.476 --> 00:07:17.846 Really? 00:07:17.846 --> 00:07:22.486 We might need to start considering if this continues actually moving it to the other core. 00:07:22.486 --> 00:07:22.816 Wow. 00:07:23.116 --> 00:07:25.046 Now you don't want to do that too much, right? Because then you could just be 00:07:25.046 --> 00:07:28.516 bouncing around all the time. That could be a problem. So it's a delicate balance, 00:07:28.516 --> 00:07:30.196 as is everything in scheduling. 00:07:30.890 --> 00:07:34.014 But this is a huge upgrade for just it being really smart. 00:07:34.490 --> 00:07:39.176 Now, it will matter more on, you know, big systems with lots of CPU cores and 00:07:39.176 --> 00:07:43.816 that kind of thing. So you might not see as much of an impact on your single-core laptop or whatever. 00:07:44.356 --> 00:07:44.509 Yeah. 00:07:45.328 --> 00:07:47.783 But it's a big upgrade for our friend, the Colonel. 00:07:47.957 --> 00:07:52.384 Linus had made a comment that it may not be pretty, but it's the correct way 00:07:52.384 --> 00:07:55.494 to deal with, oh, that code wasn't ready and caused problems. 00:07:55.494 --> 00:07:57.175 People will probably try that again later. 00:07:57.698 --> 00:08:00.274 Is he referring to the GPU scheduler stuff that got reverted? 00:08:00.274 --> 00:08:00.884 Yes. 00:08:00.884 --> 00:08:01.544 Yeah, okay. 00:08:01.544 --> 00:08:09.621 7.2 was meant to switch the DRM scheduler from first in, first out to a new CFS-style FAIR policy. 00:08:09.946 --> 00:08:14.166 So we already have this on the CPU side. This was bringing it over to the GPU side. 00:08:15.094 --> 00:08:20.934 But then an RX 9070 XT owner hit a sustained 100% GPU load. 00:08:20.934 --> 00:08:21.304 Uh-oh. 00:08:21.304 --> 00:08:25.894 Yeah, games dropped to 10 FPS. The Wayland session froze. Had to reboot to recover. 00:08:27.074 --> 00:08:29.494 And then here's the part that maybe wasn't as good of an idea. 00:08:30.074 --> 00:08:33.427 When this happened, it wasn't just that the default was swapped. 00:08:33.752 --> 00:08:37.254 It was the only schedule. The other stuff was ripped out, and all you were left 00:08:37.254 --> 00:08:40.806 with was fair. So the first in, first out, and the RR policy, 00:08:41.374 --> 00:08:43.807 those weren't there. There was no runtime escape hatch. 00:08:44.454 --> 00:08:49.584 So five days before the release, someone from Egalia shipped a 20-patch revert, 00:08:49.584 --> 00:08:53.599 smoke-tested on a Steam Deck, and it ships first in, first out. 00:08:53.802 --> 00:08:58.457 As the defaults again, sort of unrips things out, but now it still adds fare 00:08:58.660 --> 00:09:02.823 as an opt-in experiment. So instead of doing a big change to making this default, 00:09:03.674 --> 00:09:05.226 now we're still in sort of test mode. 00:09:05.621 --> 00:09:07.854 All right. That makes sense. Thank you. Thank you. 00:09:08.314 --> 00:09:12.384 What is nice, though, is the reporter, AB, tested some of the fixes himself. 00:09:12.384 --> 00:09:17.643 This is a very active collaboration process, including, quote, 00:09:17.916 --> 00:09:19.309 the patch Claude gave me. 00:09:21.119 --> 00:09:24.524 Also, I think MongoDB users will be happy to hear this. 00:09:24.864 --> 00:09:29.044 There was some work in Linux's memory management. The multi-gen LRU memory reclaim 00:09:29.044 --> 00:09:30.140 system has been cleaned up. 00:09:30.506 --> 00:09:34.662 And as a result, kernel developers have reported performance improvements approaching, 00:09:34.830 --> 00:09:40.415 so, you know, results will vary, but approaching 30% in some MongoDB workloads, 00:09:40.984 --> 00:09:43.804 along with fewer unexpected out-of-memory situations. 00:09:43.804 --> 00:09:48.111 Ooh, I like that one. You know, some of us have a lot of tabs, and it gets dicey. 00:09:51.164 --> 00:09:55.181 I, my, my bet is not everybody's going to see that 30%, but heck, 00:09:55.565 --> 00:09:59.785 even a 10% is pretty great. Just by a better memory reclaim. 00:10:00.400 --> 00:10:03.244 And Brent, I know you're going to love to see this one. Faster, 00:10:03.584 --> 00:10:05.824 leaner swap in Linux now. 00:10:06.524 --> 00:10:07.064 Delicious. 00:10:08.243 --> 00:10:09.839 Are you guys still using the swap file? 00:10:10.390 --> 00:10:13.990 Well, I have one. I don't know that it gets used a ton. 00:10:14.297 --> 00:10:18.514 That's what I mean, really, is are you actually, uh, like when you install your 00:10:18.514 --> 00:10:23.173 Linux box, are you deploying a swap file? Not are you actively using it necessarily, but... 00:10:23.423 --> 00:10:26.624 I mean, any system that runs for a really long time generally swaps a few things out. 00:10:26.624 --> 00:10:31.654 Yeah. Yeah, I have 3.7 gigs used, 12 gigs free in MySwap right now. 00:10:32.194 --> 00:10:35.054 MySwap is zero currently, which is, I think, a little unusual, 00:10:35.054 --> 00:10:37.084 but I use it heavily for hibernation. 00:10:37.404 --> 00:10:40.489 Like, sometimes when I'm traveling with a laptop and stuff, I want to maintain, 00:10:40.982 --> 00:10:43.094 as much battery power as possible depending on what I'm doing. 00:10:43.094 --> 00:10:45.764 Maybe I'm going on a train at some point of the trip or something like that, 00:10:45.764 --> 00:10:51.082 And so having hibernate available for me is quite valuable. So that's the main reason I have it. 00:10:52.024 --> 00:10:55.104 Yeah. That's one of the reasons I have one too. Uh, I also find it better for 00:10:55.104 --> 00:10:57.444 crash debugging and also, um, 00:10:58.856 --> 00:11:02.702 Some of my systems only have 16 gigs of rewrite, so I do actually get the swap out. 00:11:04.642 --> 00:11:08.905 So it's good to see that there is ongoing rewrite of the swap table infrastructure in Linux 7.2. 00:11:09.386 --> 00:11:12.612 The improvements reduce the amount of static metadata Linux needs to keep around 00:11:12.612 --> 00:11:16.212 for swap and improves how that metadata itself is managed. 00:11:16.212 --> 00:11:21.092 Nice. Yeah, one example here. A system with a 1TB swap device can save roughly 00:11:21.092 --> 00:11:25.263 512 MB of RAM just from the reduced metadata overhead. 00:11:26.111 --> 00:11:29.564 Which is, you know, nice. It'll add up over time. 00:11:30.656 --> 00:11:33.372 That's a massive savings with how expensive RAM is right now. 00:11:33.372 --> 00:11:36.102 That's no joke. So again, a system with one terabyte swap device, 00:11:36.102 --> 00:11:37.355 which is a pretty big swap device. 00:11:38.980 --> 00:11:43.744 will save roughly 512 megabytes of RAM just from the reduced metadata overhead. 00:11:45.247 --> 00:11:48.736 Does anybody out there in the audience have a terabyte swap device? 00:11:48.980 --> 00:11:50.190 Boost.Jupyter Broadcasting. 00:11:50.190 --> 00:11:50.660 Yeah, right. 00:11:50.660 --> 00:11:52.340 You've got to tell them. And why? 00:11:52.340 --> 00:11:52.970 How'd you get there? 00:11:52.970 --> 00:11:57.119 Mm-hmm, mm-hmm, mm-hmm. All right, we've got a couple other items. 00:11:57.293 --> 00:12:00.990 There's one I really want to talk about. Is there a couple before we get to 00:12:00.990 --> 00:12:04.800 the file system stuff that you want to talk about? I know we have faster pipes. 00:12:06.500 --> 00:12:11.400 Yeah, we do have faster pipes. We covered a bit about this in 6.7.2. 00:12:11.400 --> 00:12:16.040 There was basically going to be improvements to make piping things between processes 00:12:16.040 --> 00:12:18.649 faster, kind of all around, especially on the right path. 00:12:19.021 --> 00:12:22.254 And now we have a few numbers under memory pressure in particular. 00:12:22.880 --> 00:12:28.800 Throughput is up 21 to 48% and right latency is down 17 to 33%. 00:12:28.800 --> 00:12:30.630 Whoa, this is a, yeah. 00:12:30.630 --> 00:12:35.390 Yeah, and I mean like pipes and, you know, FIFOs are used all over the place. 00:12:35.400 --> 00:12:35.989 Yeah, yeah. 00:12:36.134 --> 00:12:40.120 I find it so crazy how, you know, Linux isn't just a baby anymore, 00:12:40.120 --> 00:12:46.250 and yet we're getting insane numbers of performance upgrades in seemingly every 00:12:46.250 --> 00:12:48.782 single version. It's just bonkers. 00:12:49.223 --> 00:12:52.920 It really is like, I wonder, like, was this all just there for the taking all 00:12:52.920 --> 00:12:54.587 along? And you just, you work through it, I suppose. 00:12:55.579 --> 00:12:58.970 Brent, I'd like to see you test this next one somehow. USB 4.0 stream. 00:12:58.970 --> 00:13:03.640 So in 7.2, you now, two machines can send raw streams of data directly over 00:13:03.640 --> 00:13:05.378 USB 4.0 or a Thunderbolt cable. 00:13:05.523 --> 00:13:07.526 Well, who needs Ethernet anymore, I say? 00:13:08.042 --> 00:13:11.261 Well, it is actually pretty fast. It is quite fast, actually. 00:13:11.261 --> 00:13:14.051 It's neat, too, because before, you know, you could obviously do this and then, 00:13:14.051 --> 00:13:17.561 like, stand up your own little local network or whatever and set up IPs and, 00:13:17.561 --> 00:13:19.251 you know, get it going. You could do it. 00:13:19.448 --> 00:13:23.611 But this, you just get, like, a slash dev slash TB stream zero device. 00:13:24.051 --> 00:13:28.331 And you can just cat and read to write or DD or whatever, and it just shoots it across the USB cable. 00:13:28.891 --> 00:13:31.171 It's going to be so fast. That's going to be, you know, because you don't have 00:13:31.171 --> 00:13:32.541 the overhead of the protocol of Ethernet. 00:13:32.541 --> 00:13:34.011 Nice. Oh, my goodness. 00:13:34.011 --> 00:13:38.361 And it can also still coexist with normal Thunderbolt networking. 00:13:38.361 --> 00:13:40.305 So it's not like it's the only thing you can do on the cable. 00:13:40.815 --> 00:13:43.891 Okay. All right. Well, while we're talking about performance, 00:13:43.891 --> 00:13:47.241 I think the thing that I'm probably the most excited about because it'll have 00:13:47.241 --> 00:13:52.661 the biggest practical impact for all of us is ButterFS is getting faster in 00:13:52.661 --> 00:13:55.136 this release. ButterFS is getting faster. 00:13:55.636 --> 00:13:58.491 We talked about large folios in the last episode. 00:13:58.491 --> 00:14:04.186 They're enabled by default now. Yeah. There's also experimental support for huge folios up to 2 megs. 00:14:04.302 --> 00:14:07.561 But I think the big wins come from better direct I.O. performance, 00:14:07.561 --> 00:14:11.361 faster sequential writes, and a new interface that exposes the raw checksums 00:14:11.361 --> 00:14:14.251 to user space, so that way you can build tooling around that and all of that. 00:14:14.251 --> 00:14:18.304 Yeah, including things like perhaps deduplication, Brad. Oh, hello. 00:14:18.751 --> 00:14:22.426 So, you know, this is the beginning of an API, right? This is the beginning of an API. 00:14:23.231 --> 00:14:27.521 But the direct I.O. performance and the sequential writes are resulting in real 00:14:27.521 --> 00:14:29.011 performance improvements for ButterFS. Amazing. 00:14:29.431 --> 00:14:31.905 I think we'll probably see Laravel benchmark that pretty soon, 00:14:32.300 --> 00:14:34.831 and I'm going to keep an eye out for that. So I think that's probably the thing 00:14:34.831 --> 00:14:36.420 in 7.2 I'm the most excited about. 00:14:36.774 --> 00:14:40.651 So people upgrading soon on rolling releases, you know, do they? 00:14:40.651 --> 00:14:43.119 I don't know. You probably have to do something to upgrade your ButterFS format. I don't know. 00:14:43.241 --> 00:14:46.347 But it's something to keep an eye on. Something I'll probably play with, and I'll update you. 00:14:46.985 --> 00:14:49.870 There's one here I'm excited about, inline storage encryption. 00:14:51.001 --> 00:14:55.691 There's a new device mapper target called DMInlineCrypt. 00:14:55.691 --> 00:14:59.411 And it's designed to take advantage of hardware that can perform inline block 00:14:59.411 --> 00:15:04.963 device encryption rather than doing all of that encryption through the traditional DM crypt path. 00:15:05.916 --> 00:15:08.865 Based partly on work that happened in Android, of all things, 00:15:08.865 --> 00:15:12.965 and could eventually become a useful alternative to DMCrypt on hardware with 00:15:12.965 --> 00:15:16.882 dedicated encryption support. You know how I like that encryption. You make sense on it. 00:15:17.505 --> 00:15:21.605 On mobile, you probably do have the opportunity to have custom stuff and not 00:15:21.605 --> 00:15:22.919 as much CPU as you might want. 00:15:23.151 --> 00:15:26.975 And they have been tested in the fire. Android did not start super strong in 00:15:26.975 --> 00:15:28.788 this area, and they've made a lot of progress. 00:15:28.945 --> 00:15:32.705 And, hey, look, it turns out that having Linux on your phone actually pays dividends 00:15:32.705 --> 00:15:36.228 that the whole kernel could use, proving out the entire model that we've been saying all along. 00:15:36.390 --> 00:15:38.305 It's also worth pointing out, not that you guys don't know this, 00:15:38.305 --> 00:15:41.665 but it's worth just reminding us all that these tools like this, 00:15:41.665 --> 00:15:44.651 if they're being actively supported and contributed to by a large, 00:15:44.883 --> 00:15:46.285 well-financed group of people, 00:15:46.925 --> 00:15:51.085 it's probably something to consider just because you want tools in this category 00:15:51.085 --> 00:15:52.871 that are being actively supported and maintained. 00:15:53.862 --> 00:15:56.419 Um, there's a few other things in here, but the one that really jumped out to 00:15:56.419 --> 00:16:00.765 me is faster proc. Like when we're talking about, this is really the performance kernel. 00:16:01.619 --> 00:16:02.719 Yeah, it really is, right? 00:16:02.719 --> 00:16:05.909 This is silly, but like, what? Faster slash proc? What? 00:16:05.909 --> 00:16:07.133 I didn't know that was an issue. 00:16:07.809 --> 00:16:08.089 Yeah. 00:16:08.089 --> 00:16:12.029 Proc exposes a whole bunch of metadata and info about like the running kernel 00:16:12.029 --> 00:16:16.639 and all that. And there's, um, there's a bunch of programs, especially command 00:16:16.639 --> 00:16:19.294 line programs that want to go read that stuff to actually go. 00:16:19.492 --> 00:16:22.209 Right. Cause it's a way that you don't have to have use this calls to try to 00:16:22.209 --> 00:16:24.829 get the same information. you can get it right out of the file system. 00:16:24.829 --> 00:16:27.154 It's super handy. It's standard Unix. Yeah, right? 00:16:28.455 --> 00:16:32.379 So there's a bunch of stuff that uses it. And so if you make that faster, 00:16:32.379 --> 00:16:33.529 you make all those things faster. 00:16:33.529 --> 00:16:37.619 Yeah, absolutely. It's like Brent said. It's like we're just seeing these optimizations 00:16:37.619 --> 00:16:38.729 in areas after all these years. 00:16:39.429 --> 00:16:42.838 But who thinks Steven go down there and be like, huh, what can I do here? 00:16:42.960 --> 00:16:46.819 A clanker? I don't know. Maybe a clanker? I don't know. Some autist, 00:16:46.819 --> 00:16:48.789 I don't know. Maybe like somebody who's really passionate? 00:16:48.789 --> 00:16:51.779 I don't know. I've got a few Pats and such rock I can tell you guys about. 00:16:51.779 --> 00:16:52.109 There you go. 00:16:52.209 --> 00:16:53.269 Uh-huh, uh-huh. 00:16:54.749 --> 00:16:59.009 Well, we should probably mention we now have the safer open at command because 00:16:59.009 --> 00:17:00.279 we have a sticker based on it. 00:17:00.279 --> 00:17:00.719 Right. 00:17:00.719 --> 00:17:02.959 Or we only open it if it's a regular file. 00:17:02.959 --> 00:17:06.249 That's right. Only regular files here. Only regular files. 