LINUX Unplugged - 636: Engineering the Future

Episode Date: October 13, 2025

We're back from Texas just in time to chat with Jon Seager, Canonical's VP of Engineering, and their new era with Ubuntu 25.10. On the way, we visit System76 in Denver where the COSMIC team has surprises waiting for us.Sponsored By:Managed Nebula: Meet Managed Nebula from Defined Networking. A decentralized VPN built on the open-source Nebula platform that we love. 1Password Extended Access Management: 1Password Extended Access Management is a device trust solution for companies with Okta, and they ensure that if a device isn't trusted and secure, it can't log into your cloud apps. Support LINUX UnpluggedLinks:💥 Gets Sats Quick and Easy with Strike📻 LINUX Unplugged on Fountain.FMCOSMIC — Next generation Cosmic desktop environment. Use it on your favorite distro!System76 — Powerful Linux computers designed by nerds, engineered by experts, handcrafted by humans in the US.GitHub - pop-os/cosmic-epoch: Next generation Cosmic desktop environmentGitHub - pop-os/cosmic-applets: WIP applets for cosmic-panelPop!_OS 24.04 (with COSMIC Desktop) Just Hit Beta - OMG! UbuntuCanonical releases Ubuntu 25.10 Questing QuokkaUbuntu Engineering in 2025: A Retrospective · Jon SeagerUbuntu 25.10 Fix Pending For Broken Flatpak Support - PhoronixLINUX Unplugged 634: Config ConfessionsHybrid Pick: ncspotPick: poop (Performance Optimizer Observation Platform) — This command line tool uses Linux's perf_event_open functionality to compare the performance of multiple commands with a colorful terminal user interface.sharkdp/hyperfine: A command-line benchmarking tool

Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris. My name is Wes. And my name is Brent. Hello, gentlemen. Coming up on the show this week, it was a race against time to leave Texas Linux Fest. We had a date at System 76 to get the inside scoop on the cosmic desktop and its future. And then we had a chance to chat with John Seeger, the VP of Engineering at Canonical,

fresh off the 25-10 release, but we had to make it back to the studio in time for our live chat. We'll tell you all about that and how it went and play those chats. And then we'll round out the show with some great boosts, some picks, and a lot more. So before we get there, let's say time-appropriate greetings to our live virtual lug. Hello, Mumble Room. Hey, Chris. Hey, Russell. Hello, hello. Hello. Hello. Hello. Hello. You know, if you want to join our Mumble Room, If you're listening out there, it's pretty fun. It's got a live vibe.

Details are Jupyterbroadcasting.com slash mumble. It's an open invite. We'll have a conversation with anyone. We have an open mumble room. We just check your mic to make sure everything's sounding good enough to go on air. And then you can come and chat with us during the show even or during the pre-impos show, which is generally when the mumble room is really going. And it's a lot of fun.

That's Jupyterbroadcasting.com slash mumble. And a big good morning to our friends at Defined Networking. Go to Defined.net slash unplugged. Check out Managed Nebula. It's a decentralized VPN built on the amazing Nebula platform that Wes and I are huge fans of, and we're going to convert Brent soon. Oh, yeah. We're coming for you, Brent.

Watch out, Moose. I'm open to it. I know you are. I'm just teasing. Unlike traditional VPNs, Nebula has a decentralized design, so your network is resilient. If you manage it yourself for your enterprise for a home lab or maybe even a global enterprise, Or you can take advantage of their totally managed product with 100 devices, totally free when you go to define.net slash unplugged. Nebula is really something special.

It was originally developed in 2017 to securely connect Slack's global infrastructure. I mean, I'm talking all over the world, everybody's data, some of the most important companies in the world, has to be rock solid and secure. It had to meet those goals, that flexibility, that decentralization, that security from day one. That was 2017. That's where we were at in 2017. It has grown incredibly. Now we have the managed product, clients for end users, and Rivians on the highway are running Nebula.

I mean, I'm talking industrial grade stuff. And if you want to, you can completely own the entire stack. Or you can take advantage for your friends, your family, yourself, of their totally managed product with 100 hosts absolutely free, no credit card required. Go check it out. Go to Defined.net slash unplugged. Redefine your VPN experience and check out Nebula.

Define.com. unplugged. Well, we have made it back from our escape from Texas, so I guess not to spoil it, but we did make it. And it was such a great trip. And before we get into the story and our trip to Denver to chat with the System 76 crew, I want to just share the team's gratitude to our audience who made this possible. You got us down there and you got us back. And we didn't have any incidents. Nobody, nobody blew out a tire, Nobody got a ticket. There was no fender bender.

I mean, collect, do you know the numbers, Wes, what we drove collectively? I mean, we drove collectively a ton of miles. Oh, yeah. Oh, here. Well, we've got them live up there if you just want the miles on the tracker. Oh, yeah, right. Texas tracker is still live because Brent's actually still on the road. That's right.

So you and I team Bigfoot, we drove a total path distance of 4,732 miles. And Moose there, who's not done yet, is that 3,000. It's really awesome. We had a chance to connect with a great community down there, but then it wasn't even initially on our plan, but because we were working for the audience and not a particular sponsor, we were able to do a last minute schedule with System 76 to swing by and get a hands-on demonstration of the latest and greatest with Cosmic. I just want to underscore, not only did it mean that our entire intention was always just to create great content for you, but we had a kind of flexibility. that I just don't think would have been there. And so we got to do something really special. And it meant we were on a very tight timeline. Okay, so the goal was to be out by 9.30. It's 929 a.m.

We're doing pretty good. What are you doing right now? Well, I got to unplug the van and put the solar panel in the bag. I know if your bags aren't loaded either. Well, I staged them beside the van. You probably got to feed the kitties a little bit of food. Well, they are jumping around like they want some food. Uh-huh.

Yeah. I don't know about you, Wes, but I'm kind of ready to go. Yeah, let's hit the road. We got miles to go today. Jerks. You and I had it down at this point. Yeah, well practiced, I think. Yeah.

And to be fair, Brent has a little bit more to take care of. Oh, yeah, we were doing lean and mean, and that was definitely an advantage for us. Brent's got animals to take care of, a van to take care of, and a Brent to take care of. It was clear in that moment the difference between the types of trips we had had up to this point because I was, you know, let's say, meandering my way to Texas and very happily so. And you guys were just jetting down there. And this is the collision of those two philosophies. And it was clear that we were going to have to figure it out. And just as we're ready to go, I mean really ready to go,

Brandon informs us that he's had a little issue that he hasn't told us about. You got all those belt? I think so. Maybe. I don't hear anything. No, it's only when I really give it gas. When you give it the onions? Yeah, like if I'm pulling out of a parking lot or trying to pass you.

Uh, so maybe... Since when the engine's really rev? Yeah, maybe I'll go in there and get... You want to rev on it? A couple kicks, and you could... Well, if I were to guess, I would say alternator. Oh, no. Because it's coming from the bottom corner.

This was a theme for the day. It should have been our indication of how the day was going to go because we didn't want to push the van too hard. We weren't in a particular time crunch for that day because we just had to get from Austin to Amarillo. So we had, you know, only a seven-hour drive day. So we thought maybe we would make it in decent time, get to Amarillo, do some grocery shopping, stock the Airbnb. That's not exactly how things went. Yeah, we just had 80,000 miles on the car. All right.

So we've been driving for 395 miles, 6 hours and 44 minutes. we're at the fourth O'Reilly of the day Hey, one of those was an auto zone We're going to go with Brent And we're going to get a alternator that he doesn't need And then we'll be on our way to the Airbnb Definitely moving a bit slower today We did ultimately get there

And Brent did get his alternator Which I'm kind of in favor of now Well, okay, because you never express in real time That I certainly didn't need it It was like community voting, you know, all of us But you seem to just go with the crazy Brent plans Well, I was, you know, I was like it's a belt. I think it's a belt first and foremost.

And if there's another issue because of, you know, that, then we'll deal with that. But I was like, really, it's just a belt problem. But, you know, you got a better unit. So when you do get it installed, I actually think you're going to be better off. And we did get to our Airbnb. Nothing really stands out to me about that one. They all kind of have, I don't know about you guys, but they all kind of just merged together as one meta Airbnb. Well, there was one, I don't think we could forget.

Well, yeah. Yeah, the one where we had somebody else's stuff in there. Oh, no, no, okay, too. No, the one where you didn't really fit in the shower. Oh, that was this Airbnb, wasn't it? No, no, it was the Denver one. It was multiple. Oh, yeah.

Oh, right. It was multiple. Guys, not one of our Airbnb's had a functional bathroom for me. Not one of them. They all had something weird. I don't know what's going on with Airbnb's and bathrooms. But we got there the next morning. It was time to set off because we needed to get to Denver.

This is where the timeline got tight because Carl had a, Carl, the CEO of System 76, had scheduled him. a work from home week and he came in gracefully because we were coming into town as well as so had jeremy sawler and so they both came in and scheduled to come into the office because we were going to be there to help you know give this presentation and so we really wanted to make it on time because we were kind of taking from their planned work from home week and i i didn't intend for that to happen to happen to be here maybe something will work out and that was very gracious of them we're like okay we're in amarillo but we need to be there tomorrow by 10 a.m. Oh, I wouldn't, I wouldn't call it fixed.

