This week, we're going to respond to some of the toughest criticisms around Nix OS. And a couple of years ago, the Changelog podcast had a guest on that had a few critiques themselves. And I think it sets a theme. I want to play this for you. We'll link to the full clip. This was sent in to us, and we'll start today's show with something I think we can hopefully respond to by the end of the episode.
I like to think about Nix, for example. You know, Nix OS, Nix, like it's a very interesting, very cool idea, right? It was some fatal flaws, unfortunately, but it does so much interesting stuff, and it changes the whole game. And I like to think about how Nix is kind of like 1950s sci-fi authors. I'm rereading Foundation, right? And their vision of the future was adorable, right?
It's like, oh, yeah, I could see how this would work, like spaceships without computers, and you would actually manually align the stars and navigate that way, right? It's kind of cute. But you could totally see how that future could have been a reality. And to me, things like Nix fit into there. If Nix had won, arguably Docker never would have existed, at least in terms of like an image format, right? And so, I don't know, EVPF, Wasm, I think they're fantastic.
Maybe that means you should not buy their stock.
Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris.
My name is Wes.
And my name is Brent.
Well, hello, gentlemen. Coming up on the show today, as I teased, we'll be responding to a recent wave of criticisms around Nix, and we'll try to take on some of the toughest complaints and share our thoughts, and then stick around because the pick segment, it's worth the price of admission alone. The boys were cracking hard this week to get some great picks. And then we'll round out the show with some boosts and much, much more.
Before I go any further, let's say time appropriate greetings to our virtual lug. Hello, MumbleRook.
Hello.
And a big good morning to our friends over at TailScale, TailScale.com slash unplugged. TailScale is the easiest way to connect your devices and services to each other, wherever they are. It is a modern networking solution for connecting your devices securely, and it's powered by Wagon. Yeah, and it's secure, it's fast, and it's really quick to get up and running.
And it makes old legacy VPNs look just, well, old. And if you're in a corporation, you're going to love how it'll plug in with your existing authentication infrastructure. And when you go to our URL, you can try it for free for up to 100 devices and three users. No credit card required. You can really just try it. Thousands of companies like Instacart, Hugging Face, Delingo, I don't know, Jupyter Broadcasting, others use Tailscale.
I pull the boosts over Tailscale every week.
That's right. It's just so good. It's so good. I mean, like all of my phone's private data syncs over Tailscale. I got no inbound ports. I love it. It's privacy for everyone and every organization, and it's easy to use, and it's programmable, too. Check it out. Go to tailscale.com slash unplugged. Okay, we just have a couple of things we got to nag you about. We're not going to do it a bunch, but, you know, episode 600 is just around
the corner. It's February 2nd. Brent, did you see how many meetups we have now?
I can't believe this. Yeah, we have 11 meetups happening a little all over the world.
Whoa. Oh.
So I'm going to list them all for you because I think it's a lot of fun. And if you, as a listener, want to join one of these, like, go for it. You can go to colonyevents.com slash events to see them all listed there. You can even host your own. That platform allows you to make your own. So if you want to invite some listeners in your neighborhood to join your meetup, that's a great place to post it all. And that's where we're keeping everything central.
So we have a meetup in London. There's one in Central Florida as a listening party. There's one in Berkeley. I guess there's one in, what is it, Central MA? What's MA? Or something like that. There's one happening there.
It's pronounced, it's match-a-touch-nuts.
Ah, yes, right, right, right. Sorry. I didn't get that memo. That's happening at the Quest Archery, which sounds really cool. There's also, of course, a listening party in New York. There's one in Toronto with our good folks over there.
Awesome.
Pacific Northwest. Wait, that's ours. We're hosting a little party.
Aren't we? Yep, we are. I don't think we'll be at the studio, but it'll be near the studio.
I just don't know yet. It will be in the studio area.
Yeah, but our idea is we'll be able to wrap up the show, get it out the door, and then we'll just head up and, you know, go party. But that way we can do all the post-show stuff and not, you know, make mistakes like we sometimes do when we throw parties.
We also have listed here a meetup in Walla Walla. That's in Southeastern Washington.
Let's go to that one. You are.
All right. We'll just, we'll do the, if we had a Concord, that's what we need. We could just go to each meetup, you know, spend like 15, 20 minutes with people there and then move on to the next one.
Well, the best we can do is mumble.
It's the best thing now uh i'm not quite done with the list i'll there's one in what what is this one listening party midwest and what's mnwi wichita am i getting that properly who has the canadian guide help me out here all.
Right so you got one in the midwest there's one in michigan and there's one in pennsylvania yeah.
Bloom bloomsburg bloomsburg how about that.
So 11 meetup events in total coming up on February 2nd. And you can also just join us in the Mumble Room from your pad because we're hoping a lot of these will virtually all connect via our Mumble Room. So if you can't make it to one of the events, you can't always still join us in our virtual lug. You just need the Mumble software. We have details at jupiterbroadcasting.com slash Mumble.
So there are a couple of events that we are going to be at. First of all is Planet Nix, which runs March 6th through the 7th at Pasadena, California. That runs along Scale 22X. The Southern California Linux Expo is back, and we're going to be there. That runs March 6th through the 9th at the Pasadena Convention Center. So the nice thing about that is if you're a Nix person, you go to one spot, and you can check off Planet Nix and Scale and hang out with your boys.
And then not too long after that, April 25th through the 27th is LinuxFest Northwest, the 25-year.
That's going to come up fast.
Yeah. So those are all coming up. We just want to make you aware of it. If you like to participate, then we encourage you. And if you don't care, then we are moving right along. Thank you for your patience. So we thought we would talk about a growing chorus online of complaints about Nix. And one of the ones I think caught our attention recently was a post titled Nick's Death by a Thousand Paper Cuts.
And sort of the thrust of the piece is that Nick's probably isn't even really ready for experienced Linux users on the desktop.
Yeah, the TLDR at the top is in its current state, 2025. I don't generally recommend desktop use of Nick's OS, even for seasoned Linux users.
That caught my attention, I think caught all of our attention, because there has been a reoccurring theme in some of the complaints that we've been seeing. And we have a different perspective on some of these things. And we'll go through maybe this post and just respond to a few of them. And Brent, I thought maybe we'd start with the author talking about major version upgrades.
Yeah, the author here generally thinks it's fairly easy, but here's a quote. This does not work for everything. When in Nix OS 23.11, I wanted to try KD6, but it was not so simple. I would have to do some channel foo to make unstable the default instead of just using unstable for kde and stable for other packages as many of us do.
Yeah so this is in the so so section there's the good section which we're kind of skipping over uh for the moment there's the so so section and then there's the bad section which we'll get to and i just wanted to call this out here because i think the authors write major versions in general are a strong point of nix right you uh it's very explicit when you do it you have more safety than a lot of other systems in terms of you know hoping you're about to upgrade doesn't go wrong and all
of that uh but i think this is something we should be careful of in general and just pay attention to as we go along being able to try a new major version of a desktop is i think not something most other common desktop linux distributions really even offer as a feature so i think we just like it's fine to criticize nix and nixos and there's a lot too and if it's going to improve we have to but i think it's worth being
clear are we criticizing something that it's kind of unique or at least semi-unique in doing yeah.
One theme i'm seeing here and i think we'll see this as we go through is nixos does allow you to do new things but they don't necessarily, allow you to do new things easily. What I mean is you're doing something super experimental and you're expecting that to be as easy as everything else in NixOS when that feature isn't even available in other places.
And I think that's one of its strengths. It allows you to do some of this stuff, but it feels to me a little disingenuous to criticize it for allowing you to do these fancy things.
I understand the dream. I'll steel man the argument here. I think the dream is that with Nix and its complexity, what you get with that is this ability to mix and match a lot of things. And to a large extent, this is very true. If you're trying to do this with Nginx or Postgres, probably going to be fine. If you're trying to do this with HTOP, probably going to be fine. Plasma is a big one. It touches everything. And so it touches the display system. It touches SystemD.
It touches the login manager. Obviously, it touches all the desktop applications. And so you're pulling in dependency after dependency after dependency after dependency. So it's kind of like the worst case example. This is a little bit easier with GNOME, and then it gets exponentially easier with the simpler desktops.
