623: 50 Days of Blue

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:11):
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. Well, today I'll be doing my 50-day Bluefin check-in,
but first, we're going to have ourselves a little Linux user existential crisis.
I'll tell you about that. Then we're going to round out the show with some great
boosts and a blowout worth the price of admission alone pick that you're definitely

(00:37):
not going to want to miss and a heck of a lot more.
so before we dig into all of that let's say time appropriate greetings to our
virtual lug hello mumble room hello hey chris and hello brent hello hello everybody
joining us thank you for joining us and shout out to all of you up there in
quiet listening and of course the live chat room as well stay tuned because

(00:58):
we have a heck of a show coming up,
but before we get there i want to mention that the nix vegas con has extended
their call for papers which I believe you can still jump in on and our own Wes Payne has.
Yeah, that's right. So hey, come to Knicks Vegas, co-located, I guess, at DEFCON 33.

(01:20):
Yeah, that's right. More Knicks at DEFCON. That seems like a very good idea.
And yeah, you have to tell the, I think the end of this month to get your paper submission in.
July 31st, they have a sessionized link that we have linked to in the show notes that you can click on.
It's an easy process. So if you're just thinking about it, why not throw your hat in the ring?
And then DEF CON 33, August 7th through the 10th. I mean, people might be just

(01:40):
coming to DEF CON. They may want to come say hi to Wes Payne.
You should think about setting up a meetup or something.
Oh, that's a good idea.
No, I'm a little jelly. Ah, sounds fun.
Also- We got a scheme to how to get you to come with me.
Maybe. I don't know how we pull that off.
Got a fan. I could pick you up along the way.
Too bad it doesn't run on air. We also will be at Texas Linux Fest in October.

(02:02):
So check them out. Maybe get a taco in there and get your plans to come see us in Austin.
Yeah, that's right. You still got time for their call for papers as well.
Yeah. So there is a bit still of time in two different events,
one in the near future and one off in the distant future where we'd be happy to see you.
So this week, we want to talk about a really big macro trend that's happening

(02:25):
in the desktop Linux space.
And then eventually, we'll get into my experience with Bluefin.
But I want to take a moment and make sure that we all kind of appreciate a shift
that is happening in the Linux desktop landscape.
Like, just put it all out there.
Having now used Bluefin for 50 days, watching projects like Universal Blue,
which is like the source images behind Bazite and Bluefin and Aurora,

(02:49):
What we are seeing demonstrate out over there is containerized workflows to
build a Linux desktop experience that doesn't require forking the project.
You're building on top of images. You're building on top of layers.
Instead of, say, in the past where maybe Ubuntu would be based on Debian,

(03:09):
but it would be a fork of Debian with its own repositories and all of that.
we're seeing a shift in the way that the desktop is being built and it doesn't require forking.
And maybe there's a bit of a pulling apart or um you
know different organizations and people filling different roles
like what we think of as a distribution yes there was the stuff where you know

(03:29):
you curated a desktop environment and you kind of pick the apps that went together
and you put it all into a nice experience but you were also doing a bunch of
actual packaging and building you know the package that runs systemd and make
sure that bash exists so that you have a shell to run and like all those things.
And so I think it's interesting too to see, you know, not everyone necessarily reinventing that.
It's not a new packaging format all the time, but you do end up with a different

(03:52):
curated experience often.
Right. So to get say Bluefin, they don't necessarily need to have their own
set of package maintainers.
At least not to the same degree as like trying to, you know,
bootstrap an arch when you started with Fedora.
I mean, the maintenance overhead just right there is exponentially different.
That's huge, right just you know you don't need maintainers that are that are

(04:12):
necessarily packaging everything maybe you just have to take care of a few of the edge cases.
And i mean it's not like we haven't seen that i think it's a it's a matter of
degree and tooling especially right like you've seen ubuntu has relied a lot
on debian to have a lot of packages but then they rebuild a lot of their own
stuff and then there's some downstream distributions that pull directly from
their parent distribution with their own additional archive and some that do a bunch of rebuilds,

(04:35):
but now you've got you know very explicit technologies like with containers
that you can really just say from whatever version I like that is almost good
enough for me and then add on whatever you want.
Yeah before the show started I ran a command I bet it's still in my command
history right it should be,
I ran a really simple command that was just sudo boot C switch.

(04:58):
And then I gave it the URL to Aurora. And I said, use colon the latest image and hit enter.
And I went from having a GNOME-based Bluefin system to now a Plasma Aurora system.
And of course, this is something we've talked about being able to do in the NixOS ecosystem.
And regardless of that comparison, it's just one of those things like when you
think about doing it on a traditional distro, you just think,

(05:20):
I'd rather not. I'll install a new partition, a new system.
I don't know if I want to co-locate both of these in anything but a demo.
I mean, the whole entire process we did live in the bootleg,
I mean, it was what, five minutes?
Yeah, and that included downloading four gigs of files.
Yeah, so it's interesting. The Bluefin folks, they like to use dinosaurs to

(05:43):
remind us of the older model.
Like, you know, sometimes you've got to bury your dinosaurs.
I know there's imagery there. There's a message there.
Because I think they view this as like a post-distro model.
They don't like to call any of the UBlue-based stuff, Universal Blue stuff, a distro.
Yeah, they build images.
And is Bootsy like the meteor? Or what's the part of the image here?