No. I'd call it. Maudly improved? Not even? No. But we may have fixed on a leak on the radiator. Yeah, that was easy. Unrelated.

So that's a win. We'll stack that. Let me get down the road a little bit because we've got to get the hell out of this Airbnb. And see how it does. We'll check on her. I'm worried. Well, this is the ultimate test. I mean, you're really pushing all the components, long duration driving, heat.

miles and miles. I mean, this is it, right? If things are going to break or crack. You're talking about just getting it over to O'Reilly's? Oh, man. We did it. And I do remember this, so we have a process, right? We get to an Airbnb.

And within minutes, we try to get the Wi-Fi set up. We try to get Star Trek going on the main screen. Yeah, time to Star Trek. That's our standard Airbnb metric. Yeah, and it is an interesting one because it's like, you've got to get the network stood up. You've got to get a device working. You've got to get a client device paired with that.

And what's the TV setup, right? Can you use the tech that's there? Do you have to inject your own? How accessible is it? And every time we do this, we kind of like refine this idea of this ultimate tiny Linux box that has HTML that boots right into a Cody that just immediately starts randomly playing Star Trek and sets up an AP that all our devices connect to. Like we've been iterating on this idea.

So if you have any suggestions for a device like that, please do boost them in or send us an email at Linuxunplug.com slash contact because we want to build something. when we get to an Airbnb, this, like a little Linux router slash Cody Box slash Wi-Fi AP repeater. Also Nebula device. That's what we're going for. So that's, but it is this Airbnb we arrived at turned out to be in somebody's basement, which we didn't expect. And so I also did not fit in that. And not a particularly tall guy.

I'm, I think, 5-11, if that's, you know, it's like not that unusual. But yeah, it was, it wasn't ideal, but we made it to System 76. time. Wes and I are just outside the System 76 office, waiting for Brent to feed his cats and get the audio interface that he forgot in the van somewhere. So we're kind of on time. We're on Brent time. There's no escaping Brent time. But his cats are so cute. Hey. Hey, what? What are you guys doing over here? No, I've been just waiting for you, buddy. Yeah. Okay. And we got there and, you know, we show up. Every time we go, they've changed so much. They've, they've, they've, expanded,

rearranged the office. So it's a new entrance. A bunch of new tech, bunch of new people. It just changes all the time. A lot more lasers. There's a lot of lasers now. We got there.

It was great to see Carl. He gave us a tour of all the new stuff and some things they're working on. And it was also the day they were releasing the new Oric. So that was kind of exciting. And then we sat down with him and Jeremy in an office space to try to just get the state of cosmic because they're getting really close right now. but there's still some things that are missing. But there's also a lot of stuff regarding Cosmic

that people have not discussed, including some features that have recently been added that they touch on in our chat here that are just blow away cool. We just had a great tour. Carl gave us a tour of the factory, to get an update on everything, but we're really here to get an update on Cosmic

because the last time we chatted about it, it was early Alpha. It must have been like at a Linux Fest or something. It was a while ago. And I think things have come a lot further. Yeah, I think that was, Linux Fest Northwest, and is it Big Bertha? Is it in Bellingham?

But, no, it was in the RV. Jubes! Right, you came in Lady Jubes and gave us a demo. Right. Right. So, yeah, at that point, I think we were talking about where Cosmic was. The talk I was giving was on apps because of my second time around at Linux Fest Northwest. I want to do some new content.

So Jeremy actually did a lot of the work on the apps, and I wanted to highlight how unique they are because there's a lot of features that you just don't get in a lot of desktop apps on Linux. So we've been kind of looking at a couple of things just before we started recording, but things like fancy background switching for users, at creators, and it looks like we're getting to the polish aspect of Cosmic now. Yeah, there was a long gap between Alpha 7, which was in April, and our beta release at the end of September. I guess that's whatever, six months or something like that, five months.

And after the Alpha 7, one of the reasons that we did Alpha's frequently was because we kept getting the same bugs over and over again, even though they had been resolved. So unless we were tagging releases and those were being adopted, that's because Cosmic's package for Arch and Fedora and, and suicing, things like that. So unless we were tagging frequently, just the triage and the amount of effort to go through things that had already been solved was pretty high. But Alpha 7 was kind of a point where things improved considerably at that point. And so we didn't get a lot of issues afterwards, and there were unique ones that came after that. So the beta release marked that point where all the features, or by and large, all the features are in,

and now we're working on, yes, smaller bug fixes and polish for the final release. Jeremy's here too, and I wonder, Jeremy, if you have any comments on sort of just the state of the system and where it's at from a code development standpoint, stability, things like that. Yeah, sure. So as Carl said, we nailed down most of the new features already, except for a whole bunch of things that we're adding. Like, a cosmic store needs to have flatpack add-ons. We're doing printing support in Cosmic Edit, which is more nuts than I thought it would be.

But in general, the system's pretty stable, and we're just working through the backlog of issues that people have reported during the alpha and beta phase and trying to fix them all before RC. And we actually started shipping a unit with the cosmic beta. So the Orix Pro with the codename Orup 13 is an AMD and NVIDIA hybrid system, and it turns out that it works best in cosmic because of the way that the cosmic. handles hybrid graphics. Right. So the Orcs Pro has just been released today. It's October 8th. So it's a brand new product with StrixPoint CPU,

Nvidia 5070 graphics. And what's really unique about Strict's point is that the VGA or the controller, the integrated GPU, is not VGA compatible. Meaning that it requires way, because it will not work with X-11 because it's not a VGA-compatible GPU. So as we were working through the products

like Ubuntu 2404, if you write it on X-11, the discrete GPU is always on because X-11 can't use the integrated GPU. In Wayland, that does work correctly. And so when we're shipping Ubuntu 2404, we're also shipping that with Wayland on that product.

And we weren't going to go back to 2204 for Pop and find out what the Wayland State was there when we have this brand-new desktop that's really close and, frankly, at this point, a better product than what we're shipping with 2204. So the Orix becomes our first product to ship with Cosmic, one by necessity and also because we think we're shipping a better product to our customers. That's a pretty big milestone. So we're going to kind of play with things. things off mic, but I'm curious because you guys have been kind of running this, the longest out of almost anybody on the internet. What are people missing when it comes to what's great

about using Cosmic? Like when you hear people discussing Cosmic and arguing Plasma versus Ghanome and Cosmic, what's the thing that people are missing? The highlight of Cosmic is that it adapts to whatever your workflow happens to be and it has all this ability to do that, but none of it feels overwhelming. It all feels natural. It's really, it's easy to get get in the weeds. Sometimes I think people use Cosmic and they say, well, there's not like a lot of new things in here. I think it's because the design is just so good that when you get to using it and moving things around, it just all works and it just feels so simple and natural that the innovation is taking all these advanced concepts and making them that approachable to every single

user. So you don't have to be an engineer to take event of tiling or to theme your system and make it look exactly the way you want it to look. You can go in and without being overwhelmed by a whole bunch of settings, click a few things, and then you have a dramatically different workflow or experience based on what feels right to you. And it's so easy to switch between different styles of using Cosmic, whether you want to launch apps with the launcher and switch. between them, or if you want to use the application library, or if you prefer to use workspaces or not use workspaces, or use tiling, or use hybrid, you know, tiling in some, you know, maybe tiling on your external display and floating on your laptop display.

Whatever fits right for you, it's just very easy and natural to set it up and to try different things. And to me, there's kind of three things in addition to that. We're trying to make, or what Cosmic does is make computing fun and productive and personal. And I don't know, it's been, it kind of takes some effort to make computing fun. We spend so much time on our computers. They should be fun to spend time at. I mean, we're at them for hours. And so, for me, fun is, like, customizing my theme, changing my layout, trying different things.

I enjoy doing this stuff. For other people, it's gaming, so we want to make sure gaming's fantastic. For other people, what might matter the most is I don't have to click four things to log out. In Cosmic, you just click the button, click Logout. It's like everything is a shorter path to what you need in Cosmic than a lot of other operating system experiences. Jimmy, I'm curious if you have anything to add on what people are missing when they're comparing the different desktop environments today and they're thinking about Cosmic or they're comparing cosmic to plasma or gnom or things like that?

Well, one thing I think I see is missing is people judge. Often they judge cosmic based on its current state when they should be looking forwards to the future. Every release that we do, we will have a feature cycle where we build up a large number of features. And I don't think we've ever told anybody no. We just say, no, this is not how we want to do it, but we will try to adjust to that use case.

And there's so many cases where, like, when I think about all the accessibility, work we did. That was really one where we had to tow the line between how do we make sure everything is there for every person who could potentially use this product and every potential issue they may have with the product, but still make the default experience good for the most, the largest number of people. And yeah, that's what I wish people would consider is how would you see this modular desktop environment evolve into the future, rather than judging it by how it is right now.

Also add to that, when you have an established customer base and user base like PopoS has and System 76 has, it's not like starting from scratch with no users and building something new. You have a responsibility to what their experience already has been and what their experience is going to be when you deliver something new to them. And so when we're building Cosmic, we're building a platform that's going to be familiar to them. But once you get into Cosmic, you realize that this platform. because of the way it can be composed and the way it's modular, it can actually become anything that you want it to be.