And this is just the reality of software in the free software world where you have all these dependencies is you're going to have like this sprile that happens or whatever the word is I'm trying to look for. So I think that's kind of – this is a hard one for anybody to get right. But one of the things that Nix allows you to do is rollbacks. And so I have unstable on my system. I've had unstable on my main system at home since we started experimenting with the real-time kernels.
The whole OS is on unstable, which means I also get super fresh plasma. And if something breaks, I'll just roll back. And I'll just wait for the upstream thing to fix it. And I'll try again in a few days. And if that doesn't work, I'll just roll back. And so it allows me to maybe run a little bit more on the unstable side than i normally would have and so in a way it's just a simpler approach.
Yeah uh that makes me think you could you could solve it multiple ways right there's the version where you kind of integrate it more smoothly into your config in a permanent manner and then there's that you're just like i'm gonna yolo to unstable for a while try it out then i can go right back with very few consequences i.
Have this parachute here if something goes wrong.
It makes me think how would you accomplish this in another distribution I remember pinning packages previously to different versions, but is this something that you can accomplish elsewhere?
Well, I don't know. I just think if you added a PPA for a new major version of Plasma and got it going on your system, that would be messy.
I mean, that's kind of what Neon is. Yeah. At scale.
Yeah, there are specialty things for that.
So if you build a whole bunch of tooling around it, you can do it. You know, I mean, it's totally possible, but you're kind of no longer using Ubuntu anymore, or at least not the release version of Ubuntu. Now, something I think we do see that comes up that I actually kind of can see where people are coming from, especially if they're used to Arch and the AUR, is kind of this variability of package versions. You know, Nix is kind of famous for having this massive package repository,
and the author touches on this as well, Brent. Yep.
For example, Duplicati and Rclone are widely used programs, but their Nix integration could use some love. Both have some options, but if you want them to have declarative configs, it's time to roll your own sleeves and get dirty.
I think this is definitely true. You can have packages that are out of date. You can have packages that, you know, just install and you still have to do all the manual configuration yourself. It's kind of part and parcel of, you know, community support, unfortunately. And a lot of times these extra options for the declarative side is another case of, you know, by and large, most other systems don't have those options at all.
Yes you're not really losing anything you're just not gaining anything in this particular case, that seems like a pretty reasonable compromise on nix because in some cases you will get the option to do declarative configuration and in some rare cases you won't but you can still run those things side by side.
Yeah the author kind of goes on here to say this is why i have a notes file that enumerates all the manual steps i need to take to bring up that system with things like there's a list of you know manual options firefox extensions whatever and i think that is actually a pro-tip in disguise. You know, we often have those kinds of things, notes, commented bash script, whatever, in other distributions. And it's totally okay to have that with Nix and NixOS too.
You know, like, I think there's this instinct that you have to put it all in the Nix config, and if you can, it's often great to do so. You know, you already get half the way there if you get stuff installed and then you have a documented approach to configure it from there. I mean, it's not ideal. Maybe it won't work for a thousand machines, but if you have two home servers and a workstation, it'll be totally fine.
Well, then you have a really nice declarative base. So, you know, OK, the services are getting started this way. Docker got installed this way. And that is a really lean, mean fighting machine that is awesome to then run containers or VMs on top of. You got this nice reproducible base that has rollbacks and can easily roll forward. And then you've got the isolation of the applications and you could, you know, you could run it just that way.
I think I want to talk about granular package control because that came up in this post. And Brent, there's a quote in here about the Nix solution and how it creates all this additional cruft.
The Nix solution for this is to create an additional reference to the whole Nix packages repo at a certain checksum. Then have the package source that particular checksum the next way is not terrible but package pinning contributes to the cruft that will inevitably accumulate in your configs.
Yeah the context here is like let's say you have a new release of a program and you're like you know it doesn't work for your use case you're waiting for the point one release to follow you like so you just you don't want to upgrade that one thing and yeah in nix often you would import a specific commit or maybe this is where you pull it from stable if you're running unstable on your system i think the you know if it's cruft or not is maybe a matter of perspective and i think there's also
a factor of nix maybe makes it more apparent right like okay.
Well first why do people call it cruft what do they mean is it because there's multiple versions lying around.
I mean it's additional lines of code it's additional you know yeah copies in your disk usage and that's a theme we'll see here is cruft uh messy is another adjective used later on okay in the post. So I think there is this aspect of perceived organization, perhaps, or maybe what looks like a workaround or a hack, depending on how you're viewing it.
Okay. I guess, to me, it seems like that is a, first of all, I don't have to directly interact with that crufter mess. There's tooling around that. And then, second of all, containers often have a lot of cruft, just as an aside. But with that, you get reproducibility. You get some guarantees that you don't otherwise have. That seems like a pretty fair tradeoff, especially on production systems.
Yeah, right. So, in this case, like, okay, I have to go get a commit to the version that has the package that isn't broken or is the version that I need. And then I kind of have to like add that into my flake and I kind of have to thread it through and pass it through and then do the override so that like if I use the next package option, it uses the overwritten right version of the package and it's like a little bit of work in your config. That's definitely true.
But the author says in Pac-Man based distros, I can just roll back a package to an old version in my cache and tell Pac-Man to ignore upgrades on that package. I think that counts as cruft just as much, if not more. And it's going to be hidden in your Pac-Man config and you have to rely on the mutable state of, do you have that in your cache? If you're going to set up another system, is that in Arches cache still?
Yeah, it's inside the package manager. It's not defined anywhere in the system that you can go review and change later without having to interact again with the package manager.
Right.
Well, and I would argue as, you know, the author said, this is not really a desktop that they would recommend for a desktop Linux, even someone who's quite advanced. I think I'm fairly advanced compared to most computer users. And I've never felt the need to like pin an application like this in my everyday use of NixOS. Chris, I would imagine you haven't either.
No, I mean, we have a couple of production apps that we've set once and never really had to bother with again, but that's about it. And I think it works pretty well for that use case.
And do you think there are times where you want to pin if you're all unstable, say? Oh, yeah, sure, I can see that. Or if you have a lot of complicated changes and one package is broken and you don't care about updating that right now.
Yes.
You can also sometimes just uncomment it. Depending on how crucial it is for this particular time.
That's my move.
I mean, oftentimes you have it if it's just convenient because it's in the packages and it doesn't cost too much to have. I do think there's some truth, right? Like, on some systems, if I just wanted to upgrade one package, you can do that without updating the whole world. But again, you may be compromising your reproducibility of that system, and you now have more state. So it kind of comes down to how important is, you know, which version do you prefer?
There's also, and we might talk about this more throughout, but you don't necessarily have to put everything in your system config, right? You can also have per project or per, you know, whatever flakes and stuff that have additional tools that can be pinned to differently.
And that means you don't have to mess up your main config for that stuff.
You also don't have to rebuild your whole config if you're just, you know, you don't, you need it all the time. I don't need it in every place, but I know that when I do specific activities, I just activate this environment.
Right. So, you know, it's not a whole system thing. So that's the so-so system section. Let's move to the bad section. And I don't I don't want to make it sound like we are railing on the author here because like one of the things that they say is, quote, first off, I'm well aware that some of these issues mentioned below are likely exacerbated by my own faults. Like they're they fully acknowledge that they're not perfect.
They're not like, you know, and I think we're not we're kind of using this as a discussion. We're not using this to like particularly go after them. But there is a couple of things in this bad section that I thought we should talk about and respond to.
The author notes here it's possible that the fact that nix is always quote in the way between me and linux it may have soured me.
I've heard this sentiment before it's in the way i just i just want to do x but i have to go through nix then i have to build it.
I mean it is it that is true right i think this is where maybe it can be more difficult for experienced linux admins in particular i'll.
Give you an example right adding a new user account for example you just want to i just want to add a new.
User yeah user add i know the flags like no problem i.
Don't want to have to like go add it to my config and then build a new version of my system with that user.
And you know there's been research in this area and many people have talked about ideas of like could you do that and then have nix capture the stain and we may get to a place where you know tools ai tools who knows like can't reliably do that that may be something but for the moment you're right like you you really have to adapt to like, no, I'm not going to do that. To make this change, I have to go to my config file and put it there and then
rebuild the system and now I'll get my user. And that can seem like handcuffs.
That's, it's, I think what's at the root of that complaint is a workflow change. You know, if you talk about, The difference there, people that maybe are already living like a DevOps workflow where they're making their changes and then they're putting it up on GitHub, right? They're checking it out.