(06:04):
Yeah, it's like a rise of a new evolution of distributions. Fedora Silver Blue,
Endless OS, Bluefin, Bazite, Ubuntu Core, Nix OS, Steam OS based on Arch.
Manjaro's working on Manjaro Immutable. Arcane Linux is Arch-based. Blend OS, Vanilla OS.
Seuss has a Plasma-based and a GNOME-based Immutable. We're seeing this rise

(06:28):
across all of the distro makers of these immutable or atomic distributions.
It does seem to be like it's not a trend. It's something everybody's getting involved in now.
And I think when we first started talking about it, it was like,
hey, check out this new cool trend.
With this immutable base, there are also a lot of these distros are going for
this experience where the user doesn't even really notice it.
You're not even really aware of this, you know, the fact that slash user is

(06:51):
read-only or something.
Right. I mean, so part of it, right, is it's changing the customization layer.
So maybe instead of adding a PPA and then installing Zoom or downloading a tab,
right, you're pushed to use stuff like Flatpak or Homebrew or,
you know, other third-party options that don't mess with that part of the file system.
And as a result, if you do that, then they can say you don't need to locally

(07:14):
compose that part of the OS anymore, right? You don't have to worry about stitching
it together into a working base sort of Linux system.
We'll do that ahead of time as a single atomic unit. We'll ship that to you,
and then you just switch between those, and you can keep your layers on top.
So this is where I think we have a divergence, is I never really had this.

(07:36):
I had a different experience when I went from NixOS to Bluefin.
With NixOS, I'm composing the entire system.
I'm doing these atomic updates where I switch into a different build it felt
like its own unique beast and so the immutability and the atomic updates were

(07:56):
just part of this unique beast.
And it's hard not to notice your use in Nix OS it's very different,
you did it very intentionally and the way you interface with it is not really like anything else.
So when I went to Bluefin where it felt like traditional Linux in that it has
the standard file architecture that you expect, the file system architecture you expect.
The old FHS. Yep.
It was more of a... It felt more of a experience shift.

(08:23):
It was more realized that this was a system that was designed to be immutable,
that I wasn't really supposed to be modifying this directly.
And it sometimes was a problem.
But it made me realize that it is a bit of a post-Tinkerer's distribution, these types of systems.
In the sense that...

(08:44):
You know, this thing, Bluefin, for example, it ships with TailScale integrated.
But what if I wanted to swap TailScale out for Nebula?
That's actually a surprisingly large amount of work with a big amount of learning
to figure out how to customize that image.
You very quickly end up in, I need to customize an image territory.

(09:04):
If you just want to, say, swap out a component like TailScale.
Right. Yeah. And that is where if it's baked in, there's not necessarily as
easy of an option to unbake it. And so if you're doing it at the layer where
you want to switch out a flat pack you installed.
Easy peasy.
Yeah, exactly. But to make your own customizations, that's where you start having
to engage in the build pipeline.

(09:25):
So either this is for people that really like to tinker or it's for people that don't want to tinker.
And, you know, it's not for people that like to mess with computers.
I think it's for people that want their computer just to work.
And there is a lot of people that started with Linux to play around.

(09:45):
Their computer was a toy.
They got to learn. They experimented with different desktops.
They distro hopped quite a bit.
As time went on, they got busy with jobs, perhaps family or other responsibilities.
And that toy turns into a tool. And then you extend that far enough out,
and it gets to the point of you don't want your computer to have problems at all.

(10:06):
Anything that stops you from just getting work done is very unwelcome.
And for that category of user, which I suspect is actually the vast majority of real-world users...
Immutable distros, especially ones like Bluefin, they're totally built for these people.
The developer that just wants a really solid workstation that is container-first,

(10:29):
man, is this perfect for them.
See, that's where I have a bit of a question. I think for developers specifically, I'm less sure.
There's a lot, because that's where I feel like they're trying to market at
two different groups. And that's why maybe Bazite has a lot more users than
any of the rest of these, right?
Because people running Bazite, that's what they want, right?
I mean, you want a thing that just runs games and doesn't break.
Absolutely.
And I think there's a class of developers, and Bluefin especially seems pretty targeted that way.

(10:52):
And if you can fit into their model where you don't need to rebase, then that works well.
But it seems like if you are a developer and you do need to make any changes,
you are going to have to engage.
So I guess it just feels like on one hand they're saying, we want to build this
in a way where people who have modern cloud container DevOps skills can change

(11:12):
it. but that's a different class than people who just want their computer to work.
I think the people who want their computer to work don't really want to also
then build a whole bunch of container pipelines for their computer,
but there's not a lot left in the middle between run exactly the image you get
or fully engage, and that's where I wonder about the developers specifically.

(11:33):
It is actually, you're making it, the point you're making really is that it's
a very targeted type of developer where it just works out of the box.
Yes. And that might be the majority of them.
It might be.
Yeah.
It might be, but you're right. Like, if you want to swap out a few core components,
maybe for whatever reason, you got to go with the Docker instead of the Podman.
That's a lot of work. That's no joke. So you're right.

(11:58):
If you fall into this world where you get most of your stuff either through
brew or flat packs, and you're good with the stuff they include,
I think these immutable distros have this very attractive set it and forget it,
actually semi-bulletproof a quality
to them that i suspect makes them a stronger competitor to commercial desktops

(12:24):
if you can fit and that's why it's nice we have a lot of options here but if
you can fit within their defined parameters i think these immutable distros
are a stronger contender.
Well i mean i think they're spot on the most people devs included unless you
are someone who does want to dev on the desktop, you don't care about doing the composing, right?