00:17:06.249 --> 00:17:10.029 And this has shown up in a few different vulnerabilities. This prevents an application 00:17:10.029 --> 00:17:14.159 from unexpectedly being redirected to something like a pipe or a device node 00:17:14.159 --> 00:17:15.880 when what you want is like a text file. 00:17:16.541 --> 00:17:18.519 Yeah. I think that's a good thing 00:17:18.519 --> 00:17:21.809 to call out. There's so many things we could get into here with this. 00:17:22.129 --> 00:17:27.049 It's like this kernel isn't a headliner banger feature. This is the kernel feature. 00:17:27.049 --> 00:17:32.679 It's all these things from like CPU schedulers to GPU schedulers to file systems 00:17:32.679 --> 00:17:35.296 to drivers and networking that have seen improvements. 00:17:36.312 --> 00:17:39.209 Or just new little features like the fact that if you had Thunderbolt or USB 00:17:39.209 --> 00:17:42.808 4, you can now plug machines directly into them. Like that just kind of got a little bit better. 00:17:43.336 --> 00:17:45.879 So it's, I mean, 7.2 is a really good release. 00:17:47.049 --> 00:17:49.977 Is there any word if there's LTS talk around this one? 00:17:50.959 --> 00:17:55.694 I don't know. That's a good question. I've seen discussion around a few possibilities 00:17:55.694 --> 00:17:59.484 from this one, or maybe 7.3 or 7.4, perhaps. 00:17:59.484 --> 00:18:02.654 Well, 7.3 would be a wild one to make an LTS. 00:18:02.654 --> 00:18:03.774 There's a lot going on. 00:18:03.774 --> 00:18:06.434 Let's just talk about this really quick, because it's an early development. 00:18:06.434 --> 00:18:08.995 The merge window just opened last Sunday, obviously. 00:18:09.489 --> 00:18:12.824 But there's a lot more scheduler work in here. A lot more scheduler work. 00:18:13.564 --> 00:18:19.474 Yeah. Right now, there's some focus, particularly on how scheduling interacts 00:18:19.474 --> 00:18:23.554 with C groups, because, of course, the control group mechanism is involved in 00:18:23.554 --> 00:18:27.174 maintaining a lot of access to different resources, which impacts how you want to schedule things. 00:18:27.174 --> 00:18:30.873 So instead of having to traverse a bunch of hierarchical run queues, 00:18:31.895 --> 00:18:36.475 it's moving to a single run queue with new methods for distributing C group weights. 00:18:36.695 --> 00:18:44.544 there's also a new c group mode knob um you can do smp up max concur tasks uh 00:18:44.544 --> 00:18:48.724 concurs the new default and this is basically a mode you can choose so that you pick how, 00:18:49.264 --> 00:18:52.824 how much cpu weight right like how much your share of the cpu is going to be 00:18:53.224 --> 00:18:56.324 gets but divided between the different okay that seems great, 00:18:57.094 --> 00:18:59.794 especially if you care a lot about you know scheduling things and especially 00:18:59.794 --> 00:19:01.624 c groups are used a lot in our buddy containers. 00:19:02.424 --> 00:19:03.244 Yes in our friend. 00:19:04.184 --> 00:19:07.634 Yeah the hierarchical version had a lot of long-standing oddities, 00:19:07.634 --> 00:19:12.714 including how nice, you know, the nice command priorities got handled across 00:19:12.714 --> 00:19:14.594 groups. So that might improve from some of this work, too. 00:19:14.594 --> 00:19:15.084 Oh, okay. 00:19:15.084 --> 00:19:20.114 Yeah. Early testing on an old Sandy Bridge machine paired with an RX 580 showed 00:19:20.114 --> 00:19:21.074 enough improvement that the 00:19:21.074 --> 00:19:24.364 work has been pitched partly as a win for gaming on lower-end hardware. 00:19:24.364 --> 00:19:25.074 Nice. 00:19:25.074 --> 00:19:28.174 So you might not start thinking about gaming on lower-end hardware, 00:19:28.174 --> 00:19:29.854 but that's where it ends up with the kernel sometimes. 00:19:29.854 --> 00:19:32.444 Well, I mean, a lot of those handheld devices, those little built, 00:19:32.444 --> 00:19:35.905 you know, purpose-built, run on Linux. Makes sense. Hmm. 00:19:36.456 --> 00:19:41.919 Okay. Now this next one, I can't even believe. So I'm going to say it out loud, 00:19:42.941 --> 00:19:44.184 but I still don't believe it. Okay. 00:19:46.029 --> 00:19:51.938 In Linux 7.3, in speaking of schedulers, there appears to be up to a 90%, 00:19:52.083 --> 00:19:57.408 I'm going to say that figure again, around 90% reduction in a nasty source of 00:19:57.408 --> 00:20:00.042 x86 scheduling latency. 00:20:00.836 --> 00:20:04.471 Here's what strikes me about this. Number one, 90% is a big number. 00:20:04.721 --> 00:20:10.708 Number two, x86 is a big platform. This isn't ARM. This isn't like some sort 00:20:10.708 --> 00:20:13.161 of maybe risk chip that nobody's using. 00:20:13.835 --> 00:20:19.187 This is x86. What? A 90% reduction in a nasty source of scheduling latency? 00:20:19.791 --> 00:20:22.352 Now, I realize this must be an edge case here or something, right? 00:20:22.352 --> 00:20:23.402 Because this is massive. 00:20:23.862 --> 00:20:26.162 Yeah, and part of it is just sort of how we've been doing things. 00:20:26.162 --> 00:20:30.642 So, like, back to, you know, you have multiple cores and you have multiple caches. 00:20:30.642 --> 00:20:33.352 And one of those things is the page cache, right? Sort of like how do you have 00:20:33.352 --> 00:20:37.151 the files, what you've copied off the disk to memory for working with. 00:20:37.862 --> 00:20:42.732 Well, those different page caches can get out of sync. And so you need to maintain 00:20:42.732 --> 00:20:44.471 coherence across all of these things. 00:20:44.837 --> 00:20:49.202 So you do what's called a TLB shootdown, which is where the one that figures 00:20:49.202 --> 00:20:52.632 out that something needs to be invalidated sends out interprocessor interrupts 00:20:52.632 --> 00:20:55.912 to all the other cores to say like, hey, you got to take this out of your cache 00:20:55.912 --> 00:20:57.462 and go get it fresh because it's no longer valid. 00:20:58.159 --> 00:21:03.812 Well, the calling CPU can wind up sitting with preemption disabled while waiting 00:21:03.812 --> 00:21:06.153 for the remote CPUs to finish that operation. 00:21:06.350 --> 00:21:10.562 And so ByteDance actually observed weights causing up to 16 milliseconds of 00:21:10.562 --> 00:21:13.972 scheduling latency on a 16-core system just from this. 00:21:14.662 --> 00:21:20.861 So Linux 7.3 changes that mechanism so CPUs can now be preempted while waiting in these situations. 00:21:21.210 --> 00:21:25.511 And in ByteDance further testing, the P99 maximum preemption-disabled interval 00:21:25.639 --> 00:21:30.922 fell to roughly 1.5 milliseconds, which the patch series described as producing 00:21:30.922 --> 00:21:35.202 as much as a 90% latency reduction. So there's your 90% in the affected workload. 00:21:36.142 --> 00:21:41.172 With one particular workload, which may not be yours, but it is aimed in particular 00:21:41.172 --> 00:21:45.620 at latency-sensitive and real-time style workloads, which you know Linux is doing more and more of. 00:21:45.887 --> 00:21:50.757 And it removes an architectural source of scheduler stalls that can affect ordinary machines, too. 00:21:51.059 --> 00:21:55.492 So you heard Wes there, too. He mentioned this is patchwork from ByteDance, 00:21:55.492 --> 00:22:01.572 and this is now merged for Linux 7.3 after 10 rounds of patch reviews. 00:22:02.036 --> 00:22:05.182 But I don't think this is the only set of patches that ByteDance has committed 00:22:05.182 --> 00:22:09.867 recently. It seems like their contribution to the kernel is kicking up. And is ByteDance... 00:22:10.947 --> 00:22:12.358 Are they the TikTok company? 00:22:12.561 --> 00:22:12.803 Yeah. 00:22:12.803 --> 00:22:16.053 Right? The TikTok company? What? 00:22:16.393 --> 00:22:20.563 I mean, you know, we've said it in the cliche way of, right, 00:22:20.563 --> 00:22:23.723 like, I don't know, meta runs Linux servers or whatever. But I think there's 00:22:23.723 --> 00:22:27.263 a real component where, like, if you're trying to do media distribution at scale, 00:22:27.263 --> 00:22:29.053 you really push what computers can do. 00:22:29.053 --> 00:22:29.458 Yeah. 00:22:29.574 --> 00:22:33.123 And that means you might need to interface with the kernel. Whether you like 00:22:33.123 --> 00:22:34.343 the content or not is up to you. 00:22:34.343 --> 00:22:37.840 So this is the world we live in now. And, I mean, that's a serious contribution. 00:22:37.980 --> 00:22:39.251 That's a real contribution. 00:22:40.093 --> 00:22:43.503 Something we've talked about on and off for a little while now, 00:22:43.503 --> 00:22:48.440 but really seems like it's coming together in Linux 7.3, is this subscheduler idea. 00:22:48.887 --> 00:22:49.723 So cool. 00:22:49.723 --> 00:22:55.603 So cool. So I think this could be one of the more consequential long-term changes 00:22:55.603 --> 00:22:57.612 because you're going to see vendors take advantage of this. 00:22:58.533 --> 00:23:02.203 Network vendors, hardware vendors. So they have now, I guess, 00:23:02.203 --> 00:23:07.101 like a subscheduler system where you can build your own subscheduler for Linux. 00:23:07.101 --> 00:23:11.703 Yeah. Yeah, so you have a root scheduler, and that can now give a C-group subtree, 00:23:11.703 --> 00:23:17.143 back to the C-groups, to another scheduler, including with revocable access 00:23:17.143 --> 00:23:19.844 to individual CPUs so you can control what it's scheduling over. 00:23:20.227 --> 00:23:23.073 So in practical terms, you could imagine different scheduling policies for a 00:23:23.073 --> 00:23:25.492 few different workloads on the same machine, right? 00:23:25.777 --> 00:23:28.673 So maybe you're trying to do latency-sensitive audio capturing, 00:23:29.013 --> 00:23:31.923 but you've also got a background job doing some transcription in bulk, right? 00:23:31.923 --> 00:23:33.803 Or syncing, or whatever you just have running. 00:23:33.803 --> 00:23:36.763 Or you're trying to do games, but you've got containers that are running that 00:23:36.763 --> 00:23:38.743 are doing image processing at the same time. 00:23:38.743 --> 00:23:39.293 As one does. 00:23:39.953 --> 00:23:42.913 Or I could see this a lot, right? You might have, you're running a web server, 00:23:42.913 --> 00:23:45.913 right? And you might want a particular scheduler for your database that's different 00:23:45.913 --> 00:23:47.983 from how you want to run it for the Nginx reverse proxy. 00:23:47.983 --> 00:23:51.703 Or, you know, to bring it to the desktop, my kid, right? Constantly, 00:23:51.703 --> 00:23:55.583 he's got the Discord going with like 15 people in the call. He's got the game 00:23:55.583 --> 00:23:58.586 going. He's got OBS recording and streaming to the call. 00:23:59.062 --> 00:24:01.646 And, you know, and then you want the game to perform well at the same time. 00:24:02.006 --> 00:24:05.433 And like that, that seriously taxes the system when you're doing all those things. 00:24:05.933 --> 00:24:10.121 So you could see a world where there's a SteamOS scheduler, subscheduler. 00:24:10.880 --> 00:24:13.107 And this sort of scales with how much investment you want to put, 00:24:13.107 --> 00:24:17.457 right? Like with this SketchX stuff, that of course is built on being able to 00:24:17.457 --> 00:24:21.937 write like BPF programs that can do scheduling, which is furthermore means you 00:24:21.937 --> 00:24:23.047 can take a standard Linux kernel. 00:24:23.047 --> 00:24:25.587 You don't have to build your own modular. You don't have to do potentially dangerous 00:24:25.587 --> 00:24:30.180 and, you know, require lots of testing changes on general scheduler algorithm changes. 00:24:30.516 --> 00:24:35.437 You get to implement your own, just, you know, adding it in from user space 00:24:35.437 --> 00:24:39.659 and then get that attached to particular programs and workloads that you need to run. 00:24:39.810 --> 00:24:40.077 Yeah. 00:24:40.456 --> 00:24:41.013 That's powerful. 00:24:41.354 --> 00:24:46.789 What other OS is doing that? All right. So, Brent, you know, how much would you pay? 00:24:47.235 --> 00:24:50.677 We told you you're going to get performance updates to ButterFS and Linux 7.2. 00:24:50.677 --> 00:24:51.127 Delicious. 00:24:51.127 --> 00:24:56.187 But how much would you pay for a whole other round of large performance wins 00:24:56.187 --> 00:24:58.091 in ButterFS and Linux 7.3? 00:24:58.352 --> 00:25:01.227 Because, buddy, I'm going to give it to you for the price of free. 00:25:01.227 --> 00:25:01.607 Yeah. 00:25:02.927 --> 00:25:07.132 Another round of ButterFS optimizations will be coming in Linux 7.3, 00:25:07.132 --> 00:25:08.575 and some of them are kind of a big deal. 00:25:10.212 --> 00:25:13.392 I mean, I'm sure you went through this, Wes, but DirectIO, again, gets some attention. 00:25:14.212 --> 00:25:18.902 ButterFS can now use IOMAP to bounce buffers instead of failing back to a buffered 00:25:18.902 --> 00:25:23.857 I.O. Don't know what that means at all, but it seems good. 00:25:24.211 --> 00:25:27.192 Yeah, well, it's, you know, buffered I.O. has buffers, right? 00:25:27.192 --> 00:25:29.112 I mean, I know that. I know what a buffer is. 00:25:29.112 --> 00:25:33.812 Well, and so, and DirectIO is unbuffered, right? And so now it's basically a 00:25:33.812 --> 00:25:36.482 setup where in some situations, direct IO didn't really work. 00:25:36.482 --> 00:25:38.944 You had to fall back to using the slower buffered stuff. 00:25:39.281 --> 00:25:44.052 Now, in that same situation, Butterfest is internally using an IO map bounce 00:25:44.052 --> 00:25:47.135 buffer of its own that lets it stay on the more direct IO path. 00:25:47.292 --> 00:25:49.632 But gives like direct IO a little buff if it needs it? 00:25:49.632 --> 00:25:50.082 I think so. 00:25:50.082 --> 00:25:52.569 Huh. Okay, thank you. All right. See, now I get it. 00:25:53.512 --> 00:25:57.932 And that's why the case that was sort of the study for this reportedly goes 00:25:57.932 --> 00:26:02.132 from roughly 50% of theoretical throughput up to about 90%. 00:26:02.132 --> 00:26:02.559 What? 00:26:02.664 --> 00:26:03.162 I'm sorry? 00:26:03.162 --> 00:26:04.527 That's massive again. 00:26:05.272 --> 00:26:06.408 So a 45% increase. 00:26:06.744 --> 00:26:09.421 Are they using the same math that we're using? 00:26:10.129 --> 00:26:13.572 This is really something. I mean, just in the next couple releases, 00:26:13.572 --> 00:26:15.422 there are so many performance wins. 00:26:15.782 --> 00:26:22.510 Yeah, really. I mean, there's more removing an unnecessary one Jiffy delay in a non-SSD logging path. 00:26:22.666 --> 00:26:26.902 That's like a 5x throughput improvement in one test. And then another is interesting. 00:26:26.902 --> 00:26:30.972 it replaces the use of an x-array data structure with a local least recently 00:26:30.972 --> 00:26:36.279 used lru list for inhibited extent buffers that one's a 3x improvement. 00:26:36.772 --> 00:26:40.598 All right i mean it's funny that a three it says three to five x by the way 00:26:40.940 --> 00:26:46.641 but it's funny that a 3x improvement seems kind of tame at this point i know just so great still, 00:26:47.471 --> 00:26:51.929 ah i mean like where you at butter fs haters time to update time to update, 00:26:53.236 --> 00:26:55.550 All right, there's several other things in here we could talk about. 00:26:55.640 --> 00:26:58.222 Yeah, how do you like the name FailFS? 00:26:59.407 --> 00:27:00.010 I love it. 00:27:00.010 --> 00:27:02.360 Seems like one of our challenges is coming up. 00:27:02.360 --> 00:27:06.400 I feel an episode already. I hear the name, and I'm already picturing the episode. 00:27:06.400 --> 00:27:09.010 Yeah, it's great, right? So you know what you love about file systems? 00:27:09.010 --> 00:27:11.320 They let you open and write to files and do stuff with files, right? 00:27:11.320 --> 00:27:12.050 Definitely, very handy. 00:27:12.050 --> 00:27:15.579 Yeah, well, FailFS is a file system where every possible operation fails. 00:27:16.229 --> 00:27:16.630 What? 00:27:16.630 --> 00:27:17.860 Okay. 00:27:18.440 --> 00:27:19.940 All right, all right, cool. 00:27:19.940 --> 00:27:23.630 So what's cool about this is it's a security idea, right? So what you do is 00:27:23.630 --> 00:27:27.467 you open a file and you get a file descriptor, which is basically like a handle to the file, right? 