They have version control. And then, you know, they're pulling it down on systems like they are already living in a world where if they want to make that kind of change, they go make that change somewhere else and then push it to the system. So for them, it's not really that much of a transition. But if you're maybe, again, saying you're typically used to Ubuntu, Fedora, or Arch, and you just want to add a user, this feels like it's, quote, in the way, I think.
It reminds me, it's maybe not a perfect analogy, but I think there's at least something because Nix is a functional programming language. And in the development area, I have found that it can be sometimes more difficult for experienced object-oriented programmers to adopt a functional style than it is if you're teaching someone who's a less experienced programmer. You know, there's just less unlearning.
Yeah, there's less to unlearn and there's less bias towards how things should be working or how you expect them to work.
Right. And if you don't have that expectation, maybe you can be a little more free to try several different approaches and land on the one that's a more natural fit for the environment you're in.
I'd like to argue that my personal experience with the Nix config is that the abstraction is actually a superpower for me because I've often struggled with, oh yeah, how do I do that user thing again? And the last time I did that was like a year and a half ago, and I don't quite remember how to do that. And so I have this vast directory of notes of like just quick hints to myself of how to do these like Linux-y commands that I only do once in a while.
And I found that the abstraction in the NixOS config was actually really nice for someone like me because, A, it's there and I can see exactly what's being defined in the system. But also a new user is like a copy paste away and it's super straightforward and i don't, pretend to understand the nix language at all but like a quick copy and paste for something like a new user for me is a super nice uh use case that i find is making my linux life actually much simpler which i really appreciate.
Right if you don't have to unlearn the part where you're like i already know how to do this then just knowing that oh there's this there's gonna be one spot that i know that i will have to do this can kind of be nice and simple.
Brent there was that section in the blog, too, where he kind of tries to warn off even experienced Linux users.
I'm on the fence if Linux is better for experienced or new Linux users. But this point is driven home when I find at Nix conferences, a bunch of high schoolers who package Nix derivations for fun. Perhaps I'm not the best developer, or maybe my experience is what's actually in the way. Also, the fact that I don't like the Nix language may be due to my middle-aged brain being hard to teach new tricks? I don't know.
I think this is another area where, I mean, it may be that there is just more barrier of you expect things to work a certain way. Maybe you're not especially interested, which is totally fine. Nix doesn't need to be for everyone. And it is a very different approach than a lot of mainstream DevOps and even programming tools.
Going back to that changelog quote where their guest said, it's kind of like a 1950s sci-fi novel. The bit where I agree is i think the the switch that i had in my head even though i don't really even understand you know a part of of the nix language like.
I don't think could either of you identify a nix function or you know write a hello world or a function that takes in a number and adds one to it.
I could definitely reproduce one if i saw one yeah but i don't know i could write it from scratch no and i and i have never needed to yet maybe one day i will and i've kind of taken i had this shift when I've, after we tried the next challenge and I wrapped my head around immutability and then I wrapped my head around Nick's the shift that I had was this kind of, Oh.
We should have always been doing it this way. This is how computers should always be because now you're programming the machine through the whole stack. It's no longer kind of this black box where I go in and I adjust various vague settings throughout the entire stack. And then the machine sits there and reproduces what I want for a while. But then I change something like I upgrade. It's fragile and it no longer does the thing. That's not how it works with Nix.
And it should never have gone the way it has. And so once you kind of look at it and go, oh, this is you need to program the machine. It's a machine that needs to be told what to do and it needs to be told what to do in a specific way. And if you do it, you can reproduce that and you can tell all the machines to do it this way. And that's how we should be doing this. And so when something doesn't work for me, I don't think of Nix as doing it wrong.
I think of, oh, we've probably had a bad practice of how we've been packaging, how we've been distributing, how we've been running this thing. And that needs to be addressed. and Nix gives me the tooling and the runway to learn how to actually make it feasible and how to actually do it. And there has been multiple times now where something's not available or it's not the right version. And in the past, we never would have created a deb or an RPM out of it because of the massive complexity.
But now in-house, it's something we can actually do. And it does take a little bit of upfront work, but that works done once and then it pays dividends for years and when i had that kind of realization i stopped thinking of oh it's such a pain in the butt that i have to use you know the config file to add a user account i just want to use user ad.
No i probably should have always been defining them this way and then now when i lift that config and i drop it on my new machine guess what my user accounts there it's using the fish shell it's in all the right groups every single time.
I think you're hitting on you know there's a strong philosophy behind Nix and NixOS. And there are times where that puts it very at odds with traditional approaches and what people are familiar with. And sometimes it's just pure painful, you know, it could just be better outright.
So I do think if you are going to adopt Nix and NixOS, it helps to really vibe with and get and be on board with the philosophy because sometimes that explains why you may want to make certain trade-offs or be doing something and bearing the cost of doing something in a particular way. Because it's serving a larger, maybe abstract goal.
Well, the author continues here, using a Nix system often feels like using a programming language. There are many ways to do the same thing, different philosophies, old ways, new ways, etc.
I mean, certainly true. One aspect is like, I mean, it is a programming language. This is sort of embedded in a larger discourse, I think, in the world of configuration management. You have... Things that are aiming to be declaratively done in more of a configuration language, like the proliferation of YAML files everywhere, like a YAML file for Ansible. Although Ansible's done a lot on top, you know, see the article, Ansible is a lisp, for instance, for all the things you can do with it.
But contrast that with Chef, which is no longer as popular, but for a while was, which used specifically a domain-specific language in Ruby to be a programming language. And Nix is kind of on this side of the spectrum. Personally, that's my preference. I think it does mean you need to manage, you know, don't go too crazy with it. You do kind of want it to not be, you know, way out there. You want some standardization.
But I think we have seen that oftentimes if you go the other route, you end up having to re-implement programming on top of YAML, like look at CloudFormation templates.
Yeah, and you're often kicking out to bash scripts at the end of the day and whatnot to really do some of the fine touches and whatnot.
But I do think it's totally fair, right? Like Flakes is a big thing that's made this all more complicated. There does if you are going to dive into actually writing nix the language it doesn't mean you're going to have to learn all the stuff that goes along with writing a language which is developing the sense of taste and how you like to do it and which way you should do it.
And i think that is a key insight right there is you kind of look at approaching building a nick system like you might approaching maybe a software project at certain high levels and so you use the same tooling and And so a common complaint you'll see, it's in our chat room right now, probably the number one complaint is the next documentation's bad. And then adjacent to that, the next error messages are bad.
And for everyday sysadmins and people that have been using Ubuntu, Fedora, and DNF, and Apt, absolutely true. For a developer, pretty par for the course, the type of messages you get back from a system. And with tooling like Claude and ChatGPT and Llama, you can actually take these error messages and you can derive substantial information from them and quickly get to resolutions that are not actually as commonly available to say something on Fedora or Ubuntu, Arch, et cetera.
Because it's a programming thing. It's a syntax thing often. It's a language thing. And that's what these things excel at. And they can produce configurations that you can then riff off of and tweak to your needs in a way that you just don't really get when you're trying to troubleshoot a problem on a Debian system.
Yeah, this was another area where I think we had some divergence with the author because they kind of write here speaking about, you know, figuring out and debugging. And when I've exhausted all my options, I find myself turning to chat GPT to straighten things out. Sometimes it's helpful, but like most things, it ends up producing the same spaghetti that I was dealing with in the first place.
Garbage in garbage out and like that is definitely true right garbage in garbage out especially with llms i think it probably does like we've probably all experimented and found some ways work well in some ways don't in terms of interacting with nix and llms in particular i found that most of them are a lot worse at generating nix code and even nix configurations they can do like a basic rust or go module package or something like that they can do very basic nix system configs but what
they are good at doing is making changes to them tweaking them finding syntax errors yeah finding syntax areas pointing you at things where you maybe have a typo uh and if you can do the paste your config and your stack trace or you know your trace back they're also good at suggesting areas to look at.
That's just it that is just it and it's been really powerful.
Which i think is fair like nix tracing could be a lot better i'm hopeful that eventually that will improve yeah uh so i'm glad to see that there are some things, but it is a real concern that we should acknowledge.
Wes, I'm quite happily going to disagree with your first point that it's not very good at producing configs because I have found a recent superpower of mine is to use perplexity to help me solve problems to integrate into my config. So, Hey, I want to add this new application and I've never used it before. What's the best way to go about that?