And that's, I think, what will be, it's key feature and what people will really appreciate long term is I see folks say, it looks like Mac, it looks like Nome, it looks like, it looks like, it looks like KDE, it looks like this, exactly. It does, it can look like any one of those. and to get that layout that you really enjoy and that you prefer to use, it's just so simple that you can just lay it out the way you want. So, yeah, it's unique in that way. I guess I'm sort of surprised we've been talking, and the architecture hasn't come up, rust hasn't come up, the way the system settings are stored,

which is kind of different and unique hasn't come up. Jeremy, can you touch on some of that for us? Yeah, absolutely. So we try as much as possible to keep things modular. And this is another thing where I want to see it evolve over time. Like we just added third-party applets into the Cosmic Store. So you can go in. These are things that anybody can make, and they're securely added to the panel

because they're flat packs and they're sandboxed. So every applet that you get from the Applet store, you can add to the panel. You can create docks and panels anywhere you want on the system, on any display, any number. and Carl is laughing right now. I would definitely have too many. I'll be that guy, I'll admit it. You'd have the 1990s browser toolbar kind of look. Yeah, oh yeah.

And so the ability to have every single one of those be a separate process and be in its own sandbox means that they can be written in any language. So the rust aspect of cosmic actually doesn't matter as much as people want it to be, especially now that Rust has become like a political statement somehow. I don't care of my own politics, but Rust is just a programming language, and the purpose of it is to drive memory safety. And the reason we're using it is because it allows us to say these types of bugs

cannot exist in Safe Rust Code. And so long as we're writing Safe Rust Code, then we eliminate those bugs or at least reduce the prevalence of bugs overall in the entire desktop environment. The config mechanism is another interesting part of it. It's one that was driven by portability and modularity. It's how can we make it so that the same config mechanism could be used by cosmic apps, regardless of what operating system they're running on.

So they're running on Windows, MacOS, Linux, BSD, it doesn't matter. They don't need a third-party service to exist. They don't need a Damon setup. They don't need DeConf to exist. they just need to be able to write to files. And so we have a flat file format where they just write out in the Rust Object Notation format to those files. And you can manage those files in a Git repository if you'd like to, or in any way you want to sync files between systems. And you can diff those files very easily.

They're all plain text. And another thing that we've done is to ensure that all the state is separated out from the config. and what I mean by state is if you're in KDE and you're trying to manage your config files, you might, for instance, notice that the size of every console window is stored. The last window size is stored in the same config file that saves things like your keyboard shortcuts for the console application. So we've separated those kinds of things out. Recent files would be stored in state versus the like the theme set. settings would be stored in config, and that ensures that you're able to sync those config

settings without bringing over state that doesn't necessarily apply between computers. Yeah, I think that extends to our lack of requirement for System D as well. The portability means that Cosmic will be a great desktop for BSDs. It works on Redox, and so we try to make decisions that don't eliminate its use in other platforms. Yeah, I'm just curious. How has it been? It seems like Cosmic's spreading its wings to a lot of other platforms and distributions. Have you gotten good feedback from people trying to package it? I've gotten feedback, and it may be anecdotal that this is the easiest desktop environment to port to package overall. There were a few hiccups with rust in the start. Like Fedora wanted to bundle all the rust, take all the rust dependencies, and turn them into dynamic libraries. But then they changed their mind, and now they're allowing those to be compiled by car.

just the native rust build system. And so, yeah, it's spread across to Arch, to open Sousay Nix, to Alpine, to Fedora, and onwards, because it is very easy to port, and it has very limited dependencies on the system. All right, well, I'm just looking forward to playing around with it. Thanks, guys. Yeah, really, a big thank you to everyone over at System 76 for making time on our last-minute scheduled stop. and I'm going to give it a go.

I walked away pretty impressed. One of the things that impressed me the most is the way they've implemented applets. And that's, to me, a massive innovation in the desktop space is essentially any flat pack app with the right dimensions. It's a wayland application.

Yeah, it's long as you can draw a window on the screen and, you know, it has their right interfaces. It'll work. You can make it an applet. And the massive ramification of that is all the dependencies, your language, all of that's handled by, you know, the flat pack, but it also means

if the desktop updates, your extensions are, unless something changes in the XTG portals, your extensions aren't going to break. It's a brilliant, genius way to do extensions on a desktop that works with a desktop that might be quasi-rolling on some

systems. Yeah, stay tuned. We'll see if we can get Krista Vibe code in a little extension. So I've been running Plasma, you know, in Aurora here on my main machine in the studio, and I think we're going to Nix Cosmic it up, and I'm going to give Cosmic a on this see if I can kind of replicate a workflow I really like in the studio with it. I'm just, I was just really impressed.

Anything stand out to you, gentlemen, about Cosmic? Yeah, I mean, I still really want to sit and live with it for a while to give it a proper test, of course. But yeah, just, you know, watching Carl use his Cosmic, watching him kind of live tweak it and change it in front of us with, I mean, he wasn't even really sitting in front of a proper workstation environment. He's kind of having to, you know, show us the screen and tweak things. as part of the presentation. So if you can do it in that environment, it seems like now we have this lean, mean, portable, flexible environment. So even if it doesn't end up maybe being your daily driver on every machine,

it seems like a very useful tool that you can kind of take with you. For me, I was really impressed by their dedication to, like, the design of the architecture on the back end. And they specifically said, you know, we're trying to make production software. And so we want to do it right, even if it takes longer. And I think that philosophy is really playing out in a nice way now with cosmic, you know, coming into beta and us getting that behind the scenes look of how it's built. And that also means they've been working really closely with a bunch of upstream that I found really impressive and is just making our Linux ecosystem better overall.

So it was really impressed by all of that. I think, too, one of the things that I took away from it is there's components that people, might end up using even if they don't use the entire cosmic desktop of so on hyperland you don't really have a file manager and you get to pick and choose and one of the things I experimented with is I installed the cosmic file manager because it it doesn't require nearly as much as say dolphin or files does from Ghanome and it worked really well so there's components they're working on here that are pretty portable but as you also heard Carl say they don't have a system de-dependency not as a political statement but just as a function of how they've built it which means perhaps

this will be available for desktops that want a modern desktop that don't have system D in the distro or maybe one of the BSDs out there that obviously won't have system D. And so it has, that's kind of encouraging that way because it has some potential audience that maybe modern Ghanome and plasma are

leaving behind. I have to say I was also really quite impressed by just how many resources they're putting to the desktop. They mentioned 10 people working on Cosmic full time, including QA, of course. And I specifically Carl said, well, we actually like that to be like three or four times that size in the future.

And so they're really dedicating to this desktop, which is, yeah, really impressive. Yeah. And I love, too, that their philosophy towards a rolling release desktop, he mentioned that a couple of times. And the idea that the desktop can be decoupled from the distro, and that's the way they're building it even for Pop OS. It was a really interesting concept, and I think has a lot of really really. interesting benefits. So I'll give it a go and report back in the future after I've had some time with Cosmic. And before we even got out of the parking lot at System 76, our good buddy editor Drew showed up for a quick reunite while we were in town. And, well, I think we'll get into

what happened afterwards later. Look who it is, boys. Hey, is that Drew? Is that guy who makes us not sound like shit? Hey, hi. Hello. Hello. Did we see you? Does it have been like two years, three years, eight years? We saw him for... No, it's less for you because we were here for Denver like two years. Oh, right, yeah, for the Red Hat. It wasn't, we didn't see much. It was like, still looking handsome as ever.

Yeah, right? What are you got there? That's for you, right? Oh, it's a DC to DC Charger. Yay! Hey! That sounds like one. There you go.

After your alternating, you got a whole new project. Can we open it? That's very exciting. He needs this. Just to get you to stop bugging and to buy it. It's great, Brent. Yeah. Yeah.

It looks like it's in great shape. Oh, I don't have it. It does. Yeah, yeah, that's great. Has this thing even been used? No, it looks beautiful. It looks great. It's half the price of everything else.

Assuming it works, I think it might be okay. Yeah. Good. There you go. I'm going blue. I mean, look how clean it is. Yeah, yeah. We've got to get it dirty quick.

Yeah, seriously, that looks like that's never been used. I mean, look at the labels. The person selling it only had one other eBay item, which was like a Louis Vuitton. super high-end purse I was banking on the fact they didn't know what they were doing yeah no kidding well what do you think of the van

this is your first time seeing in person it is yeah what do you think it's uh can we get more descriptors a little bit of rest yeah it's a vibe no I mean you know it's it's

it's an old ram it's an old ram he says you know so just a quick side Again, I'm so grateful to the audience for making this trip possible, gave us a chance to reunite with Drew, which is always really good to just reconnect. And Brent then began to work his magic. He gave Drew the tour of the van, so we captured Editor Drew's post-van tour impressions to see if they improved. So now that you've had the full tour, what's the refined opinion? Dude, that's awesome.

Really? You get them. You got him to come around. I mean, you know, I kind of figured it would be one of those things where, like, you're looking at it from the outside. and it doesn't look like much. It looks like a 30-year-old van. But then you get inside, and it's like, hey, look at this. This thing's cool.

This thing's cool. Every direction you look, there's something cool. There's something weird. It's just, yeah, it's awesome. So you got him. You turned him around, Brent. Good job. Well, thank you.

Yeah, I was a little, I got to say, underwhelmed by his initial reaction. I was like, this won't do. We got to work harder here. And sure enough, I got him with the home assistant in the cover there. He's like, what, how do you fit that in there? And he's really impressed. So thanks, Drew. Yes, it was really great to see you, Drew.