00:27:27.820 --> 00:27:30.730 And after that, like you've done the security checks, the kernel scene, 00:27:30.730 --> 00:27:33.180 if you have permissions, all of that stuff has already happened. 00:27:33.580 --> 00:27:36.250 And then you have a handle you can use to read and write from the file. 00:27:36.656 --> 00:27:40.168 What FailFS lets you do for a security system perspective is you can, 00:27:40.324 --> 00:27:43.559 if you know all of the files that something needs to access, 00:27:43.820 --> 00:27:48.487 you pre-open them and then you launch something and you give it a FailFS file system. 00:27:48.701 --> 00:27:52.110 And that means there's no way to ever escape its sandbox. All it can ever use 00:27:52.110 --> 00:27:55.104 is whatever FDs you pass it when you spawn it. 00:27:56.120 --> 00:27:59.121 So it's an idea to pre-restrict it. So it's not in a true root. 00:27:59.260 --> 00:28:03.290 It's not stuck in some folder. It's not in a weird mount. It's just the only 00:28:03.290 --> 00:28:06.419 file system it knows about is not the root file system at all. It's fail.fs. 00:28:06.661 --> 00:28:07.899 That's so funny. 00:28:08.954 --> 00:28:10.921 It's such a funny way to solve the problem. 00:28:11.421 --> 00:28:14.141 It's interesting because it builds on a few, I think in 6.7.2 maybe, 00:28:14.141 --> 00:28:18.291 we talked about null FS, which was something cleaning up sort of early pivot 00:28:18.291 --> 00:28:21.611 routes, switch routes stuff, where you had like a, just a super minimal file 00:28:21.611 --> 00:28:24.791 system that can be the root of the whole tree of file systems. 00:28:24.791 --> 00:28:25.071 Right. 00:28:25.071 --> 00:28:28.331 So fail FS is kind of a similar, like not really quite a, it's like more of 00:28:28.331 --> 00:28:31.599 an abstract file system, but they play important roles in the kernel. 00:28:31.965 --> 00:28:33.281 That's a good one. 00:28:33.281 --> 00:28:35.521 We've been talking about file systems a lot this year because there's been a 00:28:35.521 --> 00:28:40.061 lot of really great news. And I've got to say, FailFS seems like the easiest, 00:28:40.521 --> 00:28:46.631 file system to implement. You just make a lot of mistakes for every single implementation of a file. 00:28:46.631 --> 00:28:47.301 Easy testing. 00:28:47.301 --> 00:28:48.101 Exactly. 00:28:48.101 --> 00:28:51.116 In this case, if you don't implement a bug, it's itself a bug. 00:28:51.586 --> 00:29:00.151 I see BPF is getting some new fancy BPF-MT-MISC. Is this a mount? 00:29:00.151 --> 00:29:03.541 No. Okay, so it's bin format MISC, which is maybe you're familiar. 00:29:03.541 --> 00:29:07.371 their bin format is what lets you basically register different types of binary 00:29:07.371 --> 00:29:10.021 files. Like you can do this with an app image, right? Or you could do it with, 00:29:10.682 --> 00:29:14.061 a wine or you do it with like a jar file to launch a Java application. 00:29:14.357 --> 00:29:17.648 And so this is where the kernel looks at the magic bytes at the start of a file. 00:29:18.245 --> 00:29:25.271 That helps it determine what that is. And then you can register executions or run times for that. 00:29:25.431 --> 00:29:25.831 For BPF. 00:29:26.211 --> 00:29:29.391 Well, so there's two pieces. There's actually a NixOS angle here. 00:29:30.111 --> 00:29:33.151 Oh. Uh-huh, yeah. So right now it's all fixed, right? And maybe you know with 00:29:33.151 --> 00:29:39.770 NixOS, you have to embed the dynamic library interpreter, the LD, 00:29:40.165 --> 00:29:43.681 in basically it has to hard code that in all of the binaries it produces. 00:29:43.681 --> 00:29:46.191 And this is where the patch elf stuff comes in and all that kind of stuff. 00:29:46.191 --> 00:29:50.491 And it means you can't really relocate that stuff. It needs this fixed, hard-coded path. 00:29:51.107 --> 00:29:55.351 What's coming in 7.3 is a bunch of that's being made more dynamic. 00:29:55.351 --> 00:29:58.108 And in particular, you could implement an algorithm that was like, 00:29:58.427 --> 00:30:03.741 hey, wherever I am, go look up in like dot dot lib from me and just find the 00:30:03.741 --> 00:30:04.391 interpreter right there. 00:30:04.591 --> 00:30:05.031 Oh, that would be really nice. 00:30:05.031 --> 00:30:08.401 So then you could start shipping the interpreter that you need to run sort of 00:30:08.401 --> 00:30:10.951 hermetically sealed in with the rest of your NixOS closure. 00:30:10.951 --> 00:30:11.423 Oh, boy. 00:30:11.586 --> 00:30:15.081 And then, so that's one level of dynamic stuff. And then on top of that, 00:30:15.202 --> 00:30:19.821 you're now also going to be able to attach BPF programs to do that role of figuring 00:30:19.821 --> 00:30:21.274 out what interpreter to run. 00:30:23.167 --> 00:30:24.392 So you could make your own interpreter? 00:30:24.757 --> 00:30:28.141 Yeah, you can have your own interpreter and you can choose very dynamically 00:30:28.141 --> 00:30:30.528 at runtime which interpreter is selected. 00:30:31.752 --> 00:30:33.341 Wow. Dang. 00:30:33.341 --> 00:30:36.811 It's pretty early and like this is going to be this is like the kernel opens 00:30:36.811 --> 00:30:39.771 the features. It doesn't actually add the stuff you do. So this will be downstream 00:30:39.771 --> 00:30:43.791 dependent on distros and users and stuff to actually register the interpreters 00:30:43.791 --> 00:30:44.854 and like write them and all of that. 00:30:45.214 --> 00:30:47.581 But I certainly expect we'll see an XOS at least. 00:30:47.581 --> 00:30:50.791 This is the stuff that makes the Linux kernel the king. 00:30:50.791 --> 00:30:51.171 Oh, yeah. 00:30:51.171 --> 00:30:55.483 It is the king. And the other kernels, they can't even with this stuff. It's really something. 00:30:55.930 --> 00:31:00.561 I suppose in order for the king to maintain its reign, it must shed some older 00:31:00.561 --> 00:31:01.962 projects from time to time. 00:31:02.554 --> 00:31:06.081 And in the upcoming Linux 7.3, not 7.2, those were just released, 00:31:06.081 --> 00:31:10.981 but 7.3, which is coming out in October, 32-bit ARM support has been deprecated. 00:31:10.981 --> 00:31:11.591 Whoa. 00:31:12.526 --> 00:31:16.212 Yeah. So I believe it will survive through 7.3, but it's officially deprecated, 00:31:16.212 --> 00:31:20.421 which doesn't mean it's gone, but it means it's now marked deprecated. So there's that. 00:31:20.688 --> 00:31:25.152 One thing we maybe should mention, too, is, which I don't know if it will be 00:31:25.152 --> 00:31:31.952 fixed in 7.3 or not, but presumably it would be, there's a NTFS vulnerability going around right now. 00:31:33.172 --> 00:31:37.093 Essentially, not the newer NTFS, but the older Paragon NTFS, 00:31:37.314 --> 00:31:38.777 this seems to be live now. 00:31:39.452 --> 00:31:44.532 Essentially, you can set basically a version of Unix security model permissions, 00:31:44.532 --> 00:31:46.770 you know, CHMOD stuff on NTFS files. 00:31:47.345 --> 00:31:49.475 And that includes the set UID bit. 00:31:50.108 --> 00:31:54.452 And the old driver in a lot of configurations for desktops will mount, 00:31:54.452 --> 00:31:58.182 will like auto mount an NTFS thumb drive with that set. 00:31:58.182 --> 00:31:58.712 Oh, no. 00:31:59.032 --> 00:32:03.702 Yeah, so you can have a specially crafted NTFS thumb drive that gets you root just by plugging it in. 00:32:04.428 --> 00:32:06.198 Well, I mean, that could be handy sometimes. 00:32:07.603 --> 00:32:13.152 You know, that's really something. But I think the largest change, 00:32:13.152 --> 00:32:16.276 maybe, it's not a feature that's being added to Linux 7.3. 00:32:17.312 --> 00:32:22.992 A third to a half of the patches look like they are AI-generated low-priority 00:32:22.992 --> 00:32:25.709 fixes that are queued up in Linus' next-next patch set. 00:32:26.365 --> 00:32:30.405 And going through that, they say, quote, we are completely overwhelmed. 00:32:30.515 --> 00:32:32.032 Yeah, these are the networking maintainers. 00:32:32.032 --> 00:32:34.149 Yeah, the networking maintainers say we're completely overwhelmed. 00:32:34.968 --> 00:32:38.732 Their counter move is AI reviewing AI with meta-funding a multi-model review 00:32:38.732 --> 00:32:39.954 for every networking patch. 00:32:40.204 --> 00:32:45.032 But the practical result seems to be really just a lot of trimming from the 00:32:45.032 --> 00:32:48.392 kernel of older network drivers that are not really maintained because they're 00:32:48.392 --> 00:32:49.927 getting poked at by clankers. 00:32:50.984 --> 00:32:55.601 And so they have a whole AI reviewing, AI submissions process now, 00:32:55.601 --> 00:32:57.771 thanks to just a ton of compute contributed by Meta. 00:32:57.771 --> 00:33:01.441 Yeah, they describe it as managing patchwork, automating common process complaints, 00:33:01.441 --> 00:33:05.721 editing commit messages, and maybe applying patches, which already got reviewed 00:33:05.721 --> 00:33:08.421 by tags from people we trust. So they're experimenting actively. 00:33:08.421 --> 00:33:10.761 Yeah, I suppose they have to with the volume they're going to get, 00:33:10.761 --> 00:33:13.811 right? Because the Linux kernel has to be the one everybody thinks of when they 00:33:13.811 --> 00:33:15.571 all of a sudden I can build software. 00:33:16.191 --> 00:33:20.831 I'm going to go to the Linux kernel, right? A lot of people think that, but it's interesting. 00:33:21.289 --> 00:33:27.741 that 7.2 was widely impacted by AI-assisted development and 7.3 is both in contributions 00:33:27.741 --> 00:33:30.188 and again in pull requests and things like that. 00:33:30.699 --> 00:33:33.961 So it's essentially continuing to escalate. And so they're responding by adding 00:33:33.961 --> 00:33:36.781 more compute on their side. I think that's going to be a story worth following. 00:33:36.781 --> 00:33:40.961 We also saw Andrew Morton talking about using Gemini to not write the code, 00:33:40.961 --> 00:33:44.781 but write some of his memory management subsystem patch summaries this cycle. 00:33:44.781 --> 00:33:45.670 Look at that. 00:33:54.773 --> 00:33:58.617 Well, for years, we've had the project of, well, thinking about first, 00:33:58.617 --> 00:34:02.283 and then also just tightening up our production infrastructure behind the shows. 00:34:02.892 --> 00:34:06.417 Big dreams. Some of it's actually happening recently. And Chris, 00:34:07.188 --> 00:34:10.707 Dropbox has been a big part of that, but I know we've had issues with cloud 00:34:10.707 --> 00:34:13.475 syncing in the last several years. 00:34:13.986 --> 00:34:18.548 And nobody's excited about cloud stuff anymore, and you're making some changes. 00:34:19.262 --> 00:34:20.807 What did you replace? 00:34:20.807 --> 00:34:26.177 Well, so, you know, I'm a fan of SyncThing. I want to be clear, I've used it for years. 00:34:26.817 --> 00:34:30.677 When it first launched, I started using it. I was initially with BitTorrent 00:34:30.677 --> 00:34:33.457 Sync, and I was sad when that went away, but I happily Those were the days. 00:34:33.457 --> 00:34:35.929 It was. It was great. But SyncThing has served me well. 00:34:36.451 --> 00:34:39.197 And I've mostly used it for, like, back-end large file transfer media, 00:34:39.197 --> 00:34:45.277 stuff like that. And I've held on to Dropbox because we got this sweet 15-terabyte deal. 00:34:45.277 --> 00:34:46.217 Yeah, yeah, yeah. 00:34:46.217 --> 00:34:50.650 It's crazy cheap. And how do you say no to 15 terabytes these days? 00:34:51.056 --> 00:34:54.621 But I hate Dropbox client. I don't really like Dropbox the company. 00:34:54.911 --> 00:35:00.967 So I kind of compensated and got by for a while by using the Mistral open source 00:35:00.967 --> 00:35:03.861 sync client that was really lightweight. 00:35:04.777 --> 00:35:07.917 You could run it headless. You could declaratively configure it. 00:35:09.161 --> 00:35:12.647 And it gave me Dropbox like back in the OG days of Dropbox again. 00:35:12.647 --> 00:35:15.204 And it just ran headless in the background and synced my damn files. 00:35:15.355 --> 00:35:15.737 That's all you want. 00:35:15.737 --> 00:35:20.667 And used their storage, right? And that worked well for a couple of years. 00:35:20.667 --> 00:35:23.517 And then Mistral announced that they're shutting down. And I thought, 00:35:23.517 --> 00:35:24.747 okay, I'm going to have to migrate soon. 00:35:25.455 --> 00:35:28.117 Right. So either you have to go back to the standard clients or... 00:35:29.517 --> 00:35:33.587 And then, as sort of like to really poke me to get going, I don't know, 00:35:33.587 --> 00:35:36.560 like an old machine joined the Dropbox. I have my suspicions. Yeah. 00:35:37.738 --> 00:35:41.430 All of the files got a sync conflict in my Dropbox. No. 00:35:41.848 --> 00:35:42.838 That doesn't sound good. 00:35:42.838 --> 00:35:46.048 A little bug in how the exclusion folders work with Mistral, 00:35:46.048 --> 00:35:49.313 and it started syncing to all my systems. Like, you know, terabytes of data. 00:35:49.679 --> 00:35:53.086 Terabytes of data. So all my hard drives started filling up, and I realized it. 00:35:53.574 --> 00:35:57.741 So at first I shut down the sync client on my workstation upstairs because I'm sitting there. 00:35:58.501 --> 00:36:03.288 And then it dawns on me, every other computer is also doing this same thing right now. 00:36:03.468 --> 00:36:03.928 Oh, my goodness. 00:36:05.078 --> 00:36:07.578 So then it's like a mad dash before it fills up their hard. 00:36:07.578 --> 00:36:08.488 Drive right. 00:36:09.136 --> 00:36:09.598 Or whatever it. 00:36:09.598 --> 00:36:10.854 Might been there been there. 00:36:11.133 --> 00:36:15.988 Yeah yeah that's a nightmare you've been looking at sync stuff this or the last couple of weeks too. 00:36:16.308 --> 00:36:19.848 Yeah i uh i guess a similar problem actually i've been, 00:36:20.528 --> 00:36:26.198 trusting cloud syncing less and thought geez with these new mesh networks certainly 00:36:26.198 --> 00:36:29.748 there's something better we could do right and not have to rely about it with 00:36:29.748 --> 00:36:32.478 a always on uh system in the in the cloud. 00:36:32.478 --> 00:36:37.662 So you pushed me towards sync thing. I've complained for years that I can never wrap my head around, 00:36:38.712 --> 00:36:43.312 it how it works and how to configure it and stuff like that but somehow on my 00:36:43.312 --> 00:36:47.855 third try around i got over that and it's actually quite simple. 00:36:48.412 --> 00:36:51.959 I think that i don't know what the issue was well the biggest head shift right 00:36:52.121 --> 00:36:59.702 is um like with next cloud sync or dropbox sync it's user login and then you 00:36:59.702 --> 00:37:01.931 have these folders and you sync these folders down, 00:37:02.390 --> 00:37:06.182 and with sync thing there's no user right i mean you can pass protect the dashboard 00:37:06.182 --> 00:37:10.512 but it's these keys that you exchange and then they discover themselves and 00:37:10.512 --> 00:37:13.251 they sync and you share these keys and that's how you set it up, 00:37:13.802 --> 00:37:16.672 which is actually great because it's accountless which means anybody that has 00:37:16.672 --> 00:37:19.172 sync thing can just grab that you can share that key with them and they can 00:37:19.172 --> 00:37:20.873 sync they don't have to go create an account anywhere, 00:37:21.331 --> 00:37:24.240 but it's a total flip of the traditional file sync model 00:37:24.663 --> 00:37:27.982 where you have an account you log in and it just syncs down your files but like 00:37:28.302 --> 00:37:31.072 i think both you and i have found recently that next cloud isn't working for 00:37:31.072 --> 00:37:34.262 us that might have been the direction i would have gone on for these files because 00:37:34.262 --> 00:37:37.522 I do really like having a web interface. I think that's really nice. 00:37:37.522 --> 00:37:40.319 Yeah, I share a bunch of stuff with other people too. 00:37:40.679 --> 00:37:42.142 Yeah, sharing is also a big one. 00:37:42.502 --> 00:37:46.552 Yeah, it's a tricky one. The other thing that got me with SyncThing, 00:37:46.552 --> 00:37:48.638 I think, was the configs for each client. 