To use an LLM to help me with that scenario instead of pouring through the documentation or people's blog posts has gotten me to answers within a couple minutes and really helped me build out my system in a way that has felt a lot more natural than it has even six months ago when I wasn't using that technique. So I think, actually, there's a lot of power in it. And I found some great success using especially Perplexity to do just that.
And it provides references. So if I want to go read the documentation to get more info, that has been a really nice path for me. So that's my new workflow.
I love it. And it's a great reminder, too, that, you know, especially with these tools, they're constantly changing and they're only going to get better next.
And as Magnolia Mayhem points out in our live chat, if you have the fancy version of chat GPT, you know, where you pay like 20 bucks a month, you can get access to custom GPTs. And there's a couple of good ones in there for NICs that are trained on NICs. So they're a little bit better than the average quality. Well, I want to jump ahead, though, because the author, I mean, it's an extensive piece. And, of course, we'll link to it in the show notes if you want to read it yourself.
But they do talk about something that I think is true and is a current problem for NICs. And that's the CDN. CDN is a huge burden right now and it's massive and we've talked about this but TLDR Wes is this is sort of like what you get with Nix is this massive massive sprawling CDN.
Yeah I mean you know you you end up having to build because you have this functional closure right where you have this sort of root and everything builds off of that and if you change something low down near the root you kind of have to rebuild everything after that so you have just sort of a you know the functional nature, the reproducible nature of it imposes,
More disk usage. And also, to make it so that you don't have to run a Gen2-like system, you know, there is a massive cache and big build servers to handle doing that so that you get closer to like an Arch-type experience in many ways. And that creates a burden for the NixOS Foundation. I do think, you know, we saw last year that one of the reasons for that is at the moment, or at least then, they literally hadn't deleted anything basically since, you know, sometime ages ago.
So it wasn't just the cache for like the current stuff or the last couple of years. It's like it's all the inputs and it's all the outputs. So they were looking at doing some fancy like extended extend the next garbage collection algorithm to work through the cache kind of thing. So I'm hopeful this can be improved. It's also, I think, one of the bigger things on the foundation's agenda in terms of long-term support for the project.
Yeah. And maybe we'll get an update at Planet Nix this year. I would hope to see one because that was sort of one of the big issues identified there. And it's a source of where they need funding, too, I think. You, Wes, found a brilliant clip that I think really encapsulates this, even though it's not about Nix at all.
Uh yeah so this is a talk from rich hickey the creator of closure from i think way back in 2011, uh and rich rich likes to get into word definitions and sort of explore the history of things we're not gonna play all of it but in this talk he's exploring the words simple and easy so.
Let's look at easy, I think this notion of nearness is really, really cool. In particular, obviously, there's many ways in which something can be near. There's sort of the physical notion of being near. Is something right there? And I think that's where the root of the word came from. This is easy to obtain because it's nearby. It's not in the next town. I don't have to take a horse or whatever to go get to it.
We don't have the same notion of physicality necessarily in our software, But we do sort of have our own hard drive or our own tool set or sort of the ability to make things physically near by getting them through things like installers and stuff like that. The second notion of nearness is something being near to our understanding or in our current skill set. And I don't mean in this case near to our understanding meaning a capability. I mean literally near something that we already know.
So the word in this case is about being familiar. I think that collectively we are infatuated with these two notions of easy. We are just so self-involved in these two aspects. It's hurting us tremendously. All we care about is, you know, can I get this instantly and start running it in five seconds? It could be this giant hairball that you got, but all you care is, you know, can you get it? In addition, we're fixated on, oh, I can't read that.
I can't read German. Does that mean German is unreadable? No. I don't know German. So this sort of approach is definitely not helpful. In particular, if you want everything to be familiar, you will never learn anything new because it can't be significantly different from what you already know and not drift away from the familiarity. Okay.
There's a third aspect of being easy that I don't think we think enough about, that's going to become critical to this discussion, which now is being near to our capabilities. And we don't like to talk about this because it makes us uncomfortable because what kind of capabilities are we talking about?
If we're talking about easy in the case of violin playing or piano playing or mountain climbing or something like that, well, you know, I don't personally feel bad if I don't play the violin well because I don't play the violin at all. But the work that we're in is conceptual work. So when we start talking about something being outside of our capability, it really starts trampling on our egos in a big way.
And due to a combination of hubris and insecurity, we never really talk about whether or not something is outside of our capabilities. It ends up that it's not so embarrassing after all. Because we don't have tremendously divergent abilities in that area. The last thing I want to say about easy, and the critical thing to distinguish it from simple, is that easy is relative. Playing the violin and reading German are really hard for me. They're easy for other people, certain other people.
So unlike simple, where we can go and look for interleavings, look for braiding, easy is always going to be easy for whom? Or hard for whom? It's a relative term.
The fact that we throw these things around sort of casually saying, oh, I like to use that technology because it's simple, and when I'm saying simple, I mean easy, and when I'm saying easy, I mean because I already know something that looks very much like that, is how this whole thing degrades, and we can never have an objective discussion about the qualities that matter to us in our software.
1Password.com slash unplugged. That's all lowercase. It's the number 1Password.com slash unplugged. Imagine your company's security is like the quad of a college campus. You have those nice brick paths between the buildings, and those are your company-owned devices. You know what I mean? The IT-approved apps and the managed identities of your employees.
And then there's the past that people actually use. You know, the shortcuts worn through the grass, the actual straight line from point A to point B. Those are like your unmanaged devices or your shadow IT apps or maybe non-employee identities, things like contractors, everyday stuff. The reality is most security tools, they only work on those happy brick paths. But a lot of the security problems take place on the shortcuts.
1Password Extended Access Management is the first security solution that brings all these unmanaged devices, apps, and identities under your control. It ensures that every user's credential is strong and protected. Every device is known and healthy, and every app is visible. 1Password Extended Access Management solves problems traditional IAMs and MDMs just can't touch. It's security for the way we actually work today.
And it's generally available with companies that have Okta, Microsoft Enter, or it's in beta for Google Workspace customers, too. I love 1Password and how they make this so straightforward, because it is very hard to strike the balance between something that makes IT happy and something that makes the end users happy and doesn't pit the two groups against each other. That's what they've done. Plus, you already know 1Password's award-winning password manager.
It's trusted by millions of users, over 150,000 businesses from IBM to Slack. And did you know 1Password has regular third-party audits and the industry's largest bug bounty? Because 1Password exceeds the standards set by various authorities and leaders in the security industry. Secure every app, every device, and every identity, even the unmanaged ones. 1password.com slash unplugged. All lowercase. That's the number 1password.com slash unplugged.
Now, the author here talks about reaching Nix nirvana. Here's a quote. You will often hear the trope that there is a high learning curve to Nix, but then you'll reap major benefits going forward. I find this to be mostly true, particularly if you are just deploying some simple services or just love Nix and want to spend your days in Nixing things. But if that's not your case, there's no real getting ahead here. It's true. You may always be seeing the amazing benefits of Nix.
But you will also constantly be struggling with or mucking with configs. This is where many of us find ourselves, Nix purgatory. You have seen the light and can't imagine going back to the disorganized chaos of the old days, but damn, Nix is painful. I guess ignorance really was bliss. Or, as one user put it, Nix OS is shit. The problem is, all other OSs are even worse.
Yeah, I was really curious to explore this. One, because we've seen other people in our community kind of comment that they resonated with some of this. And then I wanted to just, I think there's something interesting going on because, I mean, I really like Nix, but I was kind of always going to like it because I'm a weird functional programmer guy anyway. But you two are having a great time with Nix. You've deployed it for family members.
We don't come from a programming background.
No, right? Like you don't, you're not writing a bunch of modules and flakes.
And Brent's brother really loves Nix too.
Well, and I can easily explain why if you're interested. And it's because the reproducibility of previous Linux systems, you know, whenever I just mentioned that my brother, you know, I got a new laptop recently. Well, putting Nix on there was like super fast because everything was already set up and he could just like deploy it and it was done.
Where previously it was like a multi-hour process to try to approximate where you had been, and yet it'll take you two weeks to remember all the things you needed to get there.