Wes and I, however, we had to roll. We still had 22 hours to get it back to the studio, which meant driving through most of Montana, most of Wyoming, and, of course, all of Idaho and Washington. We really had to put down, like, significant miles. And when we left, it was a little sad. Okay, Wes and I are in my car. We are leaving Brent behind with Drew.

So I think we'll be out of Denver probably in the next two hours or so because traffic's pretty bad. So an hour or so. How long do you think until Brent's out of Denver? Well, he's about to go to lunch or really dinner at this point with Drew. Is he going to leave Denver at like 8 p.m. at night? Probably not. So that's at least tomorrow.

But, you know, we ended up fixing on his brakes instead of fixing on his belts or his alternator, which he also has a new one. and you know Maybe he gets working on that Maybe he gets working on that I think he doesn't leave till tomorrow I'm betting I'm betting Friday

Okay Brent's survey says Well Yeah Drew and I had an amazing time together And we might have you know sushi and sake and like lots of chats and stuff And then I was like I may as well just stay overnight But I didn't really

Have a plan for that So I have to say thank you to System 76 For letting me sleep in their parking lot. A really nice party lot. I got to say nice and quiet, other than the loudspeaker that comes on at 5.30 in the morning and says, hey, you're being recorded on video. So, highly recommended. They were kind enough to not bother me in the morning. And so I got going in the morning, but, you know, I forgot some of my stuff in Drew's car. So I had to, you know, see Drew again. And he had a whole beautiful, like, care package for me.

like van supplies and things he thought would be helpful including a giant bundle of rope and it was it was really sweet yeah wow true so great wow well so we were putting down miles uh west and i on the on the route home uh totally leaned in to take advantage of tactical wife planning she was great because she was watching us on the tracker and so she was constantly looking into okay there's an Airbnb about an hour out or two hours out from you you have these options, and then we would say, yeah, we'll make it this far, then she'd book it for us. And all the details would just be handled for us, and we'd just show up and crash. Yeah, we got a message with the door code and the Wi-Fi.

Incredible. Yeah, so we were able to move pretty quick because we had to. We wanted to make it back here in time for John. It's the final push of the trip. Wes and I are in Washington State, in eastern Washington State. We stopped at Wes's absolute favorite restaurant. Burger King. When the King tells you to get breakfast, you stop, you listen.

So we filled up with gas, ordered some fully loaded croissant witches. Our way. Yeah, always our way. And we're waiting on some French toasticks because I assume the syrup's really good. And I wanted to get that. No, Wes is grabbing the French toast sticks. So we have about five hours or so on the road. I probably have another six or seven after I drop West off on another couple hours.

So six or seven hours total for me. And then the trip's all done. No idea where Brent's headed right now. He's going in the opposite direction of home, so we'll figure that out at some point. But hopefully he's having a good time. It was a great time. And so Brent's total hours on the trip driving, you put in a massive 59.44 hours of continuous drive time. That's like active motion time, right?

Yeah, yeah. And Wes and I put in a total drive time of 75. Which actually, I turned off the tracker in Seattle, so I continued to drive for another couple hours after that. Yeah, that means you ended up driving for probably over eight hours last day. Sorry. Yeah, yeah. Really amazing that there was really no incidents or any problems and that it went so smoothly and that we got to do the meetups. So we just want to do more of it next time. We want to do more meetups on the route, maybe different routes, and things like that because it was great. And everything worked out. We made it back here just in time. OnePassword.com slash unplugged. That's the number one password.com and then unplugged, all lowercase. Go take the first steps to better security for your team by securing credentials and protecting every application, even unmanaged shadow IT.

It really does change everything. Learn more at onepassword.com slash unplugged. If you're in security, you're in IT, you know that you have a mountain of assets to protect. You know there's so many devices and identities and applications. It's a lot these days. It really does create a mountain of security risk as well. Fortunately, you can conquer that mountain of security risk with one-password extended access management. It's a real problem IT is trying to tackle right now.

Shadow IT is contractor accounts applications you didn't know, got created credentials, you didn't know we're being used. Well, thankfully, Trellica by One Password can discover all of this, secure access to your apps managed or not. trellica by one password inventories every app in use at your company think of that and they have prepopulated app profiles to assess SaaS security risks they let you manage access optimize your spend to make sure you're not doing any redundant accounts and services you can also enforce security best practices and there's a secure and simple onboarding and offboarding process it makes it easier for IT it makes it easier for your employees everyone gets more efficient and you've got a solution manage that shadow IT. That's the big thing. It's just one of the ways extended access management

from OnePassword helps strengthen compliance and security. It's not easy to really understand what your spend is and how many different applications your users are using to do their job. But Trellica by OnePassword can help you take control. It gives you visibility, gives you automated workflows, and cost optimization. You can check compliance off your list with Trellica by One password. Employees are empowered to be secure with flexible app access and you can manage. at all. It really is great. If I had this when I was in IT, you know, I think I could have stayed in. I might still be in. It really is awesome. Go take the first step to better security for your team by securing credentials and protecting every application. Even the

unmanaged Tieto IT stuff. Go learn more, support the show. Go to onepassword.com slash unplugged. That is the number one password.com slash unplugged. Well, while we were on the road, Ubuntu 2510 was released, and we had a chance to sit down with John Seeger, VP of Engineering at Canonical, and get the inside scoop. John's back on the program. It's been just since the beginning of the year we were talking, and it's a big week because Ubuntu 2510 is out. And John, in your release, you say Ubuntu 2510 is a statement on the intent for the next. Ubuntu LTS in 2026. It sounds like a big deal. Welcome back to the program, sir. Thanks very much for having me. It's good to be back. So what do you mean when you say it's a statement of intent? Are we to say, is that to say, like, what's in this release today is what

we should expect in the LTS? Or what do we mean about it? I think it's two things, right? So yes, indeed, the things that you see in this release, some of which are a little more testy. The plan is to get those stable and kind of well understood in the hope that we can ship them for the LTS. But it also marks, I think, think a, hopefully what I think will be a change in Ubuntu where we try to have slightly more exciting releases going forward. So we have to dial back that excitement a little bit for the LTS and go a little bit harder on the kind of reliability and stability that our LTS customers expect. But my observation when I came into sort of looking after Ubuntu at canonical was that

the interim releases that kind of stopped being so crazy. And I'd like to make them crazy again. Because I think we're in a position where we can use our extraordinary user base and, you know, rather excellent community to push certain things that might actually have an impact beyond just Ubuntu, right? Like, it's not just about Ubuntu. It's about trying to advance Linux in general, whether that be on the server or the desktop or devices or whatever that might be. So how does that manifest then? Because one of the things that I know you've been focusing on is sort of a faster release cadence, I think monthly snapshots. do you view it as you put those contributions into Ubuntu and hope that the wider Linux community takes them upstream? Or is it more fundamental like the tooling you're choosing and you're hoping to have an impact there? Can you expand on what you mean by that? Yeah, so the two are somewhat different. The monthly releasing is based on some wisdom that I

have picked up over the years working with software teams, which is if releasing is hard, you should do it more. And releasing was hard. releasing Ubuntu was something we only did every six months. It was a relatively dated process and I wanted to modernise that process and only getting a chance to think about that process

every six months is not necessarily very conducive to us making it any better. So the monthly releases was really a tool for the release management team that may not have felt like that for them the first time. It was a tool for them to understand their own process better and improve their understanding and it really shows. So this time

we have a release sprint when we release Ubuntu. We we get together a few of us in our London office. And this time we were all done, like hours early, just waiting to release the press release at the time we said we would release. And that has basically never happened, as far as I can tell. The tooling side of things is more where I want to push. So this is things like the Rust core utils and the Rust pseudo implementation and potentially some of the work we're doing around TPM full disk encryption and a few other initiatives like that.

I see. So I would not, I don't really know, but to me it would seem like, what you're trying to do is push on processes that were designed to improve stability and reliability of the release. So have you gotten pushback saying, John, you're changing things that we put in place on purpose? Like we

had lessons learned and that's why we do it this way and you want to change it so we can release faster. Has that been a problem? A little, but I would temper that with I am asking questions and pushing but not being reckless. Like I I've been accused of trying to turn Ubuntu into various other things, like maybe Nixosius, you've heard me waxing lyrical about that before. And that's not the purpose.

The purpose is to try and understand how we could evolve Ubuntu, to give ourselves and our customers, be those tinkerers at home or our paid customers, the assurance they need that Ubuntu is the stable platform, the reliable platform, the resilient platform they want, but also delivers on the mission that Ubuntu was started with, right, which is all about delivering the very best of open source to a really wide audience. So back in the day, I wasn't around in the next scene then, but when Ubuntu first kind of forked off from Debian or became a downstream of Debian,

the whole point was to provide an area where we could ship newer, more exciting software that perhaps was more complicated in Debian because of licensing or policy or process. And I think that's fine. I think the two should exist. I don't think Debian should do what we do, and I don't think we should do what Debian does in the same way. I don't think we should do what Art or Nix do, right? But that doesn't mean that the processes that we have used for the last 20 years are unchangeable.

A lot has changed in software engineering in 20 years, and I would like our processes to reflect that way possible. So it's mostly about more automated testing, more automation in the release process, which should, if we do it well, enable us to have more trust in that release process, not less, I think, and spend less time on the mechanics of releasing and more time on polishing what it is that we're releasing. Right. The release process should not be the exciting part about an Ubuntu release.