(12:44):
I might care that I need to add a package that I can run, but I don't care at
all about how the package manager works or the nuances of how the init ramifest gets regenerated.
That's exactly the stuff that you just should run in the background.
And I think we have seen the first signs of this. What do you think, Brent?
I have always been unsure about these immutable systems because everyone I know

(13:05):
who's been running them have been super technical and want to tinker with stuff.
but then i hear yeah this is going to
be perfect for the user who doesn't want things to break and just
wants to get things done and i wonder if like
last episode where we explored omarchie and
the whole point was to use linux for its strengths is this a version of using

(13:30):
linux for its strengths that might get like bazite a big input of users that
are looking for a specific thing that traditional distros didn't offer?
I think so. And I think, you know, Bazite is a great example of where we're seeing this.
If you, that's, Bazite out of all of the Universal Blue projects has the most traction.

(13:51):
And it seems to have appealed not because it's technically interesting and not because it,
you know, includes this or that, but simply it just makes gaming on a device
like the Steam Deck really approachable.
And it solves a lot of challenges just out of the box.
And it's created this bit of a perception shift around Linux,

(14:15):
and I think you particularly see it with YouTubers.
It's the gaming that has sort of opened the door for these popular YouTubers
to talk about Linux and open it up to a new demographic.
And I think it's winning because it's working, right?
It just works. And kind of, especially with, like, they were pretty early, right, with, like,
getting all the graphics stuff shipped, which is exactly the layer of,

(14:36):
you know, desktop dev that you can be an expert at, like, using Linux as a dev
workstation and understand nothing about Mesa and OpenGA.
And there's all these layers of how graphics and gaming should work on the desktop,
and now you don't have to care.
Here's another thing that's been a weird shift for me, is now after a while
of using NixOS and now 50 days of using Bluefin, when I do jump on an Archbox,

(15:00):
I don't remember, we were just on Arch recently for something.
Oh, Merchie.
Oh, yeah, right. When I jump on an Archbox, I feel like I'm flying without a parachute.
Like I got no safety net, even though I'm not afraid of breaking the system necessarily.
But it feels like you're using the computer in root mode almost constantly.
Wait, I can just install packages right into my root file system?

(15:22):
Yeah, it's weird.
Like an animal?
And this is where I think if this idea that these immutable distributions or
desktop experiences, whatever you want to call them, become more and more popular,
which it does seem like that is happening.
And myself, I don't really see myself using non-immutable distros and ones that
are atomically updated going forward.

(15:42):
I just, I don't know why I would go back.
Interesting. I do think Ilmarchi and the Omicube stuff makes for an interesting
thought experiment just in terms of how we classify things, right?
Like what is and isn't a distribution?
Because like in some ways, Ilmarchi feels like a lot of what Blufin,
especially DX is doing in terms of making a well put together development,

(16:03):
ready to go workstation.
they just didn't they left all the regular stuff
of you install arch or ubuntu and then run our stuff on top and then the bluefin
and you blue folks are taking going one step further back to then not be the
full distribution as we were talking about but to step into more of bottling
that and shipping it as atomic units that you can deploy and then you have all

(16:23):
the way up to like your fedoras building the whole thing.
But the universal blue stuff is is a more modern way to do it than the way the
omararche omacube stuff is right that's bash scripts and installing packages
from the package manager.
It's like a layer, right? It's kind of a, it's almost the exact analogy with containers, right?
It's, they took a lot of, because if you look at the structure,
there's a build.sh that sources a bunch of other build modules.

(16:46):
In that respect, it's very similar.
Once you're in the containerized build environment, what's actually happening
between it and omarch are very similar, I think.
Right.
But then you just stuff all of that in the OCI Bootsy layer.
But the difference, yes. But the difference is, right, it's happening on each end user system.
There's a lot of variability there. There's a lot of things that can go sideways after time.

(17:06):
A big part of this is shift left, we do the builds, and then we know the build
worked, and then we ship you that.
Exactly. So there is that difference. But otherwise, yeah, how they actually
get composed is at the end of the day relatively similar.
That's why I think it's so fascinating because it exposes like this,
probably a spectrum is not enough dimensions for this whole chart of like different
options of where you can be on quote-unquote distribution space.

(17:27):
But here's where I'm going with this is...
I think Nix OS is, if you will, a divergent path in this immutable future where
these other systems try to hide some of this from the user.
And they try to, you know, maybe they offer brew and the flat hub and they try
to, like, you know, give some options there.
Where Nix OS is like you go all in and you're managing that.

(17:50):
You're the one that's constructing that environment.
It feels like Nix OS and will be the escape hatch where you have this future,
but you're in total control.
and you're composing it yourself on the fly with the configuration versus these
image-based ones where you're basically taking their experience and then you're
building on top of that if you know how to do that.

(18:13):
It's going to be two very different paths in the immutability world and the
best way to really appreciate and understand what I'm saying is to try them both.
For sure, yeah. Because I think that's where there really is an individual layer
of like what works for you what escape hatches do you actually need which things
do you want to tweak and do they fall into the the class of how easy is it to

(18:34):
do and which how often do you need to do it and how annoying do you find the process and.
If you haven't tried one yet it might be worth doing
that just because this is such a trend that seems
like it's not going away like i said we have just about every distribution under
the sun that's taking a stab at this there's a lot of options out there so there's
some shopping you can do there's some playing around you can do there's some

(18:57):
experimenting you can do to really get a sense of this and at least have some
experience even if you you know don't make it your daily driver.
They're really nice too i mean you know bluefin is a sort of secondary uh os
on your machine to boot into if you have other problems or you just you know
whatever true it's really nice you know it'll be rock solid it'll update in
the background and right now i mean they're nice and fresh they got you know

(19:18):
recent fedora great desktops it's a good time,
and Blufin's got that fresh installer, which is also pretty nice.
1password.com slash unplugged. That is the number one password.com slash unplugged, all lowercase.
Take the first step to better security for your team by securing credentials
and protecting every application, even unmanaged shadow IT.