00:37:49.294 --> 00:37:53.552 You have to kind of put your head into different perspectives when you're configuring 00:37:53.552 --> 00:37:56.717 stuff. And also there's this concept of like, 00:37:57.897 --> 00:38:01.867 I forget what they call it now, maybe Wes, you'll remember, but of like the 00:38:01.867 --> 00:38:05.477 quote unquote cloud, but it's really just the amalgamation of all the files 00:38:05.477 --> 00:38:07.597 of all of the devices into this 00:38:07.597 --> 00:38:12.517 one like virtual file system almost that you can think of it that way. 00:38:12.817 --> 00:38:16.917 And you get to pick and choose which folders you bring down into this particular, 00:38:17.837 --> 00:38:21.434 device. So it's interesting, very powerful when you start learning the config. 00:38:22.200 --> 00:38:25.597 It just sort of slips into the background too. It does have a web UI. 00:38:25.597 --> 00:38:28.157 There are some desktop clients, but it just sort of slips into the background. 00:38:28.157 --> 00:38:32.227 So you're not really thinking about it. It's just like you just have these files 00:38:32.227 --> 00:38:34.896 that persist across all your machines. And it just sort of does it. 00:38:35.657 --> 00:38:37.811 It's best when it's at the layer of infrastructure magic. 00:38:38.286 --> 00:38:41.127 So I had this problem happening, right, where all my systems are getting blasted 00:38:41.127 --> 00:38:44.288 with files to the point where it was going to affect production. 00:38:44.816 --> 00:38:49.737 So I shut it all down and figured I got to redo this. And I wanted to replace 00:38:49.737 --> 00:38:51.055 it at that point with SyncThing. 00:38:51.624 --> 00:38:53.621 I just felt like it's been the most reliable. 00:38:54.277 --> 00:38:59.397 and it was probably my safest bet now that I'm also not using Nextcloud. Sort of the way to go. 00:38:59.519 --> 00:39:04.024 That's a big change though, right? Because you kind of get, it's a very different architecture. 00:39:04.291 --> 00:39:09.667 And in particular, if you don't have a cloud relay, you're going to need a way 00:39:09.667 --> 00:39:10.961 for these things to talk to each other. 00:39:11.530 --> 00:39:16.807 That is a thing. That is a thing. I will say, I think, at first I thought it 00:39:16.807 --> 00:39:18.699 was going to be a downgrade because of that. 00:39:19.744 --> 00:39:22.681 Because I don't have a web UI, I thought this is not going to be as functional. 00:39:23.192 --> 00:39:28.718 But what I realized is that with SyncThing, because it has an API, 00:39:29.903 --> 00:39:33.239 There's actually a lot more visibility into it than I have with Dropbox. 00:39:33.239 --> 00:39:36.899 So with sync thing, I can see which peers are connected, what folders are caught 00:39:36.899 --> 00:39:38.895 up, what's still outstanding in the sync. 00:39:39.261 --> 00:39:43.789 And I can see, does this particular host have that file or not? 00:39:44.346 --> 00:39:48.438 And I can expose that in a way that an agent can read or a script can check 00:39:48.687 --> 00:39:50.609 or Prometheus endpoint can monitor. 00:39:51.289 --> 00:39:55.049 And so then it starts to become a lot more like kind of like this black box 00:39:55.049 --> 00:39:57.923 of sync and like this production monitor thing. 00:39:58.527 --> 00:40:01.499 which is even better than i had with dropbox so then it becomes like oh i'm 00:40:01.499 --> 00:40:06.829 not just replacing dropbox i'm like improving the functionality here so i leaned in on that, 00:40:07.289 --> 00:40:10.543 i i really i wanted to lean in on making sure my hermes agents, 00:40:10.990 --> 00:40:14.019 could understand where files were at and where they're needed to be reviewed 00:40:14.019 --> 00:40:17.159 because it's something my hermes agents can do because i put this on that system 00:40:17.159 --> 00:40:21.479 as well they can write a file to a folder for them and then they know when the 00:40:21.479 --> 00:40:25.589 workstation i'm at has receive that file. That's really powerful. 00:40:25.589 --> 00:40:25.949 Yeah, yeah, yeah. 00:40:26.329 --> 00:40:29.789 Especially when you want, you really need to be, you know, the number of times 00:40:29.789 --> 00:40:32.109 where we've tried to sync something at the studio and we've assumed that it 00:40:32.109 --> 00:40:34.999 synced and then we discover later after we've left the studio that in fact it 00:40:34.999 --> 00:40:38.999 did not fully sync or like one file out of the seven we needed wasn't there or whatever. 00:40:38.999 --> 00:40:39.272 Yeah. 00:40:39.591 --> 00:40:40.648 You just don't want that. Yeah. 00:40:40.851 --> 00:40:43.119 It's come up. I mean, it's come up recently too and, you know, 00:40:43.119 --> 00:40:44.630 you need to be able to change these things. 00:40:44.810 --> 00:40:47.979 You need to be able to make it flexible too and all of that and Dropbox didn't 00:40:47.979 --> 00:40:48.907 really meet that requirement. 00:40:49.116 --> 00:40:52.431 It didn't really do it for me. So that part feels like a huge upgrade now. 00:40:53.272 --> 00:40:59.029 So I think the trickier part, really, was all of the systems are different. 00:40:59.029 --> 00:41:03.369 I mean, I have my NIC systems, but then my Aurora host here is running, 00:41:03.369 --> 00:41:04.772 you know, UBlue, SilverBlue. 00:41:05.231 --> 00:41:09.359 The soundboard machine over here is an old Mac, you know, so that's a totally 00:41:09.359 --> 00:41:11.289 different system. So that part... 00:41:11.309 --> 00:41:13.739 So it's not one platform that you have to build across. 00:41:13.739 --> 00:41:14.079 Yeah. 00:41:14.079 --> 00:41:17.247 You got to make this work across a disparate set of platforms with different hardware. 00:41:17.392 --> 00:41:17.689 Yeah. 00:41:17.689 --> 00:41:19.528 Did that mean you also had, like, 00:41:20.556 --> 00:41:24.260 like disparate folder structures across those systems as well like how did you. 00:41:24.260 --> 00:41:29.590 Implement this in a way that was quote-unquote smooth yes it was you know like 00:41:29.590 --> 00:41:30.529 especially like on the mac oh, 00:41:31.220 --> 00:41:35.280 yeah like slash users slash username like the whole path is totally different 00:41:35.280 --> 00:41:39.380 but then on some of my systems i have those scary raids and so i put the sinks 00:41:39.740 --> 00:41:43.110 on the scary raid several of them and then and then like the laptops they're 00:41:43.110 --> 00:41:46.938 just in the home folder because i don't have more than one hard drive and a laptop usually usually 00:41:47.362 --> 00:41:52.615 so it's just in my home folder so like literally And then you have my Hermes agent's host, 00:41:53.295 --> 00:41:56.621 which stores that in a completely different location too. So literally all my hosts. 00:41:58.229 --> 00:42:00.283 That guy's probably the only one who knows where all the paths are. 00:42:00.405 --> 00:42:04.279 Well, that's the thing. So I focused on the main folders that I use, 00:42:04.279 --> 00:42:06.469 not all of them, just the main ones I use today. 00:42:06.469 --> 00:42:13.009 And then I had the agent figure out where they all need to go and which systems 00:42:13.009 --> 00:42:16.319 we could just hot swap and take over the existing Dropbox location. 00:42:16.319 --> 00:42:19.999 Nice, right. Because you were working from like a unified pre-model mostly. 00:42:19.999 --> 00:42:21.829 Why sync everything again if I don't have to? 00:42:21.829 --> 00:42:22.144 Yeah. 00:42:22.371 --> 00:42:26.079 But that could also go very wrong on some systems. So on some systems, 00:42:26.079 --> 00:42:28.689 I had to make the call not to do that. And I had to create a new folder pattern. 00:42:28.689 --> 00:42:29.339 Start from scratch. 00:42:29.339 --> 00:42:30.219 Yeah, yeah, yeah. 00:42:30.219 --> 00:42:32.564 Okay, well, that sounds like maybe you ran into some issues. 00:42:34.529 --> 00:42:36.929 Some of this sounds like you've handled it. You've obviously been using the 00:42:36.929 --> 00:42:40.689 same thing for years. But I'm wondering, did anything go wrong? 00:42:40.689 --> 00:42:43.949 Did you have any scary moments? What was the process like? 00:42:43.949 --> 00:42:47.221 I think the scariest moment was I forgot to include the Mac. 00:42:47.889 --> 00:42:48.569 I don't think about the Mac. 00:42:50.354 --> 00:42:52.129 But you need a lot of soundboard files. 00:42:52.129 --> 00:42:56.999 That's the soundboard files. Yeah, exactly. said i was like oh right but what 00:42:56.999 --> 00:43:00.827 is really fantastic with this setup is everything's documented all the endpoints 00:43:01.239 --> 00:43:04.263 i mean the clanker can even connect in and read the database yeah 00:43:04.689 --> 00:43:09.488 so it is really straightforward for for just you go in there and say deploy it on a new host, 00:43:10.039 --> 00:43:13.899 and it uh i have some guidance in there like it caught okay well this folder 00:43:13.899 --> 00:43:17.519 is your backups folder that's 20 gigabytes do you want to sync that to the soundboard 00:43:17.519 --> 00:43:22.206 machine i do not i do not want to sync that so that was kind of nice so i kind of i have, 00:43:23.042 --> 00:43:26.829 i'm really only dealing with 10 folders that i actively use and i'm considering 00:43:26.829 --> 00:43:31.479 just keeping everything else on dropbox removing all the dropbox clients and 00:43:31.479 --> 00:43:35.867 then using our clone to mount it when i do need something from the archive, 00:43:36.534 --> 00:43:40.089 because i'm paying such a cheap price it's like 50 a year or something for 15 00:43:40.089 --> 00:43:43.959 i got like this grandfather deal it's a cloud backup i mean so nuts for 15 terabytes 00:43:43.959 --> 00:43:47.509 of cloud backup So our clone could be a good way to interface with that. 00:43:48.109 --> 00:43:48.779 Can I get access? 00:43:48.779 --> 00:43:49.866 So the Mac was tricky. 00:43:50.882 --> 00:43:51.478 Yeah, right? 00:43:51.478 --> 00:43:51.778 Yeah. 00:43:51.778 --> 00:43:55.415 The Mac was the scary part because I almost completely forgot it. 00:43:56.019 --> 00:43:57.138 I just totally blanked it out. 00:43:57.138 --> 00:43:59.874 Did it, was it smooth for you going cross-platform? 00:44:01.908 --> 00:44:04.831 It wasn't too bad. Yeah, I mean, you know, the nice thing is you got Homebrew 00:44:05.023 --> 00:44:06.028 on the Mac. You can use that. 00:44:06.988 --> 00:44:10.228 I think Aurora. So Aurora is a, like I said, it's a silver blue, 00:44:10.228 --> 00:44:12.192 uBluehost, and I went Podman. 00:44:12.888 --> 00:44:15.472 I just did a Podman container on that, a rootless Podman container. 00:44:16.122 --> 00:44:18.318 And I don't think Syncthing cares. 00:44:18.318 --> 00:44:22.208 And then I assume it's just riding back on the back of your existing mesh infrastructure. 00:44:22.208 --> 00:44:22.408 Right? 00:44:22.408 --> 00:44:25.352 So a container plus mesh, and there you go. 00:44:26.205 --> 00:44:31.068 It's really good. I'm really liking it. I find it to be such an upgrade, 00:44:31.068 --> 00:44:33.998 and I thought it was going to be a downgrade. Screw that Dropbox. 00:44:33.998 --> 00:44:38.628 Are you using Dropbox and sync thing looking at the same folders? 00:44:38.628 --> 00:44:42.088 Because I would assume that's going to cause you some issues. 00:44:43.073 --> 00:44:47.208 Yeah, yeah, yeah, yeah. No, I had to be careful about that. But it took me a 00:44:47.208 --> 00:44:49.801 second to realize I had to shut it down on one of my hosts. 00:44:50.271 --> 00:44:50.808 That's all it takes. 00:44:50.808 --> 00:44:54.567 But when I scrambled, I didn't just stop this thing. I'm like, 00:44:54.927 --> 00:44:58.468 all right, disable the service, rebuild. This thing's gone. 00:44:58.808 --> 00:45:01.231 I did not want. So I just removed the client. 00:45:01.632 --> 00:45:04.296 So I don't have to worry about that except for I think it was the Mac. 00:45:04.801 --> 00:45:06.746 In fact, I think Dropbox is actually running right now. 00:45:08.048 --> 00:45:12.188 I even disabled it from starting, and it still starts. I hate macOS so much. 00:45:14.861 --> 00:45:17.938 so yeah it's still this it's it's such a pain in the butt and i don't think 00:45:17.938 --> 00:45:21.778 about the mac so the mac is an issue for sure i'm just gonna uninstall it after 00:45:21.778 --> 00:45:25.588 the show off of here and i don't know i don't really need to access the archive 00:45:25.588 --> 00:45:28.048 from from this thing one you know what i realized. 00:45:28.048 --> 00:45:29.252 I think i got our clone on the mac. 00:45:29.623 --> 00:45:33.678 We could put linux on this i've been waiting to replace this with a linux box 00:45:33.678 --> 00:45:37.616 right an actual but we could just put linux on this because it's an intel mac. 00:45:38.521 --> 00:45:38.849 Yeah. 00:45:39.389 --> 00:45:42.409 There's the whole soundboard thing. Isn't that the reason? 00:45:42.409 --> 00:45:44.109 Yeah, we have to, no, we have to build a soundboard. 00:45:44.109 --> 00:45:45.119 But we need to do that anyway. 00:45:45.119 --> 00:45:45.849 We have to do that anyway. 00:45:45.849 --> 00:45:47.299 We've been saying it for like, what, 10 years? 00:45:47.299 --> 00:45:50.219 Well, like, what's the point of starting the project if you don't have any hardware to run it on? 00:45:50.219 --> 00:45:53.115 Fair enough, fair enough. Well, let's do that after the show. 00:45:53.642 --> 00:45:54.555 Right, right. 00:45:55.349 --> 00:45:55.849 Hmm. 00:45:57.846 --> 00:46:00.989 So that's been an issue. You know, but being able to run it in Podman worked 00:46:00.989 --> 00:46:03.825 out great. And then it's just installed natively on the Nix systems. 00:46:05.049 --> 00:46:07.535 So are you managing your configs through Nix as well? 00:46:08.876 --> 00:46:10.106 Yeah, for those hosts. Yeah. 00:46:10.350 --> 00:46:11.101 Yeah, for all those hosts. 00:46:11.101 --> 00:46:11.341 Everything. 00:46:11.341 --> 00:46:14.261 That's how I started doing it. It's so nice. I even... 00:46:14.381 --> 00:46:15.174 It is very nice. 00:46:15.487 --> 00:46:15.831 Yeah. 00:46:15.831 --> 00:46:18.558 And it's positive pressure to use NixOS for more and more things. 00:46:19.421 --> 00:46:22.431 I would... One recommendation, if you're going to throw, like, 00:46:22.431 --> 00:46:24.317 a mobile device in the mix, 00:46:24.874 --> 00:46:29.521 which I don't know if you want to at this point, but I found it really nice 00:46:29.521 --> 00:46:35.181 to have the clanker just, like, make a config file for the Android system and just suck it in. 00:46:35.981 --> 00:46:36.482 You can import it? 00:46:36.825 --> 00:46:40.381 Yeah, you can import them pretty nicely. And that keeps you from like having 00:46:40.381 --> 00:46:43.256 to thumb all the config options and stuff on the phone. 00:46:43.623 --> 00:46:45.968 So that would be one of my recommendations if you get there. 00:46:46.484 --> 00:46:47.561 I wonder, presumably you could 00:46:47.561 --> 00:46:51.453 also maybe get some port forwards going on your Android and have it API. 00:46:52.941 --> 00:46:54.601 You're right. Or just it'd be on the mesh too. 00:46:54.601 --> 00:46:55.111 Yeah, right. 00:46:55.111 --> 00:46:56.648 Yeah. So I think, 00:46:58.941 --> 00:47:04.051 if you do dabble in the clankers, I really want to make a strong recommendation 00:47:04.051 --> 00:47:06.751 that if you need to solve a sync problem, you look at sync thing. 00:47:06.751 --> 00:47:12.439 because my favorite part of this is I built a wrapper called Hermes sync thing. 