The features that we ship and the tools that we ship should be the exciting piece. Amen to that. Well put. Okay, so kind of along the same lines of hurting cats a little bit. I feel like there's more communication along the development process, in part from you, but others on the team were more on the form, blogging. Has that been, I don't know if it's mandates the right word? Is there been encouragement there?

What's that communication like to people that are busy working on Ubuntu? and focused on actually doing the development or doing the QA, how do you convince them we need you to be more public about what you're working on? And has that gone over? Yeah. So I'll be honest, it kind of was a mandate to start with. I think what happened is canonical as it grew and became more grown up as a company. We adopted things like Google Workplace and Mattermost,

and a lot of the communication that would previously have been public in IRC and on forums kind of naturally ended up in Mattermost and on Google Docs. because that's where we were all hanging out as canonical employees, right? And so I think lots of the process, lots of the discussion around Ubuntu suddenly became quite opaque. And one of my hopes was that I could change that a little bit. And so I did actually essentially mandate that every team that I look after. So that's the Ubuntu desktop server foundations

and more recently the Deb Crafter's team. I mandated essentially there should be at least a post every week from one of us. And to start with, they just did like a ram robin. So I said, I'll go this week. you know, Foundation is the following week, server the next week, Deb Crafters the next week, and we'll just go on a rotation. And I think that was quite tough for the team, and I'm really proud of how they rose to that.

But what's been interesting is that other people on those teams have seen the engagement that those posts were getting and the interest that it drives to the project and started following suit without me asking, right? And now I think it is quite clear that we have a group of developers in Ubuntu Engineering who want to talk about what they're doing because they find it interesting and they like it when people ask them questions about it and give discussion

and thoughts for improvement and things like that. So it started out as a mandate. I do, I put my hand up, there is still a bit of a mandate. I do expect there to be something every week from Ubuntu engineering, but there's much more than that happening, which I'm really pleased to see. Well, it's great to see from the outside. It feels like life. Yeah, I mean, it's good for us, just for our business,

but it feels good for the community too, right? There's more to pay attention to, more to keep Ubuntu fresh in your mind. Yeah. All right, so let's talk about 2510. We've got a new Gnome 49 in there, right? so that means we're going to get a lot of nice stuff like HDR. What should we talk about with 2510? What's the big thing on your mind?

The big controversial thing, perhaps controversial, I don't know, is the Rust Utilities, right? So we replaced Canu-Core Util's with the U-U-U-TILs implementation, and we replaced Sulu with SOURS. And they, to varying degrees, represent the most risk, I suppose, but I feel comfortable about it, particularly Sulu-R-S, which is a much tighter in scope, and therefore it's easier for us to kind of reason about the impact it's likely to have. But the relationship that we've got with the UUTils project and the Trifect Foundation has been phenomenal. We've been speaking with them pretty much weekly, if not more regularly throughout the entire cycle.

They've been super responsive to bugs and feedback. And so really I see 2510 as a bit of a proving ground. There are still a handful of utilities that we have diverted back to the good implementations where we see that there are issues still. Because at the end of the day, as much as I would love to push this. I'm not going to do it irresponsibly and just push broken stuff to people's machines, right? So the idea is that we would iron out all of those issues throughout this cycle and for 2604 be as close to a full replacement as we possibly can. What do you consider to be a major issue with the rest tooling right now? So we have an issue that I think was reported on feroics where there is a slight discrepancy between how the MD5 checksum tool behaves in

certain cases, which impacts make self-archives, which is used as part of the fronics test suite. there is an issue with directory traversal so some of the code for the tools that have the ability to recourse through directories maybe that is trying to the example anything like LS or CP or move or anything like that the directory traversal code was written quite naively and could could be left vulnerable so we kind of waited for that to be fixed before we moved across there's been a handful of issues some of them really tiny some of them less tiny there was one that bit us where the date command output the date in a slightly wrong format and broke a huge number of the package builds in the archive. That's a classic programming thing, right? I mean, who wants to work with dates?

There is a link. I wrote a blog post last week saying kind of a retrospective about this, and there's a link which shows the list of all the utilities that we've diverted back. And the bug tracker is really active, and the maintainer is in there helping us out, along with what appears to be a really nice, like, vibrant active community around that project. You know, you mentioned your relationship with Upstream. And when you hear last time, I think you're pretty hopeful about how this could be good, not only for pushing Ubuntu and the Linux desktop forward, but also for these upstream projects. How has that been going? Yeah. I mean, from my opinion, it's great. So the way this sort of unfolded is I had the idea that I would like to do it. I reached out to Sylvester,

who's the kind of project leader for Utils, and said that I was interested in doing this, and did he think the project was ready? And would he need some funding to help out, which we agreed and kind of move forward. And similarly, for Sulu, right, I reached out to the folks there and said, I'm interested in doing this, what do you think? Both responded with huge enthusiasm. We did a bit of back and forth on some feature work that we would need doing or some targeted effort on bugs.

And yeah, I think they're certainly getting increased interest. We don't know exactly how many of them who use us there are, because we have very limited telemetry anyway, and anyone who has kind of behind a corporate firewall or opts out of it, we don't get it. So we don't know exactly how many there are, but there's definitely more than 10 million desktop users alone. I shudder to think how many server users there are.

So you've got to imagine a pretty enormous increase in their daily users as a result of this. And I mean, I think it's showing, right? Like one way to view this is maybe, oh, okay, it's not fully ready if we're still finding these issues. But another is, okay, the massive scale that Ubuntu brings means you actually finally do find a lot of these edge cases and can really iron out your software to make it, you know, robust. And hopefully they then have some funding. Like, if they can't fix the issue themselves, they have some funding to pay other developers to help them out, right? Like pay bounties.

A lot of the issues have been performance issues so that I found an issue with word count, which made it 10% slower than the Ganoo implementation. I found an issue. Someone found that interesting. Two days later, it was 11% faster than the Ganoon one, right? So there definitely are going to be places where there are rough edges. We certainly haven't exhausted every code path, right? but actually overall, I'm surprised at how surface level much of the problems we've found have been. Yeah, that's my takeaway.

Before we move completely off of Russ, one last question is, it seems like some people that are building on Russ find themselves in this contentious spot where it's a hot topic online. People love debating sort of like these hot button issues in the Linux community, and Rust has become one of them. And it seems like you want to push this forward at the same time not really engage in that. How do you walk that? is that a challenge? You know, is it, is it a brand risk association at all? Those types of

things. I have tried to steer clear of those sorts of conversations not to not address it here. So one of the biggest criticisms that we got for this or I got for this was the change in license of the core utilities package. There were lots of people that were very concerned that the Rust implementation is, it's not GPL, it's MIT. The opinion I've come to is that as far as I can tell, the values of the Utils project are very well aligned with that of Ubuntu. And so in reality, anybody who installs Ubuntu on a machine is entering a sort of trust relationship with Canonical, right? Like, every time we run app, we're root on your machine. That's how apt works. You kind of have to trust us. And so I think my view is that

it's not like we wrote the core Utils code in the first place. We were shipping Ganoos core utils. Their values aligned with ours. The code was, you know, provided something we needed. and now we're shipping a different implementation. And if something were to change, and I felt like, or we felt like the Uyutils project was not representative of the values of Ubuntu or was not servicing our community,

well, either we'd ship the gunny one again or we'd ship something else. Do you know what I mean? Like, we're a distribution event. We're curating what we think is the best for our users. And I think particularly in the case of core utals, I don't see it as a huge moneymaker for anyone in a sense. Like, I think for things like Elasticsearch

is one of these famous cases where the licensing put them in a, complicated position commercially and they've changed a licensing terms and that's that's gone how that's gone. I don't see as a matter of a risk with core utilities personally. Like canonical is certainly not making a move to try and monetize core utilities. We're just shipping something that we think is going to be a bit faster and it's safer in the long term, right? And it's been easier for us to maintain in a long term. And that's really all there is to it. I've heard you talk about building, like building with a 20 year mindset. And that's where this plays in, right? It's maintaining

software the enterprises run for just a bonkers long amount of time. Yeah. I mean, we do 12 years extended support at the moment. There's talk of possibly extending that. And in reality, I want us to be maintaining code that's as easy for us to build, test, find contributors to as we possibly can. Obviously, we want to do that. We want to make that task as easy as possible, not only so we can get the fixes out sooner, but the easier each piece of software is for us to maintain, the more pieces of software we can maintain at the end of the day. Okay, I want to ask you one other thing that sounds pretty neat. I believe this is the release where Enit Rammifest tools has been replaced.

Yes, on the desktop. So in 2510, we're using Draycut. I don't know how to pronounce that. I think it's Drakeut. So we're using that on 2510 for desktop. Work is underway to do that for the server, but there were a few use cases where we hadn't quite ironed things out, so we haven't rolled it out to server just yet.