(19:42):
If your employees bypass security to use unapproved apps that they feel they need.
To do their jobs? Well, you're not alone there. Fortunately,
with 1Password extended access management, security and productivity,
they don't have to be at odds.
Now, I know you're careful about security. You listen to this show,
but would you bet that everyone in your company is as careful as you are?

(20:03):
See, that's where Treleka comes in. It's by 1Password, and it makes security
simple for every employee with every application they use, even the apps that
IT didn't necessarily know about.
There are so many SaaS applications floating around companies right now.
And if you can't keep count of every one of them, I know I can't. You're not alone.
That's the superpower that Treleka gives you by 1Password. It helps you discover

(20:25):
and secure access to all those SaaS apps, even the ones that are unmanaged.
There really is more than just securing passwords these days.
Managed and unmanaged applications, for instance, are a big issue.
So what I'm saying is you can check compliance off your list with a system of
record for your app inventory and employee lifecycle workflows.
With Treleka by 1Password, employees are empowered with secure,

(20:48):
flexible app access with a self-serve app hub, simplifying the process for users
to gain or request access to tools, which enhances security and productivity
for your entire SaaS ecosystem.
You can really reduce unnecessary costs and get a handle on app usage and identify
and eliminate unused licenses and redundant apps.
So go to 1password.com slash unplugged. Take those first steps to better security

(21:12):
for your team by securing credentials and protecting every application,
even the unmanaged shadow IT stuff.
I've been there. I know how hard that is. Go learn more. Support the show.
Go to 1password. That's the number 1password.com slash unplugged. All lowercase.
It makes the difference. Check it out at 1password.com slash unplugged.

(21:36):
Well, we've heard from some of you listeners who've been using these immutable
systems, I don't know, since they first came out.
But I want to know, are you going to try this? Have you been using it for a
while? How has that journey gone? If you want to boost in, let us know. That would be amazing.
But mostly today, Chris, I'm curious, how has your 50 days of Bluefin gone?
Can you believe we're here at this party?

(21:58):
No.
Like 50 days.
Oh, I guess that means it's been 50 days since Red Hat Summit.
That's right. It was Boston. I was inspired by all the Bootsy and cloud-native buzz.
So I had a Knicks book that Olympiad Mike had just recently gave us,
and I thought, all right.
Sorry, Mike. This is the most sacrilegious thing I could do, but here we go.
So I deployed Bluefin on that laptop while we were at our Airbnb.

(22:21):
And then when we got back, we deployed it on what had been a Neon workstation,
which is now, well, currently an Aurora workstation.
So, of course, I played around with it here and there live, but I mostly just
put it on the two systems I use the most, which is my laptop and the studio workstation.
And it's also not only my 50-day mark of using Bluefin, but it's also the four-year

(22:43):
anniversary of Universal Blue, the project.
Which congrats to them.
Yeah. And they started with Bluefin, so that technically makes it four years
of Bluefin. So I'm using a distribution that just turned four years old.
There's about 25,000 weekly check-ins of Universal Blue systems,
Bazite being the most popular, Bluefin and Aurora are up there.
Of course, that includes Silverblue and Kino Night as well.

(23:05):
You know, regardless of if it's your community or your distro or image or not,
it's just, it's great to watch folks building and finding software and Linux
stuff that works for them.
And there's like a lot of organic excitement around UBlue and the team behind
it. And I think that's great.
And the thing that's neat too is a lot of the original team still involved.
So that gives it a unique vibe.

(23:26):
In fact, you could argue they're still very much the center of their community
and their community really takes place i think on their discourse,
they don't have like a reddit presence or a big social media presence outside of that,
but it gives a kind of a unique vibe where these people are still very much
involved they're approachable and you know they're making news and they're still

(23:46):
there it's it's fascinating it's it's definitely has its own dynamic unlike
unlike other distros and i think it's good,
so let's start with the good and i'll get to the bad uh of using bluefin now for 50 days,
I think my favorite feature has been the auto update and forget it feature,
which you can turn on. It's not on by default, but it's nice.

(24:07):
And after about a month of not rebooting my system, because it was just rock
solid, I did get a notification that said, you should probably think about rebooting.
Nice.
That's a good touch. Yeah, yeah, it was. The Ujust command, which comes with
all of these Universal Blue-based images, is really nice.
It makes it easy to update everything. It lets you do lots of system tweaks

(24:27):
and additions like setting up distro box or virtualization, gaming,
turning on the command line bling and a lot of things like that.
It also has like a 2E mode where you can see all the options and get a brief
description and see what they do. So that's pretty nice.
And of course, it is immutable. So it has a read-only root file system.
So only var and etsy are writable by the user.

(24:48):
And that really hasn't been an issue once.
It's really been nothing but positive. I don't know, Brent, if I told you,
you know, you couldn't write to anything but Varanetsy, you'd think,
well, that's not going to work for me. But it's been totally fine.
I actually can't remember the last time I wrote to anything but Varanetsy, to be honest.
So as soon as you say, oh, no, those two folders are writable,
I was like, well, what am I? Am I losing anything? It doesn't sound like I'm losing anything.