00:47:13.101 --> 00:47:17.861 And so the agents are not doing like raw credential passes over the network 00:47:17.861 --> 00:47:19.921 and like looking things up in the LLM every single time. 00:47:20.490 --> 00:47:24.401 So the agents aren't using like the raw credentials and connecting to the API 00:47:24.401 --> 00:47:25.661 without instruction every single time. 00:47:25.661 --> 00:47:28.031 And the LLM is just kind of figuring out how to do it every time when I want 00:47:28.031 --> 00:47:31.334 to check on the status of stuff. I've created this little wrapper and this wrapper, 00:47:31.983 --> 00:47:35.031 automatically has all the presets to go see if all the hosts is connected, 00:47:35.031 --> 00:47:38.751 if the folders are caught up, what might be outstanding from any individual hosts. 00:47:39.621 --> 00:47:41.901 You know, like, for example, what does Aurora still need before we can go on 00:47:41.901 --> 00:47:43.701 air? Because we've got to get the clip synced up. 00:47:44.541 --> 00:47:45.121 I like that. 00:47:45.121 --> 00:47:49.401 And I can ask it, did this file reach this host? And all the clinker is doing 00:47:49.401 --> 00:47:52.149 in the background is calling this Hermes sync thing. 00:47:52.990 --> 00:47:55.240 And that's a wrapper that just has the endpoints you want. 00:47:55.240 --> 00:47:58.720 Yes, predefined setup. And so it's not, and the advantage there, 00:47:58.720 --> 00:48:02.005 it isn't creating the workflow every single time. So you're saving tokens, 00:48:02.238 --> 00:48:06.170 you're reducing the impact of hallucination because it's fairly straightforward. 00:48:06.170 --> 00:48:09.150 You can do like a multi-stage, right? So it's like you have a skill describing 00:48:09.150 --> 00:48:12.600 this, how to work with it, and telling it how to use your wrapper. 00:48:12.809 --> 00:48:15.250 And then the wrapper can embed all of the implicit knowledge of like how to 00:48:15.250 --> 00:48:16.440 get those core tasks done. 00:48:16.440 --> 00:48:21.540 And the wrapper can query the destination machine's own sync thing database. 00:48:21.540 --> 00:48:22.220 Nice. 00:48:22.220 --> 00:48:25.700 And see, okay, is it present? Is it valid? Is it synced? Was it ignored? 00:48:25.700 --> 00:48:29.420 Was it not deleted? Whatever. So like you get beyond just proof of delivery, 00:48:29.420 --> 00:48:31.697 you can get proof of removal and all of that. 00:48:31.935 --> 00:48:35.010 And it's just using the wrapper every single time, which I could go down on 00:48:35.010 --> 00:48:37.386 the command line and run myself if I wanted to. 00:48:37.513 --> 00:48:40.520 But the nice thing is, is when you do something like that in that wrapper is 00:48:40.520 --> 00:48:42.808 you can use a free model, a cheaper model. 00:48:43.173 --> 00:48:46.060 It doesn't have to be like a big, you know, frontier class model. 00:48:46.060 --> 00:48:48.642 It can be an open source, even eventually self-hosted model, 00:48:48.967 --> 00:48:51.610 but it can execute these little queries for me and tell me if everything synced. 00:48:51.610 --> 00:48:54.510 There's no debugging, there's no code writing, it's just column tools. 00:48:54.510 --> 00:48:56.211 Yeah, it's pretty great. 00:48:56.720 --> 00:48:59.760 So I was going to ask if you were going to try and come up with a web interface 00:48:59.760 --> 00:49:03.514 replacement, but it sounds like for the most part, this is good enough. 00:49:03.932 --> 00:49:08.400 Yeah, I still toyed with the idea of putting some sort of web file manager in 00:49:08.400 --> 00:49:09.300 front of the sync thing directly. 00:49:09.300 --> 00:49:10.800 Just for like upload-download or something like that. 00:49:10.800 --> 00:49:12.550 Yeah, or this one-off machine. 00:49:12.550 --> 00:49:12.840 Right. 00:49:12.840 --> 00:49:16.450 Or mobile, that could be another way to do it on mobile. I don't know what sort 00:49:16.450 --> 00:49:18.000 of state of the art there is. 00:49:18.607 --> 00:49:23.107 Like, I don't think it's PHP file manager anymore. So if anybody has recommendations 00:49:23.107 --> 00:49:26.077 for that, I would definitely like them. Because I do think that could be useful, 00:49:26.077 --> 00:49:30.488 but I don't think it's as necessary as this. It's been pretty nice. 00:49:31.133 --> 00:49:34.737 So are you done? I mean, is there anything left that you need to build in this 00:49:34.737 --> 00:49:36.891 thing besides the web interface? 00:49:37.135 --> 00:49:39.367 Are you happy with the deployment? Is it on all your machines? 00:49:39.367 --> 00:49:42.907 Yeah, everything's deployed. It's gotten really nice. You know, 00:49:42.907 --> 00:49:46.077 I think the trick was is deploying from one central machine and just having 00:49:46.077 --> 00:49:51.578 it reach out over SSH and just blast it onto everything with their various esoteric setups. 00:49:52.757 --> 00:49:55.723 I think the missing piece still, like I mentioned might be the R clone mount. 00:49:56.048 --> 00:49:56.535 Right. 00:49:57.110 --> 00:50:00.349 And just coming up with a better way to take advantage of that 15 terabytes while I have it. 00:50:01.499 --> 00:50:04.747 Not sure. But I think there's, that's the work still to be done. 00:50:04.747 --> 00:50:06.572 There's no errors. Everything seems to be sinking. 00:50:07.530 --> 00:50:09.817 And then I do have a couple like this machine. I got to go clean up Dropbox 00:50:09.817 --> 00:50:12.226 on. Same thing is just, it feels like this, 00:50:13.771 --> 00:50:16.813 a sleeper of a project that we don't talk about, that we take for granted. 00:50:17.469 --> 00:50:21.388 I think it's just assumed that we've all implemented it. 00:50:21.544 --> 00:50:24.765 Like that's the feeling I get is that it's been around forever. 00:50:24.765 --> 00:50:26.984 It's been really stable for a long time. 00:50:27.458 --> 00:50:31.895 It's just kind of, it's like SSH. It's just there when you need it. 00:50:32.655 --> 00:50:40.205 But for anybody who hasn't tried it, you have any suggestions or tips or pitfalls 00:50:40.205 --> 00:50:41.563 to avoid or anything like that? 00:50:41.953 --> 00:50:47.039 And if you have SyncThing, do yourself a favor. Seriously, listen to Chris's hot tip here. 00:50:47.695 --> 00:50:51.905 Download OpenCode. Use the free aux model that's in there. Use any of the free 00:50:51.905 --> 00:50:56.483 models. Point it at your URL for your SyncThing and have it tell you what this API can do. 00:50:57.238 --> 00:51:02.908 Because this is where it really exceeds what Dropbox can do for you. 00:51:03.374 --> 00:51:07.975 And it becomes a complete system that you have observability over that doesn't 00:51:07.975 --> 00:51:10.160 have a cloud, doesn't have any big tech login. 00:51:10.653 --> 00:51:14.636 And I think so many times we deploy things like Home Assistant, 00:51:15.036 --> 00:51:17.125 Plex, Jellyfin, and SyncThing. 00:51:17.567 --> 00:51:21.205 And then a lot of us, and I'm definitely guilty of this for a long time, 00:51:21.205 --> 00:51:26.025 just ignore that there is this beautiful API that allows us to do everything 00:51:26.025 --> 00:51:31.445 we could do through the UI with our Clanker or OpenCode or whatever through this API. 00:51:32.182 --> 00:51:37.545 And so you can start writing little scripts or automations or check on the sync status of things. 00:51:37.545 --> 00:51:41.371 Think of the poor developers who implemented these APIs. They want you to call them. 00:51:41.371 --> 00:51:41.679 They do. 00:51:41.888 --> 00:51:43.645 Don't let them atrophy just there on the side. 00:51:43.645 --> 00:51:48.425 Take advantage of those APIs. Because I stopped thinking of sync thing as just 00:51:48.425 --> 00:51:50.467 this thing I self-host and it moves my files. 00:51:50.787 --> 00:51:52.825 And now a full system I can manage. 00:51:52.825 --> 00:51:53.965 It's a sync architecture. 00:51:53.965 --> 00:51:54.975 It's a framework. 00:51:54.975 --> 00:51:55.704 Nice. 00:51:56.495 --> 00:52:00.195 That's what it is. That's what it is. And it's way better than any stuff you 00:52:00.195 --> 00:52:03.665 could get with the commercial stuff. And really it's taking advantage of that API to its fullest. 00:52:03.665 --> 00:52:06.245 And you can craft it to fit in with whatever you need it to be. 00:52:14.360 --> 00:52:18.470 Well, as usual, we got a little feedback this week. Alex sends in a little note. 00:52:18.470 --> 00:52:20.322 Hey, Chris, Wes, and Brent. 00:52:20.931 --> 00:52:21.450 I'm Alex. 00:52:21.870 --> 00:52:27.950 I'm Alex. That's A-L-E-K-S for, you know, that's a much cooler way to do it. I'm Alex. 00:52:27.950 --> 00:52:28.819 Alex with a K? 00:52:29.946 --> 00:52:31.512 I don't know, an X is pretty cool, dude. 00:52:31.512 --> 00:52:32.772 Well, that's true. 00:52:33.692 --> 00:52:37.832 X is pretty cool. K is different. I like it because it's unique. 00:52:37.832 --> 00:52:38.712 Yeah, yeah, yeah. 00:52:39.147 --> 00:52:41.399 I mean, it's pretty rare to get an X in your name. 00:52:42.485 --> 00:52:48.464 So Alex with a K says, I'm a solar developer working on Local Comet, 00:52:48.789 --> 00:52:56.121 an open source Windows-first local AI app built with Llama CPP, Tori, and Rust. 00:52:56.945 --> 00:52:58.932 All right. Oh, you don't know. 00:53:00.972 --> 00:53:05.222 It gives people a simple desktop way to run local models with more privacy and 00:53:05.222 --> 00:53:07.932 control without depending on a cloud service. 00:53:07.932 --> 00:53:12.552 I'm actively developing it and would be grateful if you'd consider mentioning 00:53:12.552 --> 00:53:18.232 this on this here podcast and sends a little repository. 00:53:19.236 --> 00:53:20.850 Well, this is right up our alley. 00:53:20.960 --> 00:53:26.002 Yeah, we got Python, TypeScripts, Felt, and yes, Rust. Neat, 00:53:26.002 --> 00:53:28.492 a private local first Windows AI assistant. 00:53:28.492 --> 00:53:30.392 Yeah, and you don't see Windows getting a lot of love in this camera. 00:53:30.392 --> 00:53:33.702 That's what I was going to say. I'm not about to go make myself an EXE, 00:53:33.702 --> 00:53:35.576 but it's clearly out there. 00:53:35.901 --> 00:53:40.852 And the people, I mean, if we're going to be using these models to help us with 00:53:40.852 --> 00:53:43.552 our sync thing, I mean, as we already talked about on the show, 00:53:43.952 --> 00:53:45.782 do you want to be touching Windows or do you want your model to do it? 00:53:45.782 --> 00:53:49.712 I mean, honestly, why not, right? And then there's probably somebody out there 00:53:49.712 --> 00:53:53.212 listening to this show that has WSL and has all this running and then they use 00:53:53.212 --> 00:53:55.441 Windows MCP and all of that. 00:53:56.212 --> 00:53:59.292 But if you're just getting started, something like Local Comet could actually 00:53:59.292 --> 00:54:02.962 probably help you get WSL up and going. I mean, I don't know what these things 00:54:02.962 --> 00:54:05.054 can do on a Windows platform, but it is really nice to see it. 00:54:05.773 --> 00:54:08.182 And maybe it's an easy way, like if you have a Windows gaming machine that does 00:54:08.182 --> 00:54:11.340 have a GPU, then maybe it's an easy way to go experiment with some local models too. 00:54:11.515 --> 00:54:13.407 Yeah, I bet a lot of Windows, you know, that's a great point. 00:54:13.952 --> 00:54:16.592 Don't lock them away in the Windows world. You can still get to play with the fun tech. 00:54:16.712 --> 00:54:18.442 They've invested in those GPUs after all, right? 00:54:18.452 --> 00:54:18.952 That's right. 00:54:18.952 --> 00:54:22.532 So they might as well use them. Yeah. Local Comet. So there you go. 00:54:22.532 --> 00:54:25.622 We'll put a link to that in the show notes. If you got a Windows box with a 00:54:25.622 --> 00:54:31.030 Goopoo and you want to play around with it, you'll find it at linuxunplugged.com slash 681. 00:54:36.120 --> 00:54:40.624 Oh, our first boost comes from the one, the only Greg, the lawyer. 00:54:40.624 --> 00:54:43.484 He was in our mumble room for a little bit earlier today. He's often on the 00:54:43.484 --> 00:54:46.277 live streams and frequently a big supporter of the show. 00:54:46.737 --> 00:54:51.839 And he comes in with a baller boost of 33,333 Satoshis. 00:54:55.770 --> 00:54:59.924 And Greg says, inspired by the release of Oma Archie Atra or Quattro, 00:55:00.264 --> 00:55:02.804 I have been clanking my own Nix OS Senate. 00:55:02.804 --> 00:55:06.254 Wait, wait, wait. I got to stop you there. clanking as a. 00:55:06.254 --> 00:55:07.324 Verb i've. 00:55:07.884 --> 00:55:09.364 Not seen till now i like it. 00:55:10.207 --> 00:55:14.824 I do like it uh i've converted from kde to hyperland and quickshell nice making 00:55:14.824 --> 00:55:17.144 my own desktop so much more fun oh, 00:55:17.904 --> 00:55:24.334 this is i love seeing reports like this greg um we are entering a new era of customization on linux, 00:55:25.019 --> 00:55:29.844 and uh hyperland is just so far it's so right i'll speaking of oma archie and 00:55:29.844 --> 00:55:32.948 hyperland um an oma archie foundation of types, 00:55:33.430 --> 00:55:37.824 was launched with eight million dollars of backing by jack dorsey the ceo of 00:55:37.824 --> 00:55:42.634 dell and many others whoa and um one of their first expenses as a foundation 00:55:42.634 --> 00:55:43.542 if you can call that i'm not sure, 00:55:44.053 --> 00:55:47.734 is they're backing vaxery the hyperland developer for the next three to five years. 00:55:47.734 --> 00:55:48.574 Oh that's that's. 00:55:48.584 --> 00:55:51.434 Incredible that's huge yeah at least the next couple years i mean i don't know 00:55:51.434 --> 00:55:55.304 how much vaxery is going to take as a paycheck but um that's massive so he can 00:55:55.304 --> 00:55:57.962 focus on the hyperland desktop which i think is going to be really good. 00:55:58.206 --> 00:56:03.104 A quick shell i was looking like quite the yes toolkit kit as well so i put 00:56:03.104 --> 00:56:06.966 a show note a link in the show notes to a great lwn write-up of that recently. 00:56:07.894 --> 00:56:10.936 Very good thank you greg appreciate hearing from you appreciate the support too. 00:56:11.105 --> 00:56:15.011 Tomato boots in with 22 222 sats, 00:56:17.268 --> 00:56:21.015 Catching up with the missed episodes from this summer. Loved the backwards episode. 00:56:21.015 --> 00:56:22.735 And how it broke your brains. 00:56:23.615 --> 00:56:27.075 Regarding virtual machines in the kernel, I'll just lean into my recent NetBSD 00:56:27.075 --> 00:56:30.469 trolling here and point out that kernel already has a Lua VM in it. 00:56:30.995 --> 00:56:32.239 It's honestly pretty fun. 00:56:32.477 --> 00:56:33.743 That is fun. That does sound fun. 00:56:34.455 --> 00:56:37.365 Great show on entropy and the cold card mess. I tried looking into their firmware 00:56:37.365 --> 00:56:41.265 before moving my BTC to it and found the whole thing overly complicated and 00:56:41.265 --> 00:56:43.275 confusing, so I decided I didn't trust it. 00:56:43.275 --> 00:56:43.872 Nice dodge. 00:56:44.012 --> 00:56:48.085 Despite having bought one, my coins are still on a Sparrow-created wallet on an offline machine. 00:56:48.085 --> 00:56:49.370 Yep, turned out to be a better call. 00:56:49.631 --> 00:56:52.845 Wish I'd shared my discomfort publicly, and I hope all the JB listeners got 00:56:52.845 --> 00:56:55.958 their coins out in time. Especially in this new era of clanker-assisted hacking, 00:56:56.284 --> 00:56:59.065 those of us in free software who have the expertise to autocode should really 00:56:59.065 --> 00:57:00.495 speak up if something smells funny. 00:57:00.855 --> 00:57:02.635 Yeah, there's something to think about there. 00:57:02.635 --> 00:57:03.345 That's a great point. 00:57:03.345 --> 00:57:06.075 Definitely worth checking your own stuff, too. And at least checking the stuff 00:57:06.075 --> 00:57:08.425 you're going to use, even if you don't go file a clanker report. 00:57:08.425 --> 00:57:10.256 Although, if you find something serious, you probably should consider it. 00:57:11.051 --> 00:57:13.519 Yeah. Thank you, sir. Appreciate that boost. 