Work is ongoing. My hope is that we would do that for 2604, but again, not if we can't be very sure. The difficulty there is in cases where people need to load kind of weird drivers early on in server, boot perhaps for like a fancy networking card or whatever it might be. We just haven't quite got there yet. But yes, if you install questing for your desktop, you'll be using TreyCat, not the NITROMFS tools. That seems exciting. Yeah, I'm curious. Just firstly,

what was the motivation, you know, to make the switch? And then secondly, I believe this is just an Ubuntu layer switch. This isn't something Debbie and was also doing, right? Yeah. So the motivation there is, I think, stability of kind of boot. I mean, it's a few less bass scripts in the boot path. also enables us to use the system dnits scripts, which theoretically gives us a slight performance advantage and the regeneration of the kernel image at kind of, if you apt upgrade and you get a new kernel, the regeneration of the kernel image should be slightly faster. But broadly, it was seen as a way of getting more bash out of the boot, essentially. So just a slightly more robust way to boot a system. Yeah, and less badge to maintain, hopefully.

So the next question I have for you, and I imagine this probably really won't be much of a story a couple of days after release, but what's going on around release day with flat packs and 2510? Are there problems there right now? Are those being resolved? Yeah, indeed. There was a problem there. So you're aware that we use snap packaging with Cyanobuntu and the security confinement for snap packages is provided by app armor. But app armor doesn't only work for snaps, right? So app armor is a generic Linux security module in lieu of SE Linux or anything else. And part of an effort that's been going on in canonical for the last kind of year or so is an effort to just ship way more app armor profiles for everything in the archive so that, you know, increasingly as you

install things from the archive, they can only access the things on your system they're supposed to access, right? So if you ship, I don't know, a PDF reader, should that thing be able to automatically get access to your camera? Probably not, right? So that's a maybe a contrived example, but the idea is we will ship more and more app armor profiles to make sure that every piece of software you get from the Ubuntu archive can only read file system paths and devices that it should be able to read. And in this case, I believe the root cause the issue was we shipped an app armor profile for, I want to say, fuse, and we got something wrong. And essentially, that stopped flat pack from functioning correctly. There is a fix underway. If it's not already

complete, it was almost finished on Friday. And so that will get kind of backported back into crusting as soon as it's ready. We have been really impressed with the fact that you stuck to the plan of shipping, really recent kernels, hot kernels. The fresh stuff, we love that. I mean, 617, right in 2510 as it's released. That's great. Yeah, we have a policy with both Gnome and Kernel. The policy is basically to ship it, right?

Like, even if it's in RC, there's a motivation behind this. So our colonel team is big. It's one of the biggest engineering teams at Kernanical. And in the desktop, it's important, right? People want things like HDR support. They want better accessibility. We're a little bit beholden to other people's release cadence, except that our releases are literally named after a month we release it,

and so we can't really delay them. And so the deal that we have done with our kernel team and the agreement we have in the desktop team is basically we will always ship the latest canome or the latest kernel, even if it's in release candidate when we release. And then as soon as the thing is out, we will make sure the fully released version is in. And the reason is, in reality, we have to maintain the thing for 12 years anyway. It may as well be the latest, most modern thing we possibly can. And there's nothing to say just because it's an RC that's going to be a bad kernel. Like in the past, we've picked kernels that have been stable, and they've still been horrifying to maintain, for whatever reason, rather. We picked a bad kernel to choose.

And so the view is basically, well, we may as well just ship the very latest we can, right? It's most likely to have the most fixes, the most hardware support, the most features. And we're going to have to maintain it either way. So why not? Yeah, it just makes so much sense to me. Okay, I want to shift gears and ask you if there's anything you're kind of taking as observable, or lessons from the relative popularity that a new meta distro or distribution called OMA Archie has gotten in the last six months. My rough, you know, napkin math puts them around

30 to 40,000 users. So, you know, put it in perspective with the Ubuntu desktop. But I'm curious if you're noticing anything there, trends of the types of users or what seems to be appealing to people. I think what's interesting about that is precisely what you're saying, It is perhaps because of the people who make it and their background, it is capturing a certain kind of audience, right? Certainly, I'm sure, Ruby developers, Rails developers, but there's a whole bunch of people who have been using Macs for a certain kind of development who are now trying out Linux

because people in their circles are trying out Linux and finding it great. And I would honestly love to make Ubuntu a place where those people could have a great life. I think the mythical new user story is important. I care about new users. I care about inexperienced people using computers. But I actually think for Ubuntu, I do believe that for the desktop, a huge part of our market has to be developers, right? It has to be people who are either developing professionally.

So we see huge growth in our WSL numbers, like monumental growth in WSL, the number of people who are using Ubuntu on Windows to do their development work. Or maybe it's just people messing around with ESP home or whatever it is, like tinkering at home. And so we actually have a whole tool chains team, which has grown a lot in the last year, focused on trying to make the development experience on Ubuntu

as good as we possibly can, whether you're a Rust developer, a Java developer. And I think Amachi and others are kind of proving the point that even if you've been a Mac user for 10 years, at this point, there's probably enough maturity on Linux for those people to have a great experience developing on Linux too. Yeah, it seems to me one of the things that's appealing, although it's not something that hasn't been tried before,

is an opinionated, ready-to-go set up that make starting for developers easy. That's one of the things I see on social media a lot is, oh, I was able to just get going. And we've seen Ubuntu take cracks at this. We've seen a lot of distributions take, you know, one or two commands to set up an entire development environment. So that's not necessarily a new idea,

but yet it's resonating this time. Do you think it's a desktop environment aspect as well, or do you think it's just a matter of Windows 10 is going away, Windows 11 isn't satisfying people or the requirements are too high, and the Mac desktop seems to be ignoring developers. Do you think it's just the proprietary platforms are blowing it right now? Or is there something that Omar Archie has nailed in the presentation

that maybe the other distributions have missed, even though they've taken a stab at an opinionated out-of-the-box setup? I think it's both. My experience with MacOS is it maybe hasn't got the attention it used to get from Apple. I was a Mac OS user for years, and I remember very proudly telling people that it was the best computer I'd ever had, and I would be very surprised if someone could convince me otherwise. And my feeling of that over time, this is many years ago now,

but was that it started to go a bit stale. And I think we see that a lot with the focus on iOS and iPad OS. I don't think there's the same level of investment there personally. I also think what's interesting about Omachi to me is that they chose something like Hyperland or they chose Arch. And I wonder how many developers are experienced technical people who have always been curious about Arch, but have heard how hard it is to install.

They're curious about Hyperland because they see these amazing screenshots on the internet. And this is just a really, really nice way to get going, right? Even just in a VM. Like, you install it, oh, wow, that was really easy. I can now play with it right away. So it's just made a whole collection of quite niche, ultimately, technologies, really accessible. And they are interesting, right? They're exciting things to play with.

So I think it's a combination of factors. I wonder if it's not the trend, but an indication of a trend that we're going to see. Like, this could be the leading edge. ultimately I think if you're going to maintain an arch box it's probably good that you understand how that arch box is set up long term I wonder if this isn't a leading edge and if you aren't getting that puck skating

if you'll excuse this really horrible metaphor if you aren't beginning to get that puck skating for the Ubuntu desktop where this trend might be going we may be just seeing the very beginning of this and is this kind of idea been in the back of your head as you've been pushing to make the interim releases more interesting and engage more with the community. Is that sort of the strategy here? Yeah. What I would like is that for people on Ubuntu or people who are on Linux for work

and it happens to be Ubuntu, for example, that when they hear about something new or exciting or useful, it is available to them in an up-to-date and well-maintained form. Now, some of those are going to be snaps, sorry, you know, like the people who, lots of people have opinions about those and that's fine. But SNAPs give us this superpower, basically, to ship software to all manner of different Ubuntu releases over a really long period of time and keep backporting features to those releases, right? So we have some workup coming on kind of like portable developer environments that will come out next year. But also our toolchain teams are working on things like they have these things called DevPacks. There's a DevPack for Spring.

So if you're a Spring Boot developer, you can get like the latest Java development kits in the Graal VM compiler. and all these shiny things that Java developers would want by installing a single snap and the environment is complete set up for you to just get on with your work and work in spring. And we're doing the same for Go, we're doing the same for Rust, for dot net. And so I would like for not just the tool chain

to be up to date in the archive, which is kind of necessary for us to build the software we ship, but also for developers to be able to get all of the LSPs and static analyzers and formats that they hear about really, really easily on their Ubuntu machine without having to like curl to bash, add some GPG key to your machine.

That's an odd trust relationship at the end of the day. You see a piece of software, add a PPA to your machine, and then you run the maintainer scripts and that thing as root every time you update the package. Until it's no longer available. Until it's no longer available. And, you know, sure, you're giving root to canonical when you run app. But like I say, if you kind of have to trust canonical

if you're going to run Ubuntu, right, in a sense. And so it would be, in my view, if you're going to have to give that ultimate trust to somebody, you want to limit the number of people you give it to, right? And so we can publish, you know, security maintained kind of verified snaps on the store and we can push things into the archive. But there are some kinds of software that it's much simpler for us to ship to a much higher number of users with something like snaps, for example, or as a container image perhaps.