(25:12):
And I think they make a lot of good out-of-the-box app choices.
They just recently decided to change from GNOME Software to,
I think it's Bazzar, which I've been trying and I liked a lot.
You do end up needing to, I think, add a lot of additional software and set
up a lot more stuff, but the base stuff they include has been really good.
See, this is one of my big hesitations, actually. You mentioned in the previous

(25:33):
segment how difficult it is, let's say, if you want to switch to Docker from
Podman, which is built in by default, and I'm just curious...
why not have a version of bluefin that's so vanilla but they have flat packs
that are pre-installed because that would be a lot easier to modify for instance
i'm assuming it's because these some of these choices need deeper system i don't

(25:58):
know integration is that the case.
Well some things are like like firefox and whatnot those
are a flat pack uh in fact in the build you can
actually see at some point it uninstalls the firefox rpm and then it installs
the flat pack of firefox like when you go through and watch a custom build of
it right so that's been interesting but i think some things like brew have to
be installed in the way they are so that way you can put packages on the system

(26:22):
so some things do require that level of integration.
Right like if they want to deliver the appliance part
then they need some stuff to build with so i think that's to some
extent right like if you want to install a bunch of the bling well some of the fancy
you know shell helpers come from brew or if you're gonna you know ship images
and you need a bunch of escape hatches then you probably need a container runtime
so then you're going to have one and it's pretty tightly integrated and hooked

(26:44):
into the terminal and then so if you want to opt your options then do you have
to how much work is that to swap out or parallel maintain so.
They've tried to make you just handle a lot of that stuff.
But you can definitely you know build from copy their build and from the upstream
fedora stuff and tweak stuff out that way if you'd like to.
If you really want to learn how to.

(27:04):
We can talk more about that yeah.
We'll talk more about that But what I will say, it really has been very minimal maintenance.
I mean, the least fussing, the least maintenance I have ever had with any desktop system, period.
Chris can't break it approved. You heard it here, folks.
Right. It is solid. So here's the bad. That's the good, here's the bad.

(27:26):
What has worked amazingly well
on my studio workstation has made me feel a bit boxed in on the laptop.
This was most felt when setting up the system.
Some changes require pulling down very large images for what seems like a simple change.
And, I mean, I'm talking gigs and gigs of images, which, not a problem on the
studio workstation on Ethernet.

(27:47):
Bit of a problem when you're on the go with a laptop on LTE.
But after the system's mainly set up, you will run into this from time to time,
but it's pretty livable.
So I don't want to overstate it, but it's definitely worth having a solid Internet
connection when you first set up your box. You will feel that.
I will admit some of my bias here is I'm not a big brew fan.

(28:09):
I never really have been.
In my opinion, it's slow. It uses a lot of computer resources when it is running.
And the package availability, while decent, nowhere near AUR or Nix.
What, you didn't want a package system written in Ruby? I don't get that, Chris.
And so I found myself really suffering during the TUI challenge.

(28:30):
I mean, if it was something that I couldn't run via Podman or Flatpak,
I had to play this game of DistroBox roulette.
Which package manager out of the available DistroBox images can I actually get this installed in?
And man, was that tedious and a real pain in my butt.
And, you know, they have this thing called developer mode where you kind of

(28:51):
like kick it up into a more advanced user mode.
And it made me think, well, they're not opposed to having more advanced users and modifications.
So it'd be really nice if there was like a ujustnix command.
that set up the Nix package manager or the Determinant Systems Nix installer
in like a Bluefin best blessed way where you had a slash Nix.

(29:12):
I mean, the software availability would be incredible.
It would open up, you'd have Brew, you'd have Nix. It would be amazing.
You could get just about anything in the free software world installed.
But I did a little digging around and it seems like the project not really a big fan of this.

(29:33):
I kind of find this fascinating i was looking to because i don't know if you
remember they used to ship it.
Yeah that well if you go look it up on google google ai answers still claims they do.
And they also ship dev box which might have been why they were shipping it um originally,
yeah so there's kind of some interesting um history there you can see some users
sort of like oh why did this disappear and it's another case where people are

(29:56):
kind of confused because kind of just low-key change log notes and then i guess
most of it happened on discord conversation around it,
and so there's multiple people sort of asking and digging up how it happened and where.
And there's definitely been lots of asks over the years. So I'm kind of getting
a little bit of spicy pushback.
Yeah, I mean, it's usually it starts with, sorry, this is out of scope for us.

(30:16):
But then if the person pushes back at all, it kind of gets a little contentious.
Like, I think it's totally fair to have, you know, they get to say no,
they get to make this project.
They're totally right to point out that you can, you know, they publish a whole
bunch of docs on how to go make your own images on top of this and have a whole
workflow designed to let you do so.
So that's all totally valid. But I do wonder about that, like,

(30:37):
A, just that a lot of people keep asking, that's a signal.
And then B, like, if you have to add brew, is it really out of scope?
Like, I know they've had issues with SE Linux. None of them use it,
so they're not really, like, they don't want to deal with the maintenance burden
or any maintain any of it.
It's just sort of, if something feels a little inconsistent that it used to,
did it used to be in scope? What part of the scope changed where it's no longer in scope?

(31:01):
And as a user, it would be such an unlock. It would be such an unlock.
And my kind of next best is an Arch Distro box with the AUR,
but I'll tell you, the AUR hit rate ain't what it used to be, kids.
Like, I have a lot of stuff failed to build with AUR that I have no problem
installing from Nix packages.
Brew also seems, right out of the box, like a strange choice for a Linux distribution

(31:26):
to rely on as its primary place to get, you know, packages for user land.
I think they specifically like that it is distribution agnostic that's because
they don't really consider themselves a distribution right there they're built
on top of these core images and then they can layer in brew which works in user
space and is distro agnostic itself.
And a part of the issue with nix was sc linux compatibility uh and homebrew plays nicer.

(31:51):
With that out.
Of the box i believe.
I just feel like when's the last time you had brew on a linux distribution period
like i don't think i've ever run into it so it seemed like a an odd choice i'm
glad you got to uh at least give it a taste chris and tell us.
It is funny i i installed something recently and
the air output was about being able to undo some unable to do something on mac

(32:14):
os like it had some sort of post install script for max and it tried to execute
it and it failed so okay so say a guy wanted to add nix just the package manager
so you know you still had bluefin or Aurora.
You had the file system standards. You really are in a kind of a Red Hat Fedora
experience, but you just wanted the Nix package availability.