00:57:14.482 --> 00:57:20.054 Turt Ferguson also sent in a row of McDuck's 22,222 sets. 00:57:24.060 --> 00:57:30.167 I enjoy the peek into the future of IT defense, but you did not answer a key question that I have. 00:57:30.742 --> 00:57:35.096 Do you think AI security tools are actually useful for regular Linux admins 00:57:35.096 --> 00:57:38.586 yet, or are they still too unreliable to trust? 00:57:38.886 --> 00:57:42.856 I think that depends, maybe, because you experimented with a meta model that 00:57:42.856 --> 00:57:46.166 did hallucinate some of the reporting, which maybe would be like, 00:57:46.166 --> 00:57:48.226 okay, that's an example of when you can't fully trust. 00:57:48.691 --> 00:57:53.406 Yeah, I think it's kind of like a lot of things, you know, defense in depth. 00:57:53.406 --> 00:57:58.326 So you probably want to have, that's where you like a diversity of models or 00:57:58.326 --> 00:58:01.486 harnesses or processes so that you do have some way to have a, 00:58:01.486 --> 00:58:03.366 like, you know, grounding is very important here. 00:58:03.598 --> 00:58:06.816 And so if you have a way that you can actually do some fact checks or you have 00:58:06.816 --> 00:58:10.239 a council of several models, each doing it with different unrelated contexts, 00:58:10.500 --> 00:58:13.297 things like that to try to be able to figure out when that's happening. 00:58:13.552 --> 00:58:18.046 I think the biggest barrier right now to getting actually good results out of 00:58:18.046 --> 00:58:22.946 these tools would be paying the cost for the higher-end uncensored models that 00:58:23.246 --> 00:58:24.849 you've got to pay kind of per token for. 00:58:25.261 --> 00:58:30.146 So that's the gating factor right now, right? For them to do anything more advanced, really, I think. 00:58:30.915 --> 00:58:34.146 And that, yeah, that kind of pattern, too. You can have a grunt model and then 00:58:34.146 --> 00:58:37.886 another model that sort of checks on things, does its due diligence, 00:58:37.886 --> 00:58:42.078 and has a little more intelligence but doesn't have to do all of the original work. 00:58:43.425 --> 00:58:47.390 Nathan from Brissy Boosin, 4,102 cents. 00:58:50.221 --> 00:58:52.172 Come on, Lions. Hat trick time. 00:58:53.670 --> 00:58:56.710 Yeah, that probably works. You're getting a few people thinking about it. 00:58:56.710 --> 00:58:58.470 That puts it in a good direction, I think. 00:58:59.270 --> 00:59:03.103 All right. Thank you, Nathan. Appreciate you. Right, you want to take Monty there? 00:59:03.207 --> 00:59:05.802 Yeah, Monty sent in a row of ducks. 00:59:06.214 --> 00:59:06.980 Ah. 00:59:08.414 --> 00:59:14.230 As for the Frigate system overview, I have been running Frigate for several 00:59:14.230 --> 00:59:17.335 years, and I haven't touched the config in a very long while. 00:59:17.874 --> 00:59:23.825 So very interested in some of the newer features, including how to drain your own detections. 00:59:24.271 --> 00:59:28.540 Yeah. And Friggin has recently added more detection capabilities in the freer 00:59:28.540 --> 00:59:30.730 stuff. That's something I'll play around with, you know, and maybe we'll do 00:59:30.730 --> 00:59:32.033 an updated segment. I'll dig into that. 00:59:32.411 --> 00:59:36.140 One of the things I have recently done is I've expanded beyond person identification 00:59:36.140 --> 00:59:39.840 to animal identification. Because again, keep an eye out for coyotes and yeah, 00:59:39.840 --> 00:59:42.065 the birds and stuff like that. And that's been interesting. 00:59:43.010 --> 00:59:45.310 So, all right. Thank you, Monty. That's good feedback. 00:59:45.850 --> 00:59:46.630 Quick interruption. 00:59:46.630 --> 00:59:46.890 Yes, sir. 00:59:46.890 --> 00:59:50.400 I think you'll love this. You should go pull up our live chat. Our Matrix. 00:59:50.400 --> 00:59:51.370 I'm going over there right now. 00:59:51.410 --> 00:59:53.302 Because check out what Greg the lawyer just did. 00:59:54.695 --> 00:59:58.730 While I was listening about SingThing, I clanked a QuickShell SingThing widget for my system. 00:59:59.810 --> 01:00:00.750 That's so great. 01:00:01.290 --> 01:00:02.190 I clanked it, he says. 01:00:03.370 --> 01:00:06.040 And look at this lovely little widget, right? It shows devices connected. 01:00:06.040 --> 01:00:09.750 It's got some folders there. Web UI rescan, refresh stop buttons. 01:00:09.750 --> 01:00:10.907 QuickShell is so cool. 01:00:11.279 --> 01:00:12.393 And clearly it is quick. 01:00:12.579 --> 01:00:15.910 Yeah. And it's not going to break with Hyperland desktop update, 01:00:15.910 --> 01:00:18.402 which is nice, too. So you can build that and it's going to work for a while. 01:00:19.040 --> 01:00:21.500 Well done, Greg. That's why you dropped out of the mumble room. 01:00:21.500 --> 01:00:23.430 You had to go work on that. Plank that up. 01:00:24.967 --> 01:00:31.323 I'll take Thomas. Thomas English comes in with 16,345 Satoshis. 01:00:36.172 --> 01:00:42.157 What are the best methods to get my agent to do my entire NixOS setup for me? Because I am clueless. 01:00:42.859 --> 01:00:46.796 I've only ever Vive slopped Linux servers together, but going to take the plunge 01:00:46.796 --> 01:00:48.652 and smash out this Windows box. 01:00:49.117 --> 01:00:54.144 NixOS for me. Any tips for a sailor that can't swim? I've never seen a line of code in my life. 01:00:55.549 --> 01:01:00.076 Agent first NixOS setup. What say you? I'm going to paste in your YoloPie and close my eyes. 01:01:00.076 --> 01:01:00.556 Oh, my goodness. 01:01:00.656 --> 01:01:01.916 I'm going to paste it into YoloPie. 01:01:01.916 --> 01:01:04.166 It's like a prompt request via Boost. 01:01:04.166 --> 01:01:04.976 It sure is. 01:01:06.496 --> 01:01:09.206 I mean, you could, I mean, I don't know if you'd want to have it, 01:01:09.206 --> 01:01:11.006 you could point a clinker at Hypervibe. 01:01:11.163 --> 01:01:14.316 I was going to say, and we've had some, the whole collection of listeners who 01:01:14.316 --> 01:01:16.590 have shared in past show notes. 01:01:17.067 --> 01:01:20.096 So you could perhaps the first prompt is go get it to harvest some links from that. 01:01:20.096 --> 01:01:20.426 Yeah. 01:01:20.426 --> 01:01:22.696 And then from there, it could go build out a config for you. 01:01:22.696 --> 01:01:28.366 I think that's a good starting spot. I would, I'd start probably in a web chat 01:01:28.366 --> 01:01:32.086 with one of the free models to just sort of build out a plan. 01:01:32.086 --> 01:01:36.745 talk about, you know, kind of ramble about your intentions, build out a plan, link it to Hypervibe. 01:01:37.134 --> 01:01:39.946 I think there's a re-spin of Fedora that looks really good. You can even point 01:01:39.946 --> 01:01:42.596 it at Oma Archie and see if there's some, you know, design tips you want to 01:01:42.596 --> 01:01:45.736 pull from there and application ideas because it can look at all that stuff 01:01:45.736 --> 01:01:47.176 and sort of build you a master list. 01:01:47.176 --> 01:01:51.036 And then you could have it from there, produce a master prompt that you then 01:01:51.036 --> 01:01:53.330 put into the final clanker on your machine that does the work. 01:01:53.812 --> 01:01:55.449 And you probably get pretty good results with that. 01:01:56.192 --> 01:02:00.246 Yeah. And, you know, the main kind of things too is once you get like the installing 01:02:00.246 --> 01:02:03.596 Nix OS part It can be just pretty much like installing a normal, 01:02:03.596 --> 01:02:04.996 any old Linux distribution. 01:02:05.596 --> 01:02:10.726 It's just getting the, once you have it running and you have a way for a clinker 01:02:10.726 --> 01:02:12.916 to control it, then you're off to the races. 01:02:12.916 --> 01:02:14.586 You are. You are indeed. 01:02:14.586 --> 01:02:17.617 I think I would recommend a very different approach, actually. 01:02:18.656 --> 01:02:21.512 I think you should definitely use a clinker to help you. 01:02:21.959 --> 01:02:27.340 But point it at some of the documentation and then just have it almost like 01:02:27.602 --> 01:02:29.956 interview you about what you want the system to do. 01:02:30.516 --> 01:02:35.317 what kind of things you want in it, and have it explain what it's doing. 01:02:36.854 --> 01:02:39.702 I think the long-term benefit of understanding what's happening under the hood 01:02:39.702 --> 01:02:43.732 for NixOS is really good, even though you don't have to worry about the implementation 01:02:43.732 --> 01:02:48.324 details, like all the syntax and stuff like that. It's really good to understand what's possible. 01:02:48.859 --> 01:02:52.362 And I think starting with some of the documentation, or at least projects that 01:02:52.362 --> 01:02:54.188 are well-regarded, is a good way to go. 01:02:55.308 --> 01:02:58.535 Two more tips. When it's done or when you're at a stopping point, 01:02:58.866 --> 01:03:03.336 have it create like an agents MD that explains your system so that the next 01:03:03.539 --> 01:03:07.312 fresh LM context, when it starts up, it kind of knows your system. 01:03:07.312 --> 01:03:10.692 So a basic layout of where your config's at, where your preferences are, 01:03:10.692 --> 01:03:14.412 the desktop choices, the things you prefer the clanker do, you put all that 01:03:14.412 --> 01:03:17.422 into like an agents MD or some sort of description file about your system. 01:03:17.422 --> 01:03:19.728 And then when you start a new session, you have a point at that. 01:03:20.030 --> 01:03:23.642 So that way it's not rediscovering your next config where it's at every single time and all of that. 01:03:23.642 --> 01:03:25.731 It knows the rebuild command. It knows all the, yeah. 01:03:25.835 --> 01:03:31.572 Yeah, so I think that's probably maybe the most important tip is just have it build one of those. 01:03:31.572 --> 01:03:36.168 And like, yeah, the other thing that would maybe be good for you to consider is, 01:03:36.632 --> 01:03:41.922 second tip, you could ask it to document in a human reviewable format decisions 01:03:41.922 --> 01:03:45.812 that are being made and just kind of it could append and update a markdown file 01:03:45.812 --> 01:03:48.702 or a text file for you that you could review at the end. 01:03:48.702 --> 01:03:52.622 You could have another LLM review, something you could have it also produce 01:03:52.622 --> 01:03:53.702 two documents if you want. 01:03:53.702 --> 01:03:57.002 You could have it produce a technical document for an LLM with no context, 01:03:57.275 --> 01:04:00.833 a fresh LLM session with no context, a technical document that explains everything, 01:04:01.129 --> 01:04:04.132 a fresh LLM with no context needs to get started and work from the beginning 01:04:04.132 --> 01:04:07.368 to modify, change, upgrade, anything to my system. 01:04:07.583 --> 01:04:10.672 And then a second document that's for a human audience that needs to understand 01:04:10.672 --> 01:04:12.767 the decisions that are made and the design choices that were made. 01:04:13.063 --> 01:04:17.498 And it will just continue to update both of those documents for you as it works in that session. 01:04:17.696 --> 01:04:20.652 As long as you maintain the consistency and it sees that every session. 01:04:20.652 --> 01:04:21.231 So remember that. 01:04:21.462 --> 01:04:24.312 Yeah, a devlog is one word you can use for some of those concepts. 01:04:24.312 --> 01:04:24.737 Ooh, okay. 01:04:24.923 --> 01:04:31.001 I have two more quick tips that I think are good for someone starting out. If you are... 01:04:32.115 --> 01:04:35.616 comfortable enough to install things, uh, like in open code, 01:04:36.092 --> 01:04:38.742 having an MCP, uh, is really helpful here. 01:04:38.742 --> 01:04:43.562 So there's MCP dash Nix OS, which is a really great program just to help like 01:04:43.562 --> 01:04:48.562 query what's even available in Nix OS. And it saves a lot of time and mistakes 01:04:48.562 --> 01:04:50.691 and just like querying the internet. 01:04:50.964 --> 01:04:55.283 So, uh, I would recommend that. And then if you are up for it, 01:04:55.608 --> 01:04:59.259 writing yourself just a little skill for NixOS with like, 01:04:59.997 --> 01:05:04.002 common patterns that you particularly like once you get down the road is really 01:05:04.002 --> 01:05:06.386 really helpful or at least has been for me. 01:05:06.722 --> 01:05:10.062 Yeah and if it's not obvious what they might be they will become obvious to 01:05:10.062 --> 01:05:11.602 you as you use it they will become. 01:05:11.602 --> 01:05:13.662 Obvious if you keep having to tell it the same thing yeah. 01:05:13.662 --> 01:05:15.032 There you go if you tell it three times. 01:05:15.032 --> 01:05:16.138 Then you should put it in a skill. 01:05:17.242 --> 01:05:19.712 Moon Knight comes in with two thank you very much for the boost good luck let 01:05:19.712 --> 01:05:25.262 us know how it goes Moon Knight comes in with a boost to 2500 sats and he says 01:05:25.262 --> 01:05:29.048 to all the new Fiat boosters you now have boost power, 01:05:31.522 --> 01:05:35.683 nice i like that that's true thank you moon and i thanks for the sound clip too. 01:05:36.990 --> 01:05:40.368 Uh pabby boosts in with 7108 sats, 01:05:43.961 --> 01:05:44.722 time travel. 01:05:44.722 --> 01:05:48.240 Boost episode 671 love hearing from the time travelers. 01:05:48.512 --> 01:05:52.872 Thanks for the ulanzi smart pixel watch tip i've been looking for one to integrate 01:05:52.872 --> 01:05:57.602 into Home Assistant since I'm not getting around to building the Autrix kit I've had for years. 01:05:58.682 --> 01:06:02.772 Also, on 1.5 speed, Forescan sure sounded a lot like Forescan, 01:06:02.772 --> 01:06:05.652 which had me quite confused for the segment. Had to look at the show notes. 01:06:05.652 --> 01:06:06.629 Yeah, not a great name. 01:06:06.971 --> 01:06:07.302 It's the Ford scanner. 01:06:07.642 --> 01:06:11.462 But, you know, that is the risk you take with the 1.6X. 01:06:11.462 --> 01:06:14.062 I think it's, what could else, yeah, that's right. What else could, 01:06:14.382 --> 01:06:16.613 what did they do? Ford would come after them if they put the D in there. 01:06:16.857 --> 01:06:18.952 They can't put the D in there, so they just had to go with Forescan. 01:06:18.952 --> 01:06:20.332 Forescan? No, that doesn't help. 01:06:20.332 --> 01:06:24.601 Yeah, it doesn't work either. yeah it doesn't work thank you though appreciate it pappy. 01:06:24.961 --> 01:06:27.817 Well zach attack sent in 6 000 sets, 01:06:30.737 --> 01:06:32.774 Loved the episode on hacking yourself. 01:06:33.250 --> 01:06:36.527 I did dislike the episode where everything was in reverse, though. 01:06:37.487 --> 01:06:42.427 Need to still watch your Clanker Therapy video? Yes, that's linked at every 01:06:42.807 --> 01:06:45.667 episode recently, right at the top of the links. For now. 01:06:45.667 --> 01:06:47.258 For now. Almost done linking that, too. 01:06:47.374 --> 01:06:52.817 Something you could take a look at is Rockbox. I'm running it on a SureFans 01:06:52.817 --> 01:06:59.617 F20 for podcast and music. It has better audio than the phone and increased privacy, too. 01:06:59.617 --> 01:07:02.847 That's true. I haven't heard of Rockbox in a hundred years. That's really cool. 01:07:02.847 --> 01:07:07.667 This sounds vaguely familiar. Is that a piece of software you can throw on some 01:07:07.667 --> 01:07:09.787 MP3 devices? I believe so. 01:07:10.027 --> 01:07:10.127 Yeah. 01:07:10.726 --> 01:07:11.007 Yeah. 01:07:11.007 --> 01:07:14.797 I think I might've tried to play with this a long time ago. Long, 01:07:14.797 --> 01:07:16.757 long time. It's still around. There you go. 01:07:16.995 --> 01:07:20.577 Zach, thanks for the memories. Appreciate that. Appreciate that. 01:07:20.577 --> 01:07:23.300 Thank you very much. Sir, I'm here with a row of ducks. 01:07:23.921 --> 01:07:26.887 It's pretty fitting. Thanks for breaking our brains with the reverse episode. 