Yeah, yeah, that's true. I think there's also, I've heard this before, the sentiment is echoed in this quote that Brent read here about how he's always mucking with his configs. And so when I read that line, I went and looked, and the last time I updated my Nix config was like the first week of December of 2024. I don't, I think maybe when I first was getting my system set up, maybe I mucked with my config, you know, every couple of days.
But once the system's set up, I don't really ever touch the config.
No, I go through periods of like, oh, I'm trying to get X to work and then I figure that out and then it kind of just sits there.
Yeah.
I also, you know, and maybe this is another matter of how you want to use your machine. I think you do, I in practice run my NixOS desktops a little different than I did my Arch.
You know art to be kind of you just didn't get in that loop where every time you open a terminal you do a syu yeah uh i do not do that with nyx really at all uh maybe like once a week you know yeah and i think that can help especially if you're on unstable just in terms of random broken packages but also just reducing the amount of you know big old closure that you have to download, so i think there's i don't know i find that there's a promise of nyx and nyx os and it's like
very pure and very declarative and in truth we probably are only 60 70 80 percent of the way there and i find personally that a key is kind of recognizing where you're okay with those trade-offs because you can get a crazy amount of benefit and then if you're just okay with like drawing the line sometimes i mean like i will come back to this i don't care about that this much the author has this example of they were kind of fighting
with the declarative integration of flat pack management with Nix OS, that's one area where that's not as crucial to me, right? Like that's one area where a bash script with all of my Flatpaks in a loop that installs them is probably going to be about as good in a lot of ways. And that would be fine if I was having problems with the thing in there and I wouldn't lose a lot of sleep over it.
That is a great example. There are folks that come up with ways to declaratively install their Flatpaks, but But I just use my Nix config to enable Flatpak, and then I just add the FlatHub repo, and then I just kind of go shopping when I set up a Nix system. And I guess I sort of – I have like a core set of packages that I install from the Nix repo. And then a whole bunch of stuff – and this is just how I've approached all immutable distributions.
A whole bunch of the user space stuff I just do via manual Flatpak installs because, well, I'm feeling like doing Brave this install, or this time I'm feeling like doing Zed, you know, and I just use – So yeah, for me, I think what you're getting at is, You can take it to the exponential level where your individual text editor is getting its extensions and its fonts and everything is set by Nix.
And the further you take that, the more you're going to tinker, the more edge cases you're going to hit and the more frustration you'll hit. But you can back that line up a little bit to Wes's point. You can back that up a little bit and say, you know what? I'll just install the flat packs that I want and then I'll go into Zed and I'll add the line about my font size.
It kind of makes me think of automation. I mean, I think we all are big fans of automation, but we've probably also all been in cases where, like, the desire, the programmer, the nerd, the sysadmin desire is to automate it. But if you think about it from, like, a business context, like, it probably just makes sense to have a Google Calendar reminder that does this, and we pay someone to handle it the two times a year it needs to happen.
Like, there are just some times where you draw pragmatic lines, and it's not necessarily the most pure or beautiful, but it's what, you know, lets you progress to the problems that are actually higher up on your list.
But I think in this context, someone might argue, well, I want to configure something this one-off one time, but I can't find it in Nix because they do things so differently. So what would you say to that argument?
I mean, I think it would really depend on the specifics. There probably are cases where Nix OS makes it harder with particular workflows.
I'll give you an example. When I first started using Image, it was not packaged up for Nix. And so the only practical solution, really, unless I wanted to go off on an adventure, was just to use the Docker container. And then as time came along and it got packaged up in Nix and there's a module and all that, and I really liked what they did with it, I made the decision to transition from a Docker container to the native Nix installation.
And I can kind of go back and forth like that as I like. And so I would say probably 65% of the software I run in my home lab is in a Docker container and the rest is installed and managed by Nix. And I just sort of pick and choose based on, you know, the 30 minutes of Googling and reading that I do before I install the thing.
Right. You can kind of get a sense of like, what options does this NixOS module have? How mature does it look like? Is it going to fit the use case I have in mind? And yeah, you can really productively, I think, mix and match containers and Nix. Heck, you can even, you know, run the database with Docker and run the service with Nix or vice versa. I do think maybe to Brent's point, there are things you can learn.
And this might be different if you're talking like home manager and user stuff versus system stuff. More and more, I have seen Nix OS modules adopting having an extra config parameter where you can just basically write out the raw config for the thing that you want.
So between that and the basic pattern is usually like under var there's going to be the state for each service and a lot of those times you know the, If you're not managing declaratively the config via Nix, you can drop config files imperatively after the fact in there. It is a state you have to keep track of, but that's the same with other systems.
Yeah, it's just like everything else.
Yeah.
Yeah, it's just back to how you were.
But you're right. There are probably, to use escape hatches, you have to learn about and be able to figure out the escape hatches. So that can be difficult.
Before we wrap up, he does touch on kind of where he's going to go from here.
The author says, what I learned from Nix is we should have good reproducible systems by now. We are past having to rely on the golden images or VMs, but this should be doable with standard or mature Linux tooling.
Yeah, I don't know about the last part, but I definitely agree. Reproducible systems should be, you know, something we continue progressing on. It should be a goal. It is achievable and it has a lot of benefits.
The author continues here. I could cobble some things together with Docker, Stowe, or Shemua, Ansible, Flapax, but none of that would light a candle to the promise of Nix.
Yeah.
They continue, for now, I will likely keep NixOS on some home servers. On more active workstations, I may switch to using the Nix Package Manager and Home Manager, or maybe I'll just go back to the old days for a while before taking that red pill. It's going to hurt, but at least I'll get some work done.
You know, it's interesting because what's kind of implied in here is like the sort of, oh, I got to do it the right way. I might not do it the right way. I might just do it my way. I've never really been caught up on that with Nix. It's always felt like a set of tools to just do it the way I want to do it. I don't even consider it being too many options. It's just this is what works for me.
I think it is worth pointing that out. For folks coming from maybe a standard like this is a best practices playbook and there's this one way to do it. But Nix isn't necessarily, I mean, some of it is, right, with the Nix OS services and modules it is because you just kind of, they have a set of options, you can use them or not. But once you go outside of that, you are kind of handed a programming language designed to build packages and operating systems and some recipes to put them together.
But it is on you after that. And that may not be everyone's cup of tea, which is okay.
You know, I wonder if the way to kind of dip your toes in with Nix is to start with a particular project, right? Don't move your whole desktop. Don't move your whole home lab. Set up a NIC system. And if it's a desktop you want to do, just set a goal of trying Plasma on NICs. Or if it's you want to run Plex or Jellyfin...
Maybe just try a Nix box that just is doing that and don't try to bite off, you know, this whole I'm going to do all of my work on this Nix desktop or I'm going to have my entire infrastructure on this Nix server.
We didn't cover it, but the author talks a lot about development environments, which we've talked about on the show, can definitely be one of the trickier environments in Nix. The authors even use a dev end, which is one of the things we recommend for that among a lot of great tools.
But that's like i i use nix on my personal laptop on kde neon for a long time before i switched to use a nix os on it as a daily driver and i didn't even think about trying to use it in like a work context besides maybe adding a flake here or there to help myself with a very you know not a super complicated environment and that's one area too where it's you know you can like with python you can set it up where every single
dependency in your you know that pip installs nix knows about and it does itself. Or you can just make a single virtual environment that Nix only knows about that. And you lose some of the reproducibility, you lose some of the fine grade rebuilding, but it's simpler. Or you can just have Nix manage Python and activate a development shell and then you just install stuff with pip normally.
Figuring out how to choose between those different levels and which ones really matter for you when I think is very useful. It also makes me think, it's kind of like the same advice we give to people, switching from Windows to Linux, right? Like, don't put Linux in the way, if you can avoid it, of causing you problems when you're trying to get work done.
Let it put it in situations where it has a chance to succeed and you're not going to be fighting with it when you are in the right headspace and have the right time to get it done.
And this is where I wanted to punt to the audience for a second and have you boost in your thoughts on this, because kind of from where I'm sitting, it does very much echo my experience early in the industry. And check me on this audience if I'm right or wrong here. When I first got into IT 20 years ago, I was constantly told we shouldn't be deploying Linux on these servers because the way we've always done things is on Windows. And we have these applications that do this specific thing.
And your Linux box, while cheaper, doesn't exactly do it the way we've always done it. And you can accomplish the same end goal, but you don't accomplish it the same way. And I cannot really convey to you in 2025 how much resistance that alone created in the adoption of Linux. And if I would have listened to those people, I would have ended up deploying Windows in all of those systems. And so to me, a lot of what it comes down to is workflow friction with Nix adoption.