(32:36):
It's a great place to run it, really, in that sense, because you don't have
to care about the OS. You can just use it for isolated dev environments or for
running whatever package.
This is an example of where you need to start making your own custom images.
And so we thought, well, this would be a neat opportunity to try to build Blue
Nixfin, which, Wes Payne, you gave a go.

(32:56):
I did give a go. We'll also have some links. There's a few different folks out
there who have taken stabs at this as well.
Yeah, it was kind of fun to go learn more about how you go build your own images.
I tried both from cloning the whole repo for Bluefin and building that just
to see what a base from Silverblue, from Fedora, building the whole thing looks like.

(33:17):
But if you don't want to make crazy changes or not add, if you don't want to
remove stuff that's deeply tied in, they've got an image template repo that
you can just fork on GitHub or clone locally.
It's all powered by Just with a Just file. So that's your entry point,
just like with Ujust on Bluefin.
And what this does is it's got a container file set up that runs a build.sh

(33:40):
script that's ready for you to add your own scripts or whatever you want into.
So is this where you would say, like, create a slash nix directory at the root?
Exactly. You can install extra software. You can add extra directories or manipulate stuff.
Because, like, one of the core complications here is you need a slash nix,
which you don't get unless you modify the image.
That's also been made more difficult recently. I believe
as of 42 they're now moving to use composefs as

(34:04):
the thing that like composes and puts together this read-only root file system
which has better integration I guess it also allows like hashing and checksumming
for security around like what the final file system looks like and stuff but
it means it's even harder to do hacks at like runtime without modifying the
image to make it have a and especially for Nix which is like really hard.
all the stuff that's built is assuming a slash nix at the root. So you really need it.

(34:29):
So you get in this build.sh, you make your modifications.
Yeah, and so that's run as part from the just file as building.
It starts from, like, it's just a container file at the start that says from,
and then Bluefin, Bluefin DX, Bluefin GTS, Aurora, you know,
whatever the base image you want.
And then it handles running your build.sh for you in the right environment that's

(34:52):
set up so that like RPM OS tree can run and commit your changes at the very end automatically.
So you don't have to think about that. So base one of just getting Nix going
is, yeah, you just need to add a make dir slash Nix so that that's there to work with.
And then shout out to the determinant Nix folks because they've done a lot of
work and continue to do so to not only get Nix to play better.

(35:16):
I mean, you don't get a whole bunch of extra like verification,
but just to play better out of the box with SE Linux-enabled systems, which is great.
And then also specifically to make their installer work on OS tree systems.
Now, this gets tricky if you're trying to do it. I was building my own version
of their installer, which obviously has a flake, so it's easy to do,
which is great, because they know to figure out, basically, you use slash var, right?

(35:39):
So you make slash var slash nix or something like that, and then you bind mount
that over the empty slash nix that you make in the image.
And so if you leave them a slash nix and you run it at runtime, that'll all just work.
Now, it gets trickier if you do what I was trying to do, which was build one
where you could like fully bake Nix in.
I got that working in single user Nix mode where you'd have to use sudo or like

(36:01):
grant yourself permissions.
With the multi version one, it's tricky because when you're in that build environment.
Systemd isn't running, but you want systemd to work.
And I think right now some of the setup process sort of assumes that they can
like write stuff and then reload the daemon and then check that that worked
and other stuff that you do at runtime.
So you're telling me to really just make it work with the Terminant Systems

(36:23):
Nix installer, you just needed to create slash Nix?
And it could either be var Nix that's bind-mounted or whatever.
Well, that can happen at runtime because slash var is writable.
You need the right name existing in the image at runtime.
So it seems like the bluefin image just needs slash var slash Nix to exist and then a ujust option.
Not even that, just slash Nix at the base.

(36:45):
Okay. Well, I was just thinking then a Ujust image that bind mounts it or something.
But yeah, just like a Ujust option that just has an image that has slash nix.
Or, and I think this is what they would push for, right, is some community folks
who want to copy all, you know, they have, it's all integrated on GitHub Actions.
You can set it up to build automatically and build a downstream version that adds it on.

(37:06):
And I just feel like, yeah, yeah.
If you look at some of the issues, their perspective is that even adding that
sort of implies support.
And I think it's folks that recognize, let's say, that Nix is complicated and
that they don't, you know, it's not a tool that they're engaging with.
So I think maybe they've had some bad experiences, perhaps. I don't know.

(37:28):
See, this is where I feel like, though, I am going to struggle to keep using
Bluefin on my main laptop because, like, I would like to be able to swap out some of the components.
But I don't want to go through the process that you just went through.
Where with Nix OS, I would just add it to my configuration. I need to know the
syntax, but I can find that pretty quickly and then I rebuild and now that exists.

(37:50):
Yeah, there is.
So that's what I'm struggling with. Whereas like on my workstation here in the
studio, I don't make those kind of changes.
That is where I think there's kind of an interesting, maybe gap isn't the right
word, but yeah, right? It's like it's optimized for different parts.
It's optimized for the appliance user and then it's optimized for like the distro builder.
Like it would be a great, it would make total sense if you were going to deploy
a fleet of machines and you wanted to bake Nixon and engage with this process.