Yeah, it seems like, I mean, snaps are kind of, at least for the dev stuff you were talking about, snaps are allowing you to bring containerized development and deployment workflows, but in a, I guess, you know, a different flavor than you see in tools like Docker and Podman. Right. John, I could ask you thousands of more questions. We should probably, we should check in again in a few months. Yeah, for sure. Is there anything we didn't touch on that you want to touch on before we wrap up? There is a really nice step towards a better TPM full disk encryption setup. So this is essentially where you want to have an encrypted disk, but you don't want to enter your password twice when you boot up. So it basically allows you to store that disk encryption key in the TPM and assuming that the kernel hasn't been tampered with and the command hasn't been changed and the machine hasn't been messed up. with it should unlock the discourse somatically. We've had that as an option, an experimental

option in our installer for some time, and I'm pushing really hard for us to have that not as experimental for 26.04. It's in better shape than it's ever been. There's been a huge amount of work on it, which you can read a bit about on our discourse or in my blog. But I'd really encourage people to give that a go, particularly on machines with slightly simple partitioning set up. So at the moment, it's kind of single root partition kind of thing. But I'm really proud of the way the team have come together. That is an effort across SnapD and Colonel and desktop, like, it's a huge effort to bring that across the line. And the more feedback we can get on it, the better. So I think that's a really interesting piece of work that is a genuine

win right. Like, you get better security and better usability. Like, we don't often get to have both of those things at the same time. Yeah, I'm curious, is that kind of what was driving it as an effort for just the general experience? Or is this something that you've also seen maybe asks from, you know, people doing enterprise desktop deployments, that kind of thing. Yeah, I mean, firstly, as a user of it myself, I would like this. And secondly, for sure, if you look at things like the macOS file vault or Windows Bitlucker, like these are features that people have come to expect in operating systems in an enterprise environment, and we also play in that space.

So I think it's important for us to provide that. Okay. One last pet question, just because I want to be able to pick your brain on this. What are your thoughts maybe after the LTS? I don't know, but down the road, so no time commitments, but what are your thoughts on integration of a BcashFS DKMS module or giving BcashFS access to users during the install of one of the interim releases sometime in the future. Is that on your radar at all, John? I haven't got it planned, but I'm not against it. I've been following the whole BcashFS story

quite closely. We obviously did some work on ZFS a few years ago, which still exists. So you can still install ZFS on route and things like ZSys which are tooling around it still exist. We are actually working. So Subiquity, which is the kind of installer that sits beneath the graphical installer for desktop and in fact the actual installer you interact with on a server can support all kinds of different partitioning schemes. It's quite limited at the moment, but we're doing some work this cycle specifically to enable a couple of the Ubuntu flavors to have a slightly different partitioning layout. So that would open the door for more complex partitioning schemes, sort of guided partitioning schemes rather than manual partitioning where people wanted that.

As for BcashFS itself, I am not against it. It's not planned work right now. I don't see it as super high priority at the moment, but equally, like, we ship our own kernel, right? So we have always chosen what goes into that kernel and what doesn't, which drivers we ship, which far systems we ship, and this would just be another variable there, right? I don't think it's a particularly controversial decision for us, at least. Right. Famously, for example, we ship ZFS where lots of districts have shied away from that.

And seemingly has worked out great. so far. And many users appreciate the fact that you do. Right. So, John, thank you so much for spending a little time with us. I'd love to do it again, maybe in April, if you're available, and pick your brain again then. Absolutely. Well, thank you so much. And congrats on the release. We'll link to

your post about it, and as well as some of the community discussions, and of course, the downloads. And any other links you want to send our way, we'll throw in the show notes. And come back real soon, sir. And Linuxunplug.com slash membership. Go there to support the show and you get access to the bootleg feed, which is clocking in at one hour and 58 minutes right now as we record lots of extra content discussion, news that didn't fit into the topic of the week. All that stuff is there.

And it's a great way to support the show directly. You just put your support on autopilot. And one of the things that our members made possible, just our community in general has been really supportive over this last year. We didn't talk about it on this trip. But once again, the headsets that you helped us fund. raise for Linux Fest Northwest, I think it was? Or it was for scale, maybe. Because we used them at Linux Fest Northwest.

We use them at scale. We've used them at Texas Linux Fest. Unbelievably, how great they are. I mean, right. Right. It totally has been fantastic. Yeah, absolutely. I was able to use one when I went to Nix Vegas.

Right. I mean, they're just so much more portable than our previous setup. Everything we need fits in one backpack. I think we use them on the van rescue, too, Chris. And we use them in the van? Oh, we did. We did. This is an example of how our funds can be allocated to make the show

better, make the trips possible, and the gear we use. And we try to do it very wisely. We've been, you know, the show turned 12. And we want to be here for another 12 years, so we try to be very careful about how we do this. And your direct support either via a boost or from a membership at Linuxonplug.com slash membership makes all the difference. And thank you for your support. Well, while we were on the road, the boost machine was still pumping. And a lot of you figured we probably needed to get home with some boosts. So we've got a baller. And now it is time for the boost.

Hey, Rich Lobster! Blackhost is our baller this week with 399,998 sets. All right. All right. Thank you very much. Appreciate that, Black. They write, sending you some digital gas and wishing you a safe trip back home. Thank you for the amazing coverage on Texas Linux Fest. P.S. Drew's Notes failing.

Hopefully this one gets accepted. Oh, we'll look into that. We'll definitely look into that. And thank you for the digital gas. It worked just as expected. Thank you. We made it. We are back. Thank you, Blackhost.

I really appreciate that. Outdoor geek comes in with 100,000 cents. Not bad, thank you. Outdoor Geek. Late Texas, Linux Fast Boost. I wasn't able to get Breeze to send a fountain, so I ended up using River instead. I was messing with Breeds, and eventually it told me I needed to enter an on-chain address to get my sats. I think my Breeze wallet had gone dormant because I didn't open the app for months. And, yeah, I think that's pretty much the behavior they do, so you don't lose your sats,

but you do kind of have to go through a recovery process after your channels get closed. I see, they close the channels. Yeah. River is great. I really like River. I think we have a jupiter broadcasting.com slash river affiliate URL. By the way, I think it's one of the best ways to stack sats. In the U.S., it's all they do, and they have a great infrastructure.

And that is a great boost. Thank you, Outdoor Geek. Thanks for going through and trying it and then send it in your report. Appreciate that. Well, we've got two boosts here from Daja for a total of 86,247 sets. Hey, I hoard that with your kind. Well, I'll be dipped. Late techs stats to help with the return gas.

This is also a zip code boost. If you guys see this in time, you're more than welcome to take a pit stop at our office just off of the interstate. I'll be there pretty much all day, and we've got free soda. So he sent in a geo-hash 9x23J and a little matrix connection. So thank you to that. Oh, my gosh. Yes, zip code is a better deal. Thankfully, I have my, yeah, I have my well-tested map from our trip, which we brought along, of course.

and this would be a postal code in Davis County, Utah. Oh, okay. Well, Brent, you never know. You'll be on your way back. Next trip, we're keeping our boost dashboard open. That is the way to go. We should have thought of that. I will note that.

That is an error we will correct. So thank you, Dajah. Yeah, and thank you for the boost, I'm going to. Now, the second boost here is actually a row of ducks with a little message. Anyway, I forgot to mention that I was curious if you guys have ever messed around with getting mesh-tastic. hooked into Home Assistant. I've been doing around with getting sensor data passed over MQTT for a mesh

of buildings with no internet to one that does, that will bridge to the outside world. Super random and probably won't work as nice as I'm hoping, but initial tests seem promising. All right. Keep us posted, I suppose. I've always been curious.

I've not looked at connecting meshtastic to H.A., but very, very interested in doing so. I'd say top, top tier interested. That's such a thing. You know what I mean? Wes and I were literally talking about it on the drive. We had a lot of time to talk about things.

That was one of the things we talked about on the drive. Goose Guy comes in with 44,4,44 sats. Put some macaroni and cheese on there, too. Big old duck. Affleck! Here's a lay boost for your travels to and from Texas Linux Fest. The Texas Tracker is awesome. I'm glad you got to check on my favorite barbecue spot, and Austin.

Enjoy the sats. Hey, we're glad, too. We are very glad we did. And thank you. We appreciate it. the trip, as far as all the accounting, I'm still, of course, going through everything, and we'll have to figure all that up

because also PayPal is still holding some of the funds. So very much appreciate that, Goose Guy. Also note that the Texas tracker will still be up because I'm not home yet. So if you want to follow the crazy journey that a Brent takes on his way home, please do. Probably be for the next, I don't know, at least a week. I have no idea, actually.

Where even is home for you at this point? Yeah, who knows? Not the one boosts in with 20,000 stats. Linux, sir. something sorry i forgot to send this to help with costs for texas hey no no need to apologize we just appreciate it still struggling to work out how to properly set up nixOS need to read more any tips would help well one thing i mean maybe depend on you know are you struggling with the initial install are you struggling just to like really get a config you like it might depend a bit on

exactly where you are in the journey but if you would like some inspiration maybe check out episode 634 um there should be a bunch of links to various user and audience member configs there's a lot to be inspired by there. And if you dig around on the downloads page for NixOS, there is a graphical installer. You can grab, and they will set you up with a base config. Always a good place to start from. Yeah, maybe even in a VM.

Yeah, you're not going to go wrong there, I don't think. Good question, and do keep us posted. And if you have any specific questions, please do feel free. Send those into us. Well, Kiwi Bitcoin Guide boosted in 4567 sets. Oh, my God, this drawer is filled with fruit lobes. I've been listening and following live from New Zealand.

may the force be with you. Ah, thank you, Kiwi. Thanks for the New Zealand checking. That is so cool. Love that. Jordan Bravo is in with a row of ducks, 2,222sats. Plus one to do another episode with the Rate My Config. All right, that's two votes.