And when you bounce off it, if you bounced off it, please boost it and tell me if I'm right. It's because you had an expectation that the system should work one way and it turns out you have to do it a different way like our user ad example or other things like that.
That's kind of my assessment of it and that clip that Wes played really drove it home it's what you're familiar with I'd just like your thoughts on that and anything we've talked about in this segment to kind of check us because we tried to steel man this as much as we could internally but now it's your opportunity to kind of steel man it as well and send us your thoughts.
I think I have some closing words here, too. I came to a new perspective recently, which is that I don't think NixOS is very distro hopper friendly. And I think we're used to being able to move to a new Linux for a weekend and try it out and have these fuzzy feelings of, you know, seeing new things. Just being able to play with it for a bit and understand it quite well.
But I don't think Nix OS falls into that category because you can certainly install it for the weekend, but you have to relearn a whole lot of things. And that just takes time and expecting to have, you know, put the check mark beside your Nix OS, it's going to take a lot more than a weekend. And I don't think you can learn it all in one go. It is a process. And if you go into it, you know, enjoying the journey, then I think you'll have more success.
Annual membership is available jupyter.party and if you sign up you get one month for free what yeah we'll have a link in the show notes for the annual version for the jupyter party need special features for all the shows you support all the shows you don't have to pick a favorite child what you can love all of them i know and of course you can support this show directly to linux unplugged.com.
Slash membership, Well, we got a special boost here. Drew, our dear editor, sent in what is considered a manual boost and directly to editor Drew.
He got his own boost? All right. That's great. Yeah.
A DevFree17 sent Drew 10K sats to say, thanks for all the edits you do for Jupiter Broadcasting.
Aversaries knows. He knows. Drew makes it better.
He really does. It kind of skirted all of our systems because it was straight to the heart.
Yeah. That's great. That's nice to see. And speaking of Adversary 17, they are our baller booster this week. On sale for 38,768 sats. Woo, it's across seven booths. Checking in on pass keys, Bitwarden does sync cross-platform. I use pass key from Firefox extension on my Android phone without issue.
Nice.
That makes me feel better.
Yeah, definitely.
I'm still feeling resistant to the issue, but that does make me feel better. He says, if you ask me, a better solution to pass key is Squirrel, created by Steve Gibson. It uses a single master key that can deterministically create a key pair for each site based upon the domain.
Oh, neat.
That is. That makes so much sense. God, it does. It's such a clever protocol and system, but because it didn't come from a big tech company, it won't see the day of light. Squirrel is kind of like Noster, but way before Noster. I do wonder if Noster will one day kind of provide a little bit of a public-private key pair system that kind of gets credibility by your social media account because it's from your public key.
But Squirrel does sound really cool. That would be great. Regarding the article about Linux and why it isn't ready for the desktop, this was last week's episode. Has the person never looked at their Windows file system? I see DLLs all over the place.
Binary blobs.
Yeah, yeah. He says, sometimes the same runtime is installed from three different applications and it's all the same version. That's a great point. That is, most Linux users download and run software from the web without verifying it's malware free. Well, remove Linux from that statement, and you've just basically summed up 99% of all Windows users, too. Trust me, I work in corporate IT. Yeah. Yeah, I've been there,
too. I could go on, but most issues presented in this article are applicable to Windows, too. Microsoft does not pay people to test. At least, it sure doesn't seem that way.
Oh, ouch.
He says, we in the enterprise pay thousands to even use Windows legally, and we end up being the beta testers each month for these release updates.
Oh, man.
I purposely wait a month before installing the next Windows patch because of that. This article is just an opinion. Yeah, great breakdown. Yeah, last week's episode, we tackled the Linux is not ready for the desktop, and then this week we decided to focus on Nix. And that's this recent round of criticism and just bang these things out of the way for 2025. You know what I mean? Just clear out the decks.
Fud busting.
Yep, right off the top.
Ayo hell, Abusin, with 38,000 cents. Well, no message here, so just the sats.
Oh, appreciate the value. Thank you, sir.
But Zachely's boosting with 29,036 sats. Hey, JB, I got a zip code boost for you.
Oh!
Nice.
Oh, did you bring the...
I did. It might be frozen, though. It's chilly here. Now, since text editors are a common topic, I'd like to shout out Notepad Next, which I just discovered and is cross-platform. It's a reimplementation of Notepad++ from the Windows world. I love Notepad++, and I've been looking for a Linux replacement.
I have always been a little Notepad++ envious, I will admit. So that's good to know.
Quality editor, that is for sure. Okay, let me return to my map.
Now, I have a rough guess, but I could be wrong. Is it a postcode? I'm curious what your map says.
Yeah, I'm looking somewhere in South Carolina.
Me too. Okay.
What?
Come on, guys.
Yeah, Lexington County.
Okay.
Probably maybe some in Richland and Newbury.
Chapin.
Chapin?
Yeah.
Chapin.
Chapin? I don't know.
Did you say Japan?
Tell us. What do you think? What do you think, Brent? C-H-A-P-I-N. South Carolina.
I do love it when you include the town. You know? Please include town if you can. That is a lot of fun for us.
Well, Forward Humor sent in a boost across three boosts, actually, for a total of 9,444 sats. First one here, Rove Duck, for a little fountain feedback. Hi, friends. I noticed the live stream in Fountain allows me to boost live, but does not stream sats live. Is that pretty normal? Thanks again for the great show and happy new year.
I'll ask them. I think that technically should be working, but I will ask them. There's probably a good chance they've already spotted this if you boosted us, but I'll double check.
The next boost here is a Docker question. His current system builds with ZFS storage. Do you all use the overlay 2 driver or that ZFS driver? Or do you use a different file system for containers? Any tips?
ZFS can make a fantastic file system for containers. So I just want to acknowledge that up front. Because we're not anti-ZFS. But that said, I do not use it for my containers.
Yeah, and if we're talking the storage driver, then I think most of the time if you're doing ZFS, folks will at least be having data sets that they mount into containers often. But you can also have it where Docker itself is provisioning things via hooks into ZFS, which can do stuff for volumes, but it can also then do stuff for storage of the container images itself.
So it may kind of depend too on what pools and data sets you have available in the system and how much you care about that stuff and if you need and want the tools that ZFS offers or if you prefer to keep that just on the base.
And how good you are about cleaning up because when I use that, I kind of made a mess.
Did you say cruft?
Yeah, that definitely ended up with some cruft.
I think it's probably better now, but in the past, this was a while ago, there was sometimes some limitations where Docker would be better at fully cleaning up itself if it's using overlay 2 than if it had an issue with permissions or something on the ZFS side, then maybe you'd have to go manually clean up some of the ZFS data sets for, it's probably better now, but.
And he says in here too, like he's thinking specifically, you know, image storage and then the rollback capabilities or snapshot capabilities there with ZFS.
But yeah, I mean, if you're just talking about the actual like application data, ZFS or some kind of copy on, right?
Yeah. And let us know what you end up with because file storage for image is top of mind over here. User 47.9 comes in with a row of ducks. This is my vote to keep the tuxies. I think it gives us a good read on the direction of things and the way they are going.
I'll agree with that.
All right. I'm keeping track every episode. I'm putting another mark down.
I can see him looking for a pen, folks. It's true.
Yeah, but you've just resorted to that giant knife and you're just scratching into the wall? What's going on over there?
Well, we're out of pens.
Next is blood. And if I cut myself, I won't be able to focus. So there's only one other person in the studio with me right now.
Oh, well, good. While you're doing that, the immunologist boosted in with 3,333 sets. To add one more, Tuxies plus one.
Oh, okay.
I discovered quite some useful things to look into.
Nice.
Witcher 1, 2, 3 comes in with 3,930 sats. Listener since the mid 400s and a first time booster. Hey, congratulations. Thank you for taking the time. He says, I would like to say that A, I would suggest one episode dedicated to gaming. We have not done a gaming episode in a really long time. I predict at least four new Linux handhelds from major manufacturers in 2025. Well, Lenovo just announced one, so.
You're getting close.
You're already going closer. He says, also, if you multiply this boost by 10, you'll get a zip code from the central European country from which where I come. Oh, Wes, guess what?