(38:12):
If I was DHH and I lived in the container world, this is how I would be building
Alma Archie. I mean, how cool would it be to have this, but Arch?
Like, that'd be really something. Then you'd have the AUR.
But I agree that it is kind of, unless you're already doing a bunch of container
stuff all the time, it is, I mean,
I will say, I think you could definitely do it because for their image template,

(38:33):
it is, you could fork it on GitHub and then just add one line into the stuff
and then have GitHub run it probably even.
Yeah.
And then have an ISO file you could use or a Docker thing.
It's definitely doable.
Yes.
Yeah.
Maybe, you know, because there are things that I do like about it.
And it is a cost where like once you figured it out and you can even use GitHub,

(38:55):
it's all set up to like push to GitHub repo and stuff.
So you could have it run, bake it in, and then you could just rebase locally.
And then once you set it up, you can even set your system to just follow that for updates.
So then you go into GitHub, trigger the action or whatever to build a fresh one.
Right. That's how you'd want to do it. Yeah, it could be worth looking at because
like I've mentioned it a couple of times in this episode so far,

(39:17):
but having the file system hierarchy standard actually on your system turns out is handy.
Because what you get is every application that you run, it just sort of seems
to understand it's on a Linux box, it knows where stuff is, where this can sometimes
be an issue with Nix. True. Nix OS.
Not an insurmountable issue, and there's tools to help mitigate it,

(39:40):
but it is something you just have to think about, and it's something you don't
have to worry about with Bluefin.
Right. It might be somewhat more incompatible because things can't write to
all the places they might want to.
But for the most part, you probably don't want them writing there without very explicit permission.
And I don't have to decide which version of VS Code I want to install.
There is also something nice about the fact that they support writing to etsyudevrules.d.

(40:05):
Now, this is intended to be writable. So when you install a package,
if it's adding hardware support for a device, it can write it to rules.d.
The way this works in NixOS is you declaratively define a UDEV rule in the configuration,
and then it generates that rule for you.
Now, that's probably a better long-term way to manage a system,

(40:27):
but an example is like...
You know, you want to connect a quick hardware device to flash it or a steering
controller to play a Steam game really quick.
And you forget that, oh, yeah, I got to actually go define a UDEV rule for this,
where with this type of system where it has an actual rules.d that you can just
write UDEV rules to, those devices tend to just work because a package adds that rule.

(40:51):
And that experience is a little smoother.
Perhaps not the best long-term way to do it, but is a smoother end-user experience.
Yeah, there's some more escape hatches in that way. Which I think is kind of
interesting because there's, you know, it is more of a, like a component composed
system in that you have like this core bit and then you do have,
you still have a mutable, writable Etsy that's going to be per device on there

(41:12):
in a way that's a little different than like how NixOS does it.
Then you've got like the flat pack layer that you might install and then the
homebrew layer as well that all act together to get you your final system.
The experience I've had though with the studio workstation, I mean,
this is such a no brainer for a headless server.
if you just took out the desktop graphical environment and just ran this.
This thing, this low maintenance, if you're running everything as an image anyways,

(41:34):
man, it's just such a great way to run a server. And I know they're working on something.
Yeah, they're working on a new one. So they had one based on CoreOS that I had
not tried, but we'd been poking at.
So they have a new one called KO. Yeah, there we go. A boot C server image for
your self-hosting needs.
And yeah, you could totally see how this would make a ton of sense,
especially where on a server, you might not need to do nearly as much of that

(41:57):
kind of, you know, base level changes or customization, especially if you're
mostly just running a bunch of containerized workloads already. It'd be killer.
I feel like I'm sticking with it. I mean, I think they also.
They do note they like include ZFS out of the box too. So yeah.
Not only do I just sort of have everything set up, but I just rebased the studio
workstation to Aurora and I haven't really played around with the Plasma version

(42:18):
yet. So it seems like an opportunity to keep doing that for a bit.
Yeah. What'd you think of the process?
Super smooth. I really really liked the boot C command line stuff is really simple to understand,
one-liner did the whole thing literally a one-liner did the whole thing.
And kind of just like what you're used to from nixos right like if you reboot
you could go choose the other one.
I could roll back i could roll back that's a big deal i mean it really has been super bulletproof,

(42:42):
to the point where it had to remind me to reboot you know and that i think is
a pretty good stamp of confidence right there too it's a really solid system
and i feel like it's being built on top of,
fedora 42 you know the latest fedora stuff which is really good but always just
needs a little bit extra and they're adding that yeah.
It definitely you know uh definitely a lot closer to like a shiny workstation
than you get with with base fedora as lovely as it is.

(43:04):
And i'm a lot closer to more like a rolling fedora i mean they have they have
to do the updates and they have to rebase when fedora 43 comes out but they're
going to deliver those updates to me and i'm just going to continue to just have more stuff.
In flat pack and brew which you can update more on your own timeline.
Yeah and then you know there if you use the you just auto update not only will
it do all of the image updates for your system, but it also will do all the

(43:27):
brew and flat pack updates in the background for you too.
And the idea is, is embrace the reboot from time to time, which I need to get
better about. And then you just come into a fresh system that's with everything up to date.
Well, you'll just have to start telling us your system uptime and we can shame
you until you reboot more often.
Unraid.net slash unplugged. Go unleash your hardware with Unraid.

(43:50):
It's a powerful, easy to use NAS operating system for those that want control,
flexibility, and efficiency in managing your data.
Unraid is built on top of modern Linux, and it allows you to mix and match drives of any size.
That's a big one right there. That means what you have in your closet right
now, you could throw into production and start utilizing.

(44:11):
It also includes built-in support for tailscales, so just like check a box and
the applications on your tail net.
Static IPs, of course, are supported, but even better for those of us that can't
run it, It now also supports Wi-Fi out of the box.
But the thing that really sings are the thousands of applications and the super active community.
Not only are they great with support, but they're always coming up with stuff.