01:07:27.288 --> 01:07:30.187 I was about to blame my podcast app, But how dare you skip the hello, 01:07:30.187 --> 01:07:33.367 friends, and welcome back intro. Anyways, here's some value for the show. 01:07:33.967 --> 01:07:34.457 Thanks. 01:07:34.683 --> 01:07:37.927 Hey, if it got a boost, maybe it was worth it after all. But I think we're done 01:07:37.927 --> 01:07:39.815 doing that. I think we're done doing that. 01:07:41.057 --> 01:07:44.647 Our pal Magnolia, ma'am, comes in with a double row at ducks. 01:07:44.647 --> 01:07:45.028 Hey-oh. 01:07:45.881 --> 01:07:51.027 Two of the same message, but it says, just going off of the title here on 680 01:07:51.027 --> 01:07:55.882 Go Hack Yourself, but I just got started with syncing myself after a Git workflow of mine failed. 01:07:56.323 --> 01:07:58.807 Can't wait to see what the show's about. Ah, that must have been to the pending. 01:07:58.807 --> 01:07:59.287 Yeah, the live one. 01:07:59.287 --> 01:08:00.265 Live boost. 01:08:00.474 --> 01:08:04.157 Well done. Well done. Boosting one before we even started the show, 01:08:04.157 --> 01:08:07.823 man. Well done. Yeah, let us know how it goes with your sync thing set up. 01:08:07.991 --> 01:08:10.127 That should be very interesting. 01:08:11.764 --> 01:08:16.420 Blackhost comes in with a fiat boost of $1,000? 01:08:18.365 --> 01:08:21.067 What? What? 01:08:21.067 --> 01:08:21.627 Wow. What? 01:08:24.987 --> 01:08:25.319 What? 01:08:27.478 --> 01:08:28.726 When did that get in there? 01:08:28.726 --> 01:08:30.826 I don't know. I'm just seeing it. 01:08:31.326 --> 01:08:31.966 When did that come in? 01:08:32.586 --> 01:08:33.936 10, 21 a.m. this morning. 01:08:33.936 --> 01:08:35.386 Wow, we were already on the air. 01:08:35.386 --> 01:08:36.576 That is crazy. 01:08:36.576 --> 01:08:39.766 Wow, Blackhost. Thank you. That's really, that's... 01:08:43.581 --> 01:08:46.206 I got to pick up my jaw here this week. Yeah, thank you. 01:08:46.646 --> 01:08:47.826 No ads placeholder. 01:08:47.826 --> 01:08:48.956 Thank you, Blackhost. That's so sweet. 01:08:48.956 --> 01:08:53.306 That is really amazing. Wow. Wow, wow, wow. Really do appreciate that, 01:08:53.666 --> 01:08:55.730 and it definitely will make a difference. Thank you. 01:08:56.746 --> 01:08:59.726 David from Burlington comes in with $10 fiat Hey. 01:08:59.726 --> 01:09:03.676 You know what? We have a boost sound for that now, right? Should we roll it 01:09:03.676 --> 01:09:06.208 up? Let's try it What do you think? You ready? See if it feels good, 01:09:08.606 --> 01:09:11.616 That's not bad It's not bad Hey folks. 01:09:11.616 --> 01:09:15.456 It occurred to me that while y'all have extensive tech backgrounds I don't know 01:09:15.456 --> 01:09:20.026 if I've ever heard where it all came from Oh boy Can you guys give a quick summary 01:09:20.026 --> 01:09:21.326 of your past work history? 01:09:21.326 --> 01:09:22.009 Oh boy, 01:09:23.415 --> 01:09:24.538 Oh, boy. 01:09:24.538 --> 01:09:26.798 It's basically a virus that infected all of us. 01:09:28.663 --> 01:09:31.518 Could we do like some special on that? Or is that worth it? 01:09:32.078 --> 01:09:34.520 Like in a members we put in the show? I don't know. 01:09:34.648 --> 01:09:35.048 I don't know. 01:09:35.048 --> 01:09:37.448 We'll have to think about that. Thank you for asking, though. 01:09:37.448 --> 01:09:37.758 Yeah. 01:09:38.304 --> 01:09:39.158 And for the support. 01:09:39.158 --> 01:09:41.544 The reason why we don't have time to do it now is because we're old. 01:09:42.020 --> 01:09:43.538 And so there's a lot to it. 01:09:44.098 --> 01:09:45.091 We only remember so much. 01:09:45.898 --> 01:09:49.113 I remember before Linux. I was alive. 01:09:49.276 --> 01:09:51.308 Yeah. So, yeah, that's a great, that could be a great topic. 01:09:51.308 --> 01:09:53.378 We haven't really recapped that in a long time. I suppose some people have some 01:09:53.378 --> 01:09:54.406 questions on that. Thank you, David. 01:09:54.951 --> 01:09:57.028 From Burlington? My Burlington? 01:09:57.028 --> 01:09:57.588 Yeah, which Burlington? 01:09:57.598 --> 01:09:58.258 There's one in Ontario. 01:09:58.278 --> 01:09:59.555 Maybe a zip code boost in there. 01:09:59.967 --> 01:10:04.678 Yeah. I'm not getting beers with them then. But all right, we have some member 01:10:04.678 --> 01:10:07.735 boosts that we'd like to get to. Our buddy Anonymous is back. 01:10:07.989 --> 01:10:12.378 Do you guys really use Firefox? You talk about a record number of tabs. 01:10:12.378 --> 01:10:13.556 Have we mentioned that? I don't know that. 01:10:13.928 --> 01:10:16.638 But I don't believe it. I've used it on and off for years, and I keep having 01:10:16.638 --> 01:10:20.987 to bail back to Chrome. I'm trying again, but after a day, too, it just hangs on me. 01:10:21.405 --> 01:10:25.364 I have to kill all Firefox. I got 90 gigs of RAM, a 22-core Xeon, 01:10:25.544 --> 01:10:26.977 50 tabs really shouldn't hang. 01:10:27.680 --> 01:10:28.278 Wow. 01:10:29.258 --> 01:10:33.554 I never have it hang. And it's probably open the entire time my laptop is online. 01:10:33.723 --> 01:10:36.859 I have in the past had issues, but it's been years. 01:10:37.032 --> 01:10:38.130 Yeah, interesting. 01:10:38.582 --> 01:10:42.358 For the question of if we, you know, do we actually use Firefox, 01:10:42.358 --> 01:10:45.558 I'd say it's my permanent first choice. 01:10:46.038 --> 01:10:50.808 I mostly only use Chromium when something goes wrong and the website's not built 01:10:50.808 --> 01:10:54.927 for Firefox or for the show because I need multiple browsers. 01:10:55.734 --> 01:10:59.918 And I would say I probably have the most tabs and windows open out of all of 01:10:59.918 --> 01:11:01.558 us, probably out of you guys combined. 01:11:01.898 --> 01:11:08.268 So there's that. But I only really used to have issues on low-end hardware where 01:11:08.268 --> 01:11:09.213 I didn't have enough RAM. 01:11:09.422 --> 01:11:15.218 But never experienced the slowness or anything like that unless I was filling 01:11:15.218 --> 01:11:20.394 up the RAM. So not sure what your issue is. Hopefully you can figure it out. 01:11:21.497 --> 01:11:25.452 I love this. On my home desktop, I do a boot, Garuda, and Arch for daily driver 01:11:25.452 --> 01:11:28.631 and Windows 11 for occasional work applications that, well, I couldn't port to wine. 01:11:29.008 --> 01:11:31.512 My 15-year-old plays video games, Minecraft, Steam on Garuda. 01:11:31.512 --> 01:11:33.989 Yesterday, my six-year-old came up to me while I was working and said, 01:11:34.198 --> 01:11:36.439 Dad, can you boot into Linux? Windows is boring. 01:11:37.622 --> 01:11:39.572 That's amazing. Nice. 01:11:39.572 --> 01:11:40.062 Aw. 01:11:40.496 --> 01:11:44.612 He says, with all of the success with the reverse episode, how about a randomness 01:11:44.612 --> 01:11:46.602 episode? You could do the segments in random order. 01:11:46.602 --> 01:11:47.022 Whoa. 01:11:47.022 --> 01:11:47.486 Good luck. 01:11:48.062 --> 01:11:49.102 Yeah, that won't be harder at all. 01:11:49.302 --> 01:11:52.082 Curse on the show. I don't know. You really want to melt our brains. 01:11:52.082 --> 01:11:54.772 It'd be fun, too, if we just literally rolled a dice or something and we didn't 01:11:54.772 --> 01:11:57.904 even know it just totally ran out. There's something to that. 01:11:58.502 --> 01:12:01.122 Or each MP3 that gets generated is a unique order. 01:12:01.362 --> 01:12:03.102 Different order. Boy, if we could pull that off. 01:12:03.742 --> 01:12:05.637 Why do we do this to ourselves? Dynamic segments. 01:12:08.782 --> 01:12:12.912 Anton member boosts in. For the Web Boost anonymous username issue, 01:12:12.912 --> 01:12:16.812 the quickest fix might be a small confirmation pop-up when the username field 01:12:16.812 --> 01:12:20.242 is left blank or default and the user clicks the member boost or checkout button. 01:12:20.242 --> 01:12:23.597 the pop-up could confirm that they meant to send it anonymously. 01:12:23.777 --> 01:12:26.362 I don't love a pop-up, but that is not a bad idea. 01:12:27.282 --> 01:12:29.234 Thank you for thinking there, Anton. 01:12:29.994 --> 01:12:35.962 Well, Jimmy's here with a free members boost, which you could do at boost.jupiterbroadcasting.com 01:12:35.962 --> 01:12:36.896 if you're so interested. 01:12:37.367 --> 01:12:42.372 How do you reconcile the privacy considerations of AI? It can be amazingly helpful, 01:12:42.372 --> 01:12:45.848 of course, but they get a lot of sensitive personal info. 01:12:46.758 --> 01:12:49.352 That's what's holding me back. A local model would fix that, 01:12:49.352 --> 01:12:53.097 but of course the VRAM required for good models is quite prohibitive. 01:12:54.088 --> 01:12:59.388 How can you be running graphene, but pumping configs into the AI machine? 01:12:59.806 --> 01:13:03.550 I guess it, I mean, do you consider Nix configs to be confidential information? 01:13:03.915 --> 01:13:04.816 I generally don't. 01:13:04.816 --> 01:13:05.196 No, I don't either. 01:13:05.196 --> 01:13:06.936 Especially for something I might consider putting on. 01:13:07.436 --> 01:13:08.606 I post them on GitHub, yeah. 01:13:08.606 --> 01:13:11.996 You guys don't keep your bank account information in your Nix OS config? 01:13:11.996 --> 01:13:12.636 Not anymore. 01:13:13.596 --> 01:13:16.276 No, now I don't know. Declarative banking. 01:13:18.574 --> 01:13:23.656 So, like, if I were putting in personal information, I suppose that would be one thing. 01:13:24.616 --> 01:13:28.151 Yeah, you know, health stuff, addresses, account number, you know. 01:13:28.303 --> 01:13:32.406 So there can be ways you want to have concerns about that. And some information 01:13:32.406 --> 01:13:36.106 will leak, right, if you ask it questions about how do I set up X thing or I'm 01:13:36.106 --> 01:13:37.156 trying to solve this problem. 01:13:37.576 --> 01:13:41.906 But the question is what risks – how does that fit into your model of – Right. 01:13:41.906 --> 01:13:46.343 In my model, while that is a risk, it's actually less of a risk than like, 01:13:46.465 --> 01:13:50.076 you know, data brokers that are monitoring all of my transactions or, 01:13:50.076 --> 01:13:53.206 you know, government surveillance or Google surveillance. 01:13:53.206 --> 01:13:58.082 or, you know, like those are actually more prevalent in my day-to-day life that I already live with. 01:13:58.684 --> 01:14:05.058 I actually wouldn't mind if they did train on my Nix configs and they got better at it. 01:14:05.743 --> 01:14:10.006 So in a way, I kind of want them to, I don't think the models I'm using are 01:14:10.006 --> 01:14:13.646 using my data for training, but if they wanted to become better at Nix, 01:14:13.646 --> 01:14:15.136 I really would not have any problem with that. 01:14:15.495 --> 01:14:18.056 I think of it too as sort of like a controller agency thing. 01:14:18.056 --> 01:14:21.676 Like for me, I mean, yes, the Graphene OS sort of privacy stuff, 01:14:22.376 --> 01:14:26.026 is great, but i really like that it makes me actually feel like i control the 01:14:26.026 --> 01:14:28.316 android phone that's it and that's where it's like with the ai 01:14:28.716 --> 01:14:33.046 i might make mistakes or whatever um or be lazy but it's ultimately my choice 01:14:33.046 --> 01:14:36.216 of like do i engage with it what do i put into that prompt before i hit enter 01:14:36.596 --> 01:14:38.716 and stock android didn't really feel like that. 01:14:39.592 --> 01:14:45.152 I have uh hmm yeah various categories like you say chris um nix west configs, 01:14:45.616 --> 01:14:47.851 totally good to send those out i think they're, 01:14:48.763 --> 01:14:53.183 not private really at all. Um, but I do hesitate on some topics. 01:14:53.183 --> 01:14:58.113 Like there's a couple topics that I have that like accounting automations and 01:14:58.113 --> 01:15:02.538 stuff that I have sitting in the queue for when I feel like, 01:15:03.293 --> 01:15:08.749 there's some great option for like privacy, local AI stuff. So I can't wait. 01:15:09.202 --> 01:15:13.383 But the other thing I would say is, um, there are a couple of things you could do about that. 01:15:13.383 --> 01:15:18.533 Like I, I don't trust myself completely to to keep myself safe so i put a couple 01:15:18.533 --> 01:15:23.173 layers like there's uh an mcp that you can add i'm trying to see it here that 01:15:23.173 --> 01:15:26.187 i've added to my open code just to like 01:15:26.623 --> 01:15:33.467 pull out um secrets and bank out and credit card numbers so it just goes through and, 01:15:34.239 --> 01:15:38.223 sees if you've accidentally you know put anything like that in there and so 01:15:38.223 --> 01:15:41.663 there's a couple things you can do just for peace of mind maybe before. 01:15:41.663 --> 01:15:42.493 It hits the model. 01:15:42.493 --> 01:15:42.973 Yeah uh. 01:15:43.873 --> 01:15:47.223 Wes has talked about also like at the injector level if you have an injector 01:15:47.223 --> 01:15:51.780 sitting between your client and the cloud service you can have something that's, 01:15:52.500 --> 01:15:56.413 looks like it's the it looks like the api to the client but it's actually sitting 01:15:56.413 --> 01:16:00.063 between the cloud model and the client and it's looking for sensitive information yeah. 01:16:00.063 --> 01:16:02.983 We've talked about a few of these uh key g proxies another one there's a few 01:16:02.983 --> 01:16:06.213 options out there now and you could also consider this is a lot more technical 01:16:06.213 --> 01:16:10.443 setup but if you do uh entertain something like renting a cloud GPU and then 01:16:10.443 --> 01:16:14.323 running your own self-hosted models, that has a lot of nice privacy implications too. 01:16:14.323 --> 01:16:17.183 And maybe if you only need it for certain batch work or you want to build a 01:16:17.183 --> 01:16:19.963 certain thing and you're not going on day-to-day stuff, you're not running an 01:16:19.963 --> 01:16:22.223 agent or something like that, you're just using it for that type of work, 01:16:22.223 --> 01:16:26.506 a rented GPU with a self-hosted model, well, then you've got full control. 01:16:27.833 --> 01:16:30.198 You've got out of it. So that's a good question. Thank you, Jimmy. 01:16:30.842 --> 01:16:35.013 Pabby comes in with a free time travel member boost. 01:16:35.713 --> 01:16:37.953 So great. This is now possible, Pabby writes. Thank you, Pabby, 01:16:37.953 --> 01:16:40.393 and I love the time travel boost. But you've got to say what episode you're 01:16:40.393 --> 01:16:43.037 at when you time travel. Otherwise, we don't know from what time you're at. 01:16:45.053 --> 01:16:48.361 Digital Jolt comes in with a member boost. 01:16:48.361 --> 01:16:48.821 All right. 01:16:49.181 --> 01:16:53.221 I'm a bit late, but just wanted to say I really enjoyed the Reverso episode. 01:16:53.661 --> 01:16:56.341 It took me a minute to remember that you'd mentioned it a few weeks ago, 01:16:56.661 --> 01:17:00.021 but I smiled every time I heard one of the jingles appear in the wrong place. 01:17:00.021 --> 01:17:00.231 Good. 01:17:00.231 --> 01:17:00.741 All right. 01:17:00.741 --> 01:17:01.081 I'm glad. 01:17:01.381 --> 01:17:01.741 Love that. 01:17:02.729 --> 01:17:08.721 Well, Paul member boosted in. Where are you on the ad block is privacy topic. 01:17:09.101 --> 01:17:12.881 I don't like giving money to corporate overlords. This is the main thing holding 01:17:12.881 --> 01:17:15.784 me back from YouTube Premium right now, for instance. 