Oh, we got a deal this time. Okay, so let's see. We got to do math first, everybody. We got 3930. You got to add a zero by multiplying by 10, right? So 39300. Looks like it's a valid postcode in Hungary.
That's awesome.
Yeah, let's see if we...
I'm always hungry.
That's why we keep you locked away in the freezer up there thank.
You witcher and thank you for taking the time to get the boost going and all of that.
Oh maybe it's turkey okay i'm getting mixed results from this map here oh i gotta do a little more i'll come back to you let me double check my results.
Well nord boosted in 15 000 sats, i'll be only found a route for my previous boost to brent well thank you anyways let's try again then. I hope Linus from LTT will make Brent's YouTuber prediction come true.
Hmm. Good one. I'd love to see it. And good job routing, Brent. Good job. Proud of you. The Lightning Network apparently did not shine upon Brent or Wes and I. However, this came from my node. So it must have at least worked on my node. Because we see right in there that it came from my node.
The Moose Network seems It seems to be working perfectly fine, so I don't see what to promise.
Yeah, it always is. It always is. That's true. That's very true. All right. Well, True Grits is here with a Jar Jar boost. 5,000 sats. I agree with keeping the tuxies. Wow. All right. Okay.
Strong signal.
Even if you don't make it a multi-network ordeal, at least it could be mentioned on the other network podcast to get the voter count up. I remember in previous years I heard it mentioned on the Ask Noah show. Also, maybe split and set aside for giving the winning projects or something. Ooh, that's interesting. So let me back up a little bit. What True Grits is talking about is there's a scenario that we're kind of down to clown about, which would be kind of a cross-podcast version of the Tuxies.
We don't have any details or any plans or any specifics yet, but maybe something that goes a bit beyond just Linux Unplugs, which reflects multiple audiences. And they would also be involved in helping, you know, come up with topics and some of that and all that. It's a lot we'd have to work out. So it's all kind of in the, you know, be nice to do stage right now. But we've got some time to figure it out.
Okay, update. I'm now going, my now new guess, final guess is Spain. Let's go 39300 is a city located in, let's see here, Torre la Vega.
Hello, Torre la Vega, Spain. That's exciting.
Tell me if I got anywhere close, please.
Oh, I think Freak here has a prediction.
Oh, let's find out. Freak KVH boosts in with 8,472 sats. Ubuntu Declarative Donkey. That's a great name. They got to do it with a name like that.
Right?
I mean, that's a perfect episode name, really. Go on.
God. Well, and I just feel like, wouldn't Ubuntu just blow everybody's socks off if they came out with a declarative version of Ubuntu and they called it Declarative Donkey or something like that? Everybody would love that.
So good.
Well, Chris, you've got some sneaky inside info that the folks over there do love declarative systems like Nixos.
That's true. That's true. Gene Bean comes in with 9,912 sats. Question for the audience. Anyone want to gather for LUP600 and a watch party in, how do you suppose you say that?
Ghent.
Ghent, Belgium? It's the night before configuration management camp starts, and I'll be there for the conference. If so, hit up Gene Bean in the Matrix.
Oh, cool.
Awesome.
Gene the Traveler.
Really? I tried the Zen kernel on Nix 2411 with Plasma 6, but I had to revert back due to the same lockups that Chris had when he went full RT. Oh, interesting. So you went even kind of semi-RT and he still had problems.
Huh.
I have not.
No, the Zed's been rock solid for me too.
Not a single lockup back to totally reliable. So I'm sorry to hear that, Gene. He says, I'd love to learn more about the Mesh-tastic gate monitoring setup. If the booster is willing to write maybe something up as long as a blog post, I'd happily send them a boost and thanks. And also, please do add transcripts. Notes.jupiterbroadcasting.com would be much more useful. Stay tuned on that, Gene. Stay tuned.
This MeshTastic gate monitoring, I don't know. I can't remember which booster sent that in, but if you wanted to follow up with Gene, I'd love to know too. So make it public because gates, presence monitoring, like motion sensors and soil monitors, I would love to get on MeshTastic. I think that could be really awesome. I got a node at the top of a shed, so the whole yards would be covered, I would think. so it's gonna be really way to go.
Exception comes in with a rodox using linux on my laptop since 2006 yeah buddy external displays with lid closed has always been a problem can confirm just got a new tell with a bunch of 2204 pre-built in it was fine initially but started to act out after a recent update we linux people need to get that right man.
Nothing stinks more than when the laptop's working and then the updates come down and there's like a regression of the video driver or something like that or i wonder if he got updated to wayland or you know huh if you figure it out let us know maybe others have run into that exception.
Yeah it is especially rough if you get a new rig that has an outdated lts and you just go to the new lts that's already a big change right out of the gate yes.
Yes yeah if it comes with an image it's already a little old too you're just primed for it.
Well we have a report of extra problems here by tebby doug with a space balls boost 1, 2, 3, 4, 5 Satoshis. I recently started experiencing network issues in my Docker stack. Being that it has been around since the beginning of my home lab about five years ago, it was a giant mess. So I decided to learn something new, Kubernetes to the rescue, and holy smokes am I the only one that found it extremely difficult to learn.
I still have not quite figured out persistent volumes and persistent volume claims, But I did manage to get Olive Tin running in high availability with two replicas and a load balancer. Thank you for the show.
All right.
That is no small achievement. Nicely done.
Tabby, I wonder if you agree with me. It's like part of it's just the language is so thick. So you're reading like a half of a web page before you even really start to get to the nuts and bolts of what you're trying to figure out.
Well, and it's also, right, a lot of these systems are built for scale and also to work for our diverse use cases at different companies and stuff. So there's a lot of abstractions that you have to be able to figure out and understand the full abstraction before you can actually do the concrete case for whatever one you particularly have. OliveTen gives safe and simple access to predefined shell commands from a web interface. Neat.
Super neat. Yeah. All right.
Little command runner.
I like olives.
Yeah. I'll eat an olive right now. Thanks, Tebby. Nice to hear from you. Sohang is here with a row of ducks. That's 2,222 sats. And he writes, regarding Linux is only as good as its web client, have these people ever heard of our lord and savior, GNU Emacs? Good point.
Yep, fair.
Got him. Got him right there.
Pod bun follows on with another row of ducks. we could have an episode for linux based roasts or roasts of linux gather everyone's roasts and have you guys say them or have people join the mumble room and say their own roasts.
It could be cathartic to kind of get it out.
Right probably could yeah that's so nice all your gripes the things that you know the thorns in your side is.
That every week though really.
That's every.
Show that we do yeah that's.
A great that's a great suggestion though thank you.
A rotted mood did Send in 9001 Satoshis. All that just to say, getting caught up in old shows.
Ah, nice to hear from you. Oh, he's at 584, Captain Mesh-tastic and the Solar Cowboy. Well, welcome to the Mesh-tastic era of the show. Glad you're getting caught up, Rotted Mood. It's good to hear from you.
Solar Cowboy is PJ.
Right? So good. Yeah, that's PJ. Pink Snake's here with a Jar Jar boost. That's 5,000 sats. Testing out Breeze, which is B-R-E-E-Z, a great way to boost the show without having to switch podcast apps. And it is a lightning note in your pocket. Thank you, Pink Snake. Appreciate it.
Your test was successful. Autobrain comes in with 12,345 sats. Thoughts on the tuxies. Instead of doing a people's choice, maybe survey Linux podcasters and reporters. Maybe this would be less work and an opportunity for cross-promotion. And then here's some topic ideas for shows. Maybe Linux workflows for graphics, illustration, and presentation. And maybe a show related to writing tools. or a show on educational tools. Thanks for the great content.
You know, I'm kind of vibing with some of what Otter Brain's putting here as far as, like, you know, working with some other folks and stuff like that.
I have some deep value on this one.
Thanks, Otter Brain.
Otter's got brain.
Yeah. Oh, that's it.
That was our last boost.
All right. Well, thank you, everybody, for boosting. We do have the 2000s at cutoff for time. We keep all of them, though, and we read them, and we actually save them in our doc for posterity.
Yeah, we've got some test boosts. We've got some folks working out, moving to Albie. We appreciate all you're doing.
Also, we see you streamers out there that stream them sats to us. We appreciate you. We had 57 of you participate in the sats streaming, and collectively, you helped the show stack 118,027 sats. So when we bring it all together, episode 598 managed to stack 329,601 sats. Not bad at all. Now, you want to get in on the fun. There's a few ways to do it, including you can get Strike to get access to sats or there's so many different ways, really.