Boats are climbing. I don't know if that was a super well-received episode, The Config Confessions. Would like more feedback on that because we enjoyed it and we have a couple of configs in the kitty. It's a fun thing to do. So let us know. And thank you, Jordan.

register your plus one on this matter. Bite Bitten comes in with 2,000 sats. I like Android to be called Google's Android, and the AOSP to be Google's Android Source Project. Gasp, as it doesn't feel open when everything needs to be registered and it makes devs gasp for air. Oof, that's, that's, uh... I like you. You're a hot ticket.

That's good. Fire. I'm going to try to remember. Remember that. That's a good one bite. Thank you. Ed Broughton boosted in 5,000 sets. This is a tasty burger. I really enjoyed the road trip stories, and I'm looking forward to next week's episode. That's this one, where I hope to hear about the return trip.

Thanks for the coverage of the Texas Linux Fest and for the effort the team puts into this show. Well, thank you, Ed. Thank you for the boost. Really appreciate that. Hybrid sarcasm's back with 10,000 sats. He's a good guy. He's a real good guy. No, he's a great guy.

Tough little ship. Little. Okay, so Hybrid's got an app pick for us. It's NC Spot. He's got a Flat Hub linked here. It's a Spotify Tuey. It has a nice N-Curses interface. And as one who likes to avoid electron apps,

if I can find that, it helps. You know, a Tu-E via Flat Hub, or I mean, I'm sorry, flat pack. That's different. And kind of neat. Thank you, Mr. Sarcasm. This is great. That's a bonus pick right there.

Thank you, Hybrid. Appreciate that. It's NC Spot. I'm sure you can find it in all kinds of places. It's a cross-platform and curses Spotify client. I mean, if you're going to use Spotify... That's the way to do it. That's probably the way to do it.

Although friends don't necessarily let friends use Spotify. There is that. But if you're going to, that's probably the way to do it. Thank you, Hybrid. And please, send more boost picks. Appreciate that. Anonymous Boost In with 5,555 cents. This is the way.

First time booster here finally took the dive and set up Albi Hub. Hey, really? Nice. Also, it was fantastic meeting y'all at Texas Linux Fest. Right on. Nicely done. I'd also like to give a special shout out to a standout member of the community,

TechDev. While at the conference, I noticed how deliberately he reached out to everyone, especially those who seemed alone or were looking for a group to join. TechDev, your inclusive spirit is truly rare and deeply appreciated. I will plus one that. I will plus two, that one. Yeah. tech dev I think made it a little extra special

and I was really glad that he could make it to the event and I noticed the same I noticed the same so I'm glad he got a shout out here in this boost and congratulations on getting your own AlbiHub going yeah thank you for taking the deep dev to do it how fun is it right I mean it's a little challenging right but it's such an unlock too while Audubrain boosted in 3,500 sets

thanks for the coverage of the Linux Fest it was fun watching you guys crisscross the content well thanks for following along yeah we do appreciate that But it was fun for us. We used the crap out of that tracker. Let me tell you. We really did.

Begus Maskus comes in, probably, with us, probably not right. With a space balls boost, one, two, three, four, five sats. So the combination is one, two, three, four, five. Hey, it's Brian from Boise. We were just talking about Brian. This is my first real boost. I was the guy who rolled up at the gas station, slid to your bit chat, and then crept up on your car in the dark while it was pouring rain. Thanks for taking a few minutes to chat.

I was a bit nervous, probably because the whole stalker vibe thing. But it's not every day you get to meet one of your heroes. Oh, Wes is pretty great, I will say. No, it's all Brent. Keep up the amazing content. I'll be boosting it in regularly. Awesome. And hopefully my own new Albi build soon.

Wow. Thank you for, I was hoping we would hear from you, Brian. Yes, it was great to get to meet you. Thank you for putting in the effort, especially in the pouring rain. It was a hell of a storm. We did not expect anyone to come out. And we were stopped there at that gas station for a minute. So we must have wondered what was going on because we were like looking at the storm routing,

I mean, trying to figure out. Debating how far could we make it in this? Where do we sleep? The traction was low. I mean, there was just so much standing water, so we didn't want to drive in that all night. But we were totally happy that Brian stopped by. And we had, like, you know, the quick,

hey, how you doing, moment, and thanks for saying hi, kind of thing. So we do hope to hear from you more big and keep on boosting in. Well, Brad, at Team Toronto boost in with 7,500 sats. The traders love the ball. Plus one for more Nick's Confessions. Okay, that's three.

All right. We have a smart audience. Yeah, amen to that. And it's great to hear and see via Git what everyone is doing. Would love to see more around Nix as well. Or even other topics like Home Assistant, Utilities, Scripps, et cetera. Leverage the community for more content. Love it.

Oh, I would definitely love to see some more community contributions around their Home Assistant setups. I'm sure there's some stellar stuff. I am in a bad way with my Home Assistant. Right before I left for the trip, my Zigby Radio just seems to have gone offline. Oh, no. and yeah and so home assistant can't talk to my Zigby radio none of my Zigby devices are working and it's just totally busted and it was like the day I left and I haven't had a chance to look at it because we got right back to the studio and did the show so I haven't even really dug in yet and looked at the logs but man is it a sucky thing because that kind of thing never happens to me on this it's been so solid and it's the built-in home assistant yellow radio too oh if anybody has any tips please let me know I am not looking forward to troubleshooting that one And thank you, Brad. I will take those notes.

Those are good ones. Appreciate that. Our last boost here is from Magnolia Mayhem, 3,210 sets. Well, that's here, good, buddy. Yeah, I think for context, this is a live boost from our live stream earlier in the day during our interview with John. Nice. In our interview, there's a quote here, accused of making Ubuntu more like Nix. Well, Magnolia Mayhem says, I fail to see the problem here.

There could be some nice things to learn. Right, you got to keep Ubuntu, Ubuntu. I do agree there. People depend on that. But there's a lot to learn from all the great projects out there. Indeed. I had so many more questions for John. I mean, so many more questions.

So we'll try to get them back on soon. Thank you everybody who supported the show with a boost or a membership. And thank you everybody who also stream sats as you listen. 24 of you did that, and collectively, you sat streamers stacked 39,000 and 18 sats for the show. When you combine that with our boosters, we stacked a really very, very, very great. 745,839 sets. And we are putting that towards the Texas Linux Fest trip. We did have a little extra travel going to Denver.

And that's from almost 40 unique people out there. Isn't that crazy? That's great. Thank you, everybody. If you'd like to boost the show, Fountain.fm makes it real easy. That's a podcasting 2.0 client that just hosts all that. But you could keep your client, set up an AlbiHub, and use the podcast index or whatever you like. There's a lot of options there.

Check it out at podcast apps.com. Fountain has been getting really, really good, too. They've been making some fantastic, fantastic improvements. And thank you to our members, Linuxunplug.com slash membership, for putting that support on autopilot. We very much appreciate it. Now, our pick this week,

this is only the kind of thing that people that are absolutely exhausted and have been driving across the country would come up with. This would never make it into the show otherwise. I am aghast that we have this in here. Yeah, what's your problem? It's the performance optimizer observation platform. Yeah, poop, for short.

And I love their tagline. Stop flushing your performance down the drain. You see what they did there? I think you'll like this, though. We don't have a sound effect yet, but it's 100% Zing. Really? Yeah, yeah. Okay.

And it uses Linux's built-in Perf Event Open functionality to compare the performance of multiple commands with a, colorful terminal user interface. That's the part I like. So you're just building on some stuff that's existing in Linux to give you some really great comparisons in a very readable way. The name's silly, but the utility is there. Yeah, and folks might be familiar with Hyperfine, which is sort of a more mature project

in this space, which builds itself as a command line benchmarking tool. So that's kind of where this thing is aimed is you can, you know, maybe you've got different versions of something with different tweaks or comparing programs. doing the same kind of work. If you're on Linux, this can do it. But how much Zigg is that one written in? Well, that one's written in a combination of Rust and Python.

You see? Yeah, zero percent Zing. I asked Zieg. That's zero percent Zing. Yeah, right. This is 100 percent Zieg native pick this week. You're welcome, everybody. And it is MIT licensed,

so have added as you will like or whatnot. And we will have poop linked in the show notes. Okay, there you go. See? You could tell. We're a little, we're a little tipsy from the road. We wouldn't have put that in there otherwise. It was probably Brent idea. You're going to just go ahead and blame him on that one. You know what I mean? Just go ahead and

see you next week. Same bad time. Same bad station. All right, we will be live back at our regular time. Of course, we'd love it if you join us over at jbblive.tv on a Sunday at 10 a.m. Pacific, 1 p.m. Eastern. Links to what we talked about are at Linuxunplug.com slash 636. And there's some good ones this week for Cosmic and, of course, the new Ubuntu release. Wes, if they are in a super power user podcasting app, we have some great, I don't know, features for them? Oh, yeah, we do.

Well, we have cloud chapters, so they can jump right to the content they like or listen to the show in whatever order. What else do we have? Well, if you want to take an even deeper dive into the show, check out our transcripts. We got transcripts now. Yeah, plenty of apps,

and some of them even show our speaker diarization so you can tell which of us said the dumb thing. I love it. All right, we'll have a big show for you, so join us right back here next week. I'm going to be able to be. You know,