01:17:16.318 --> 01:17:21.971 I prefer to support a few selected creators. My willingness to spend money is 01:17:21.971 --> 01:17:23.952 limited, leaving many creators out. 01:17:25.618 --> 01:17:32.481 Yeah, that's a tricky one because the ad blocking, at least in the browser, it's not just about ads. 01:17:32.481 --> 01:17:37.761 It's also about performance and readability, obnoxiousness with the pop-ups. 01:17:37.761 --> 01:17:40.811 So there's other things now that are also a factor there. So it's not just like 01:17:40.811 --> 01:17:42.798 a statement on trying to get the content for free. 01:17:43.693 --> 01:17:46.868 And I do tend to agree. I think I kind of align with you, Paul. 01:17:47.461 --> 01:17:51.379 Like I think Pharaonics is a great membership. LWN in our space is a great membership. 01:17:52.801 --> 01:17:57.254 We try to make the Jupiter signal and unplugged core contribution a, 01:17:57.608 --> 01:18:00.870 you know, really as much high signal as we can worth it membership. 01:18:01.520 --> 01:18:05.851 But I have over the years decided because I have a family that watches YouTube, 01:18:05.851 --> 01:18:10.361 I did subscribe for YouTube premium to avoid that. and the YouTube experience is notably better. 01:18:11.001 --> 01:18:11.361 For sure. 01:18:11.361 --> 01:18:14.411 When you sit down on an account that doesn't have it, I cannot believe how much 01:18:14.411 --> 01:18:17.101 ads they're slamming in there now. And I just don't want my kids seeing all that. 01:18:18.281 --> 01:18:21.161 As for the sort of, is it piracy or not? I think it's one of those, 01:18:21.161 --> 01:18:23.181 like, you know, it's sort of a spectrum. 01:18:24.067 --> 01:18:27.700 The world exists where both people are dependent on ads and, 01:18:28.581 --> 01:18:31.411 one, local control of what your computer does is important. 01:18:31.411 --> 01:18:34.421 And, you know, ad blockers are a known technology that exists. 01:18:34.421 --> 01:18:39.187 So I think it is true that choosing that perhaps has a moral dimension, but... 01:18:39.981 --> 01:18:43.490 The idea that, you know, you only need to make one choice for every single creator 01:18:43.810 --> 01:18:44.740 doesn't seem right to me either. 01:18:44.740 --> 01:18:46.470 Yeah, it is a nuanced problem. 01:18:47.173 --> 01:18:51.817 I think for me, this is a really good one. I think about this all the time. 01:18:52.108 --> 01:18:57.720 I've kind of landed on ad blocking everything because I feel like very naively, 01:18:57.720 --> 01:19:04.010 actually, I feel like we could find a better way to fund things instead of using ads. 01:19:04.010 --> 01:19:09.650 So if they're not making money on ads, maybe they can find a better way. 01:19:09.650 --> 01:19:11.636 but I'm sure it would just be more draconian. 01:19:12.239 --> 01:19:15.270 The problem is it's always kind of eventually a race to the bottom. 01:19:15.270 --> 01:19:18.080 You always have to add more ads. It's just the way it goes. 01:19:18.080 --> 01:19:23.570 Yeah. I would say one option is, let's say you're using Adblock Plus or something 01:19:23.570 --> 01:19:28.801 like that, you could disable it on websites that you want to get the revenue. 01:19:29.990 --> 01:19:30.937 I do that on Photopea. 01:19:32.030 --> 01:19:32.190 Yeah. 01:19:34.090 --> 01:19:37.350 They ask, and I think it's worth it. Photopea, they deserve my ad revenue. 01:19:37.350 --> 01:19:38.330 That's another one I should probably... 01:19:38.330 --> 01:19:40.730 It does give you a chance to find out, too, like how far did they go? 01:19:41.490 --> 01:19:44.370 Are these like horrible ads or is it a reasonable sort of ad placement that 01:19:44.370 --> 01:19:45.640 isn't actually that bad in practice? 01:19:45.640 --> 01:19:46.280 Good question. 01:19:46.280 --> 01:19:49.040 The last thing I would add is like the internet is completely unusable without 01:19:49.040 --> 01:19:51.190 it. So there's that, right? 01:19:51.710 --> 01:19:55.680 Yeah, it's bad. All right. I'll wrap us up. Senior Smile comes in with a free 01:19:55.680 --> 01:19:58.748 member boost. Papa John was correct. I forgot to put my username last time. 01:19:59.073 --> 01:20:02.620 I said a USD boost referencing my previous anonymous boost about K8s. 01:20:02.620 --> 01:20:05.030 But you read them in reverse order. Oh, sorry. 01:20:05.030 --> 01:20:05.720 We didn't catch that. 01:20:05.720 --> 01:20:06.770 We don't always get that right. 01:20:07.130 --> 01:20:10.334 Yeah. Sorry, Senior Smile. And then Bearded Zero wraps us up. 01:20:10.938 --> 01:20:14.690 I enjoyed the backwards episode, and I got to say, I think doing the boost and 01:20:14.690 --> 01:20:16.215 feedback may have been a good option. 01:20:16.586 --> 01:20:19.540 It lets it bring the feedback into the show and kind of feels like a bookend 01:20:19.540 --> 01:20:21.004 for the last episode. Just a thought. 01:20:21.439 --> 01:20:25.213 I also felt it was kind of nice to warm up with the feedback and the boost. 01:20:25.830 --> 01:20:30.010 And I think if I were going to always do the show in the order that felt the 01:20:30.010 --> 01:20:31.771 best, that might be the order I'd do it in. 01:20:32.391 --> 01:20:35.321 But if you think about it, you see that, you know, it's like if you're new to 01:20:35.321 --> 01:20:39.071 the show and it takes us 30 minutes or whatever to get to the actual content. 01:20:39.071 --> 01:20:39.911 Not everyone loves that. 01:20:39.911 --> 01:20:43.791 Not a good impression. But I liked it too, Bearded Zero. I liked it too. 01:20:43.791 --> 01:20:47.310 Thank you, everybody, and everybody who boosted below the 2,000 sat cutoff and all of that. 01:20:47.718 --> 01:20:52.991 We really appreciate you. We really had a banger. So we had 124,000 sat sent in. 01:20:52.991 --> 01:20:57.461 But then when you bring in that fiat boost, I mean, we should take a moment 01:20:57.461 --> 01:21:00.981 again. I mean, that was really something. That was really something. 01:21:00.981 --> 01:21:04.921 And $1,000 came into the show today from one single contributor, and we needed it. 01:21:05.541 --> 01:21:07.741 It has been tight behind the scenes. It was. 01:21:09.231 --> 01:21:13.221 That works out to be 1.29 million Satoshis. 01:21:14.454 --> 01:21:16.819 Wow. Look at you doing the math. 01:21:16.991 --> 01:21:20.241 No, it wasn't me. It was hybrid sarcasm. Thank you, hybrid sarcasm in the chat. 01:21:20.241 --> 01:21:23.451 I mean, that's got to put Blackhost on the leaderboard for the year. 01:21:23.451 --> 01:21:24.471 I mean, there you go. 01:21:24.471 --> 01:21:24.841 Right there. 01:21:24.841 --> 01:21:27.051 Yeah, maybe we'll have to do like a Boosties preview sometime soon. 01:21:27.051 --> 01:21:30.321 Yeah. So when you bring it all together and you take the fiat boost and you 01:21:30.321 --> 01:21:33.941 convert them to the current Satoshi value and you just smash it all to one line 01:21:33.941 --> 01:21:36.281 item, we had 16 of your stream stats. 01:21:36.281 --> 01:21:41.151 That's 27,485 Satoshis. When you combine it all together, it gives us, 01:21:41.151 --> 01:21:49.121 I mean, a very impressive grand total of 1,458,305 Satoshis. 01:21:49.501 --> 01:21:50.081 That's crazy. 01:21:50.401 --> 01:21:50.881 What? 01:22:03.570 --> 01:22:06.268 If you'd like to support the show, you can become a member, put it on autopilot. 01:22:06.268 --> 01:22:09.978 You get the bootleg or the ad-free version of the show, and you get a free boost 01:22:09.978 --> 01:22:12.290 every single week. Or support us one-off. 01:22:12.406 --> 01:22:14.758 Like a particular show, want to get your thought on here, just support us. 01:22:14.758 --> 01:22:18.901 Go to boost.jupiterbroadcasting.com. You can send a boost with Fiat or Sats. 01:22:22.100 --> 01:22:23.494 It's crazy. Wow! 01:22:26.802 --> 01:22:30.128 All right. And thank you, everyone, who supported the show. We really appreciate 01:22:30.128 --> 01:22:33.075 it all. With that, it's time for some screaming picks. 01:22:34.812 --> 01:22:39.584 All right. Buckle up, boys, because this week we're talking about SSH Clipboard, 01:22:39.803 --> 01:22:41.864 and this could change your life. 01:22:42.212 --> 01:22:47.443 A peer-to-peer Clipboard sync server over SSH. Copy anything on one machine, paste it on another. 01:22:48.151 --> 01:22:51.309 A picture, a block of code, it doesn't matter. 01:22:51.558 --> 01:22:55.668 You copy here, you paste there. Rich text content doesn't matter. 01:22:55.668 --> 01:22:59.538 It keeps the native formats intact. Native formats, not just text, 01:22:59.538 --> 01:23:04.158 preserving images, rich content, things like screenshots, something you might 01:23:04.158 --> 01:23:07.116 want to put into a ticket maybe. You want to take it from one machine and put it on the other. 01:23:07.389 --> 01:23:09.998 Maybe you've got a clanker that's working on something and doesn't have to save 01:23:09.998 --> 01:23:13.432 it to the file system anymore. It can actually just put it right in your freaking clipboard. Or... 01:23:14.587 --> 01:23:17.923 Or if you put it on multiple machines, the Clanker could put it on the machine's 01:23:17.923 --> 01:23:19.016 clipboard you tell it to. 01:23:19.167 --> 01:23:21.923 You're saying I should be using this instead of netcatting things around. 01:23:21.923 --> 01:23:25.727 This is actually an option to reduce your netcat dependency. 01:23:25.970 --> 01:23:27.163 It looks pretty new, huh? 01:23:27.163 --> 01:23:28.853 It is very new, I should mention. 01:23:28.853 --> 01:23:32.363 From Justin Schroeder and Andrew Boyer, the contributors so far. 01:23:32.363 --> 01:23:35.008 Yeah, yeah, yeah. It is very new. It is very new. 01:23:35.449 --> 01:23:38.563 But I think there's been several attempts at various tools like these over the 01:23:38.563 --> 01:23:41.776 years. I think when I was first getting into Linux, I remember trying to roll something myself. 01:23:42.553 --> 01:23:45.933 It's nice to see something that looks like a structured, well-thought-out approach. 01:23:45.933 --> 01:23:48.373 Yeah, I went through it. I dug through it, and I was very impressed. 01:23:48.373 --> 01:23:52.743 It's early days, but very, very, very impressed with what they've done so far. 01:23:52.743 --> 01:23:56.393 And, of course, if you've got SSH keys set up, just you reuse those, 01:23:56.393 --> 01:23:57.810 so it handles all of that for you. 01:23:58.448 --> 01:24:01.233 It's cross-platforms. You can move stuff between Mac and Linux. 01:24:01.233 --> 01:24:04.713 You can reuse the copy-paste mechanism there. 01:24:04.713 --> 01:24:07.763 It won't break your other clipboard managers. That's good. Right, 01:24:07.763 --> 01:24:12.133 it's using Rust and Ratatouille, which is just turning out to be such a great library. 01:24:12.133 --> 01:24:12.533 Yes. 01:24:12.953 --> 01:24:15.713 I think it'll be packaged in multiple ways right now. Like, they're kind of 01:24:15.713 --> 01:24:17.634 the recommended path is NPM. 01:24:18.513 --> 01:24:22.707 I did look into it. You could probably package this in NIC. It's pretty straightforward. 01:24:23.253 --> 01:24:27.832 So it is very doable. And it is MIT licensed. MIT licensed. 01:24:28.459 --> 01:24:30.183 Well, that has been Blackhost Pick of the Week. 01:24:30.409 --> 01:24:33.653 Yeah. Brought to you by Blackhost. 01:24:36.293 --> 01:24:36.940 I had another one. 01:24:37.276 --> 01:24:40.113 Oh, a sneaky pic, you didn't even put it in the dog. 01:24:40.113 --> 01:24:41.963 I didn't put it in the dog, so I don't remember what it was. 01:24:41.963 --> 01:24:42.513 Though. 01:24:45.493 --> 01:24:46.973 So you're saying you have a pic for next week? 01:24:47.073 --> 01:24:47.873 It's in his clipboard. 01:24:47.873 --> 01:24:49.473 It's been a long time since we've done it. 01:24:49.473 --> 01:24:51.773 Yeah, you should have set it up. It probably is in my clipboard. 01:24:51.773 --> 01:24:53.893 Oh my god, that's hilarious. 01:24:54.353 --> 01:24:56.053 It's been a while since we've had a single pic pic. 01:24:56.573 --> 01:25:01.293 I know. Well, that is a banger of a pic. I just can't remember what my other 01:25:01.293 --> 01:25:02.873 pic was. I can't remember what it was. 01:25:02.873 --> 01:25:03.813 It's probably a single pic. 01:25:07.752 --> 01:25:10.141 No, I was pretty proud of it. It's funny that I can't remember it now. 01:25:10.141 --> 01:25:13.918 That's what I get for editing a text editor first. So we'll sneak it in the next episode. 01:25:14.311 --> 01:25:16.581 And if you've got a banger pick, you can always let us know too. 01:25:16.581 --> 01:25:18.791 We love it when we feature an audience pick. We'll put it in the feedback or 01:25:18.791 --> 01:25:20.163 we might even put it in the main pick segment. 01:25:20.698 --> 01:25:26.449 So you can boost.jupiterbroadcasting.com that or go to linuxunplugged.com slash contact for that. 01:25:26.785 --> 01:25:29.611 All right. Well, that wraps us up. Links to everything we talked about today 01:25:29.611 --> 01:25:34.131 over at the website, linuxunplugged.com slash 681 or jupiterbroadcasting.com 01:25:34.131 --> 01:25:36.340 for the links. and all the shows. 01:25:36.833 --> 01:25:39.301 And if you want to just, you know, chew away at some extra stuff, 01:25:39.301 --> 01:25:41.291 you want to know what we talked about, anything like that, tell them, 01:25:41.291 --> 01:25:43.581 Wes. Tell them about these fancy metadatas that we got. 01:25:43.961 --> 01:25:47.341 Yeah, well, the first level is Cloud Chapters. Yeah, that's right, 01:25:47.341 --> 01:25:50.783 a JSON file with all the information about where our segments are and what they are. 01:25:51.143 --> 01:25:55.771 And you know what? If you look in that, you get custom Drew-made chapters. 01:25:55.771 --> 01:25:56.041 They're great. 01:25:56.041 --> 01:25:57.621 Using all that editor cleverness. 01:25:57.621 --> 01:25:58.141 They are great. 01:25:58.841 --> 01:26:04.023 But if you need more details, check out the VTT file, or if you must, the SRT file. 01:26:04.343 --> 01:26:07.798 We have subtitles, transcripts, right? Available in the RSS feed itself. 01:26:08.366 --> 01:26:12.251 And if you want to see our ugly mugs, there might even be an MP4. 01:26:12.251 --> 01:26:15.809 There could be with screencast stuff and all that of what we're talking about. 01:26:16.203 --> 01:26:19.181 I'll just say again, thank you to Blackhost. You know, it strikes me. 01:26:19.181 --> 01:26:22.731 I was thinking about it this week, like all four of us, the three of us here 01:26:22.731 --> 01:26:25.931 on the show and also Drew, you know, we really have shaped our lives around 01:26:25.931 --> 01:26:28.857 the show. It's that level of dedication. It really drives our life. 01:26:29.661 --> 01:26:32.481 It's the defining clock and heartbeat of our day-to-day life. 01:26:32.481 --> 01:26:35.799 And so when you see support by folks, like we saw this week, 01:26:36.188 --> 01:26:38.731 it really does validate putting that kind of effort and energy and building 01:26:38.731 --> 01:26:41.491 a life around the show to make it possible. So that way we're here every single 01:26:41.491 --> 01:26:43.949 week going on like, what, 13 or something years? 01:26:44.192 --> 01:26:44.591 Crazy. 01:26:44.591 --> 01:26:48.332 Not missing a beat. So we really do appreciate the support. Thank you, everybody. 01:26:48.772 --> 01:26:51.691 Like I said, links at the website. And of course, we'd love it if you joined 01:26:51.691 --> 01:26:53.540 us live. Make it a live vibe on a Sunday. 01:26:57.308 --> 01:27:01.733 Get past the Sunday blues by making it a Tuesday on a Sunday over at jblive.tv 01:27:01.733 --> 01:27:04.963 in your local time at jupyterbroadcasting.com slash calendar or stream live 01:27:04.963 --> 01:27:07.721 in your podcasting 2.0 app of choice. 01:27:08.018 --> 01:27:10.543 And, of course, the Mumble Room's going. We've got details on the website. 01:27:10.543 --> 01:27:15.865 The Mumble Room gets a low latency, high quality Opus stream. 01:27:16.464 --> 01:27:19.813 It's pretty nice. Shout out to our members for making the show possible and 01:27:19.813 --> 01:27:22.975 keeping us going. Everybody who supports us at boost.jupyterbroadcasting.com. 01:27:23.359 --> 01:27:26.103 But that's it for this week's episode. Thank you so much for joining us, 01:27:26.103 --> 01:27:29.234 and we'll see you back here next Tuesday, as in Sunday.