Bitcoin Well is one that I've been recommending, and we have an affiliate there at Bitcoin Well slash Jupiter. And the nice thing that they do is once you get all your account stuff hooked up, you buy and it sends directly to your Bitcoin wallet.
Why are they all like that?
Isn't that great? So you don't even keep the sats on their infrastructure. So that's another great way Strike and Bitcoin Well are available in the U.S. And Canada. and strikes available in 110 countries. And then you just need an app to send them, something like Breeze, B-R-E-E-Z, or Fountain.fm. And that's amongst many options, which you can find at podcastapps.com.
And then if you just want to put your support on Autopilot and know the show's going to be taken care of, it's linuxunplugged.com slash membership. Thank you to all our members as well. We appreciate you. Yeah, we have a smattering of picks. Could have been the whole episode, really. But let's start with this really handy tool that you found, Brent, and I guess it's just called ISD, which helps you simplify system demanagement.
Yeah, I think I'm going with iced. I like that ice iced baby, you know, that kind of feeling.
Yeah.
And this really seemed attractive to me because I'm not that well versed in system demanagement. And this is a little Tui that offers some fuzzy searching for units, some auto-refreshing previews, smart pseudo handling, and all sorts of customizable interface and power user stuff. And I thought, geez, I've been sort of leaning heavily towards things like lazy git and lazy docker and btop these days just because I'm lazy.
And this fits right in there to allow me to do some stuff that otherwise I just wouldn't.
Yeah, this looks nice. well, this is a Nix episode, so one of the things that came to my mind is if you're using the setup that converts, compose files into Podman and SystemD units that all work together, with the fuzzy searching and stuff, they all have common unit names, this could be a great way to manage those.
And it just looks good.
It does look good.
It's a terminal user interface that looks good. And when you nail that, you got me. So, nice find, Brent.
And it even has a flake in there, although it's not yet in Nix packages proper.
Ooh, well, a little bird told me last night that he's working on a flake to get this going. And that's our dear Alex. So maybe we'll get that this week.
Oh, cool. Well, let us know. Okay. So the next two, kind of an interesting story. I think, did we both find Planify this week and begin using it for our task management?
When you say both, who do you mean? Because I was certainly looking at it independently.
Yeah, you. Yeah.
Oh, that's me. Yeah. You know, Chris, you and I are in this to-do management journey for the last little while. I think maybe we both adopted it as a theme.
And I was looking for something nice and simple that would play really nicely with the Linux desktop and i found planify and it's has to do with support next cloud syncing support it's designed for gnome but it seems to work wonderfully on plasma as well and i'm curious uh if either of you gave this a shot because i spent a little bit of time with it.
Yeah so i've been really looking you know my my holy grail is oh i'm out and about and i remember something and i add a reminder on my phone and then the next day or later that afternoon or whatever it might be i'm down at my computer. I'm sitting at my computer and I just want to pull up the list of things I need to do. I don't want to go to my phone. I want it on my computer. And then I want to check them off on my computer. Planify nails that because we use a mix of Todoist and Nextcloud.
So it's really nice that Planify supports both of those things. And it's a good looking application at the same time. And it has, you know, support for all the stuff you might expect, like reoccurring due dates and reminders. So it works really well for me. So that's Planify.
Yeah, I did give it a shot. I I will say I ended up off of it after a couple of days.
Oh, yeah.
Maybe that wasn't your experience. And the reason is, I think, a personal one. It was a little too simple for me, it turns out. I thought I was looking for something simple. And there were a couple of things that just got in the way for me. I was looking for something that would allow me to... Just do things really quickly without, um, the interface getting in the way and things like that.
But, uh, I found actually a couple bugs, you know, how I do that and it's, but, um, uh, but also just, it turned out that some of the design decisions didn't work for me. So your mileage may vary. It is really beautiful. I agree with all of that. And I think it would really work for someone. Chris, I thought of you instantly when I saw this and I'm glad you're giving it a try.
That said it did lead me to going back to a productivity system that well i talked about quite a few episodes ago linux unplugged 553 which is called a portably predictable productivity if you will and i used this for like six months every single day and i just stopped for some reason i think i was doing a lot of travel and that broke up my routines but planify allowed me to kind of get back into the motions and realize, hey,
that tool I was using, Super Productivity, actually does everything I need it to. So I jumped right back into Super Productivity. But you found some news this week, I think, Chris.
Well, it sounds like, you know, they've been working on the Linux version and getting things in top shape. And so the developer reached out to us and just said, hey, you know, I've been polishing up features and the Linux version has just really gotten into a great place. And if you guys want to take a look at it again and mention on the show, I really love it. And we did. It's MIT licensed, too, so it's free, and the developer says it will always be open source.
It's got every integration you could think of, Jira, GitHub, GitLab, GitT, CalDAV integration, WebDAV integration, Dropbox integration. I mean, you get my point, right?
Ooh.
Yeah.
So this, to me, does feel like a power user's productivity tool. And you could almost have a system in here, too, because there's time tracking. You have time boxing.
Oh, that is nice.
Yeah. Yeah. So I'm curious.
Especially if you're doing a small business or something.
What Planify didn't do that this is doing for you, Brent?
Well, one thing specifically was keyboard input of tasks. I realized using Planify, which is a little mouse heavy, that... I actually often do want to use my keyboard to do most of the input. And so Planify is a little more click around than enter things. And a popular use case is to allow you to define many, many, many things like to do is like this, I believe. Define tags and projects and stuff all in the text line. And super productivity allows you to do that. And I realized going to Planify
that I was really missing that. And I didn't realize that that was such an important feature for me when I was looking at Planify. But it turns out you learn things about yourself when you try software.
Especially to-do apps. It's something about to-do apps.
It's such a personal thing.
Yeah, maybe there's another great one out there that works on Linux desktop and syncs to mobile.
I mean, is this also why to-do MVC is the classic front-end demo for frameworks?
I always hear about to-do.txt. I always hear that, too. Like, oh, on Linux, you've got to use to-do.txt. I don't know. If somebody's got a to-do workflow that works really well and across machines, boost in and tell us about it. I'd love to – I've just started using Planify. I'm not married to it. I'd love to get something really great in place.
I bet there's still some good Task Warrior users out there with their own servers going.
Oh, I bet.
That was me for a while.
I daydream of an e-ink display by the door here so when I'm leaving, I can have reminders and to-dos right there.
Ooh, like that.
I just daydream of that.
Your birthday is coming.
Dude, next episode, the birthday.
Yikes.
It's on, it's an actual, it's one of the weeks where- Oh, that's right. The birthday's gonna be in the episode, so.
Oh my goodness.
That's how I'm still inviting my birthday is I'm doing a show.
Bring your party hats, everyone. Woo!
It's gonna be our birthday! Which means absolutely nothing.
The real question in my mind, is it gonna be a different bat channel?
Right. So we've been experimenting with a little bit earlier in the morning. We're doing 10 a.m., Pacific, which is what, 1 p.m. Eastern? Am I doing that math right?
Yeah, I believe you are.
And I think that's 6 p.m. UTC. I could be wrong. Always double-check Chris's math at jupiterbroadcasting.com slash calendar. Do you want to try it again next week?
Let's do it.
Okay. Brent, you agree we try the earlier live time next week?
Yeah, I do agree.
Okay. So if you're a podcasting 2-0-0 listener, we'll just mark it pending earlier in your feed and you don't have to do anything. But you can also go to jupiterbroadcasting.com slash calendar and always get those live days, hopefully. Hopefully the calendar updates. Sometimes my calendar luck is just really bad. But that'll be streamed at jblive.tv as always. So perhaps it's a new bat time and a new bat channel.
Remember, we want your feedback. Help us steal, man, the other side of the Knicks argument. And if you think maybe it really does come down to just workflow expectations, I'd love to hear your thoughts. Boost in with that and anything else you'd like to talk about and respond to, even from several episodes ago. We always like it. Don't forget, links to what we talked about, you can find those on our website, linuxunplugged.com slash 598. Oh my goodness.
Getting close.
Shout out to our Mumble Room details at jupiterbroadcasting.com slash mumble. Thank you so much for joining us. And we'll see you right back here next Tuesday as in Sunday.