(44:33):
I just noticed a community member, John M,
recently submitted a script that can automatically monitor the CPU usage of
your Unraid box and then shift power profiles so that the Unraid box is using
the most efficient power profile per CPU demand. And it can do it dynamically, too.
It's just that kind of stuff. There's just a brilliant, brilliant,

(44:56):
passionate community around Unraid, which means that when you do run into some
sort of challenge or you have a question or something you want to try,
you're generally going to find somebody else that's done it or can help you figure it out.
The other thing that's really nice and just saves you a ton of time,
the VM stuff for like passing through a GPU or sharing a GPU or getting templates
so that way you can just design something and then always build from that.

(45:19):
All really, really straightforward with Unraid. Probably the most efficient way to do it.
And then in 7.1, they took the ZFS support that's been in Unraid for a little
bit, and they really brought it to the next level by making it possible for
you to migrate from an existing Ubuntu system or a Proxmox box or a free NAS.
You know, when you're ready to go something a little more powerful that has

(45:39):
a bigger community and more applications, a little more flexible, well, that's Unraid.
And so it's really something that lets you get started with a lot of the stuff
we talk about on this show in just minutes.
So go learn more and support the show. You go to unraid.net slash unplugged. Go check it out.
You've probably heard about Unraid, but have you actually tried it?
They do a lot of neat stuff, and they have a great model over there to make

(46:02):
sure they can continue building that thing and following the latest developments in Linux.
Check it out and support the show. Go unleash your hardware without having to
lock into some sort of rigid setup.
Unraid.net slash unplugged.
Well from last episode we received a little feedback from
zach writing hey i haven't had a chance

(46:24):
to try it yet but when you guys were talking about omarchie and
having a better update mechanism to remove things that you
didn't want in there first thing that came to mind was the
common arch project seems to be
very similar to the boot c and bootable container
projects but based on creating arch and other distro
based containers instead of being tied directly to os

(46:46):
tree and like boot see is in theory you could basically run the omarch install
script at container build time and then have a bootable image that could just
be installed directly and updated atomically i haven't played with it yet but
i'll be sure to report back if i do and see if you guys give it a try.
Oh thank you zach yeah this looks

(47:08):
very interesting i don't think i've heard of common arch
before but having a similar idea
right like i think one of the strengths of these kinds of approaches
is just like containers worked to wrap whatever language
and runtime that you needed right they can work to wrap whatever linux
you've needed as long as you've got the right sort of hooks to make it
deployable and runnable so if common arch does

(47:30):
that part of it uh that sounds great
we've got a link here from zach to their github org i don't know if they have
any more docs i see iso builder system base uh core which is a docker file so
i don't know the best way to find out more but if anyone else does i think it'd
be definitely something we'd take a look at it might be pretty new you.

(47:51):
Know on a quick uh i'm archie update i saw dhh post that they're working on
um getting it down to a five minute install from a fresh system so.
Essentially like an iso.
It has like a five-minute setup. He's also been joking about creating Omicron,
a Linux conference with Vibes.

(48:12):
And he just did a...
Did you promise we'd go?
No, not yet. He did a six-hour Lex Friedman interview, and I skimmed it.
I don't think Omarchie itself came up, but his thoughts on open source and WordPress
and removing his systems from the cloud did come up.
So it sounds like there's a lot more still coming to Omarchie,
and we'll keep an eye on it. I think it's a fascinating project.

(48:36):
I suspect you do too. I don't know, what do you guys think? You think it's fascinating,
right? Like not something we're necessarily using, but it's definitely noteworthy.
Yeah, absolutely. It's great to see people messing with, building on,
tweaking, configuring, and sharing, cross-collaborating on Linux.
I think the thing we're most excited about is the demographic he's going after,
the way he's talking about it and advocating it.

(48:57):
I'm definitely going to check back in soon. I think that's something we should do.
a couple of shout outs I want to try to do this more often so I want to just
give shout outs to people that are contributing value back in time and effort
to the show and two solid folks that do this on a near daily basis are CG bass
player and Chance M in the web chat,

(49:21):
they're always working on the website just making sure things go
but recently they had to help with a migration we moved this
week in Bitcoin to be fully hosted on the jupyter broadcasting website
there's a lot of little bits there they had to pick up for us
so i wanted to give them a shout out and then i wanted to solicit
you listening if you notice somebody in our community who's extra
helpful or helping with the show in some way and sending value back call them

(49:42):
out either with a boost or a message in matrix we want to try to give those
people more attention because it's uh the community members that are doing those
things that really make a big difference we really appreciate it so shout out
to all of you out there who contribute some value back in time and talent,
And, of course, we have those of you who contribute your time,

(50:02):
talent, and treasure, and the treasure in the form of a membership or a boost.
And we have a couple of great boosts here. In fact, I'm going to say it's KS
Koba, who is our baller booster this week with 50,000 sats.
Hey, right. Hey, guys, I'm a new listener to Unplugged coming over from the self-hosted podcast.

(50:24):
I tried to get into Linux a few times over the last year to have a more work-focused
computer. I'm an academic and an emergency room doc, which means writing lots
of research manuscripts or journal articles.
This means collaborating with others on Word.doc files that has the track changes,
the comments, all the stuff, and needs a reference manager software.

(50:46):
That sounds painful.
This has been, as he puts it, my friction point, bouncing me off of Hyperland
and Ubuntu for the last two times, ultimately giving up and going back to Windows
where I can't find a comparable program.
Fair.
Any recommendations for others
doing this kind of work or ways to interact with Windows colleagues.
I dream of whizzing around my desktop using hotkeys, Tuis, and workspaces.

(51:08):
This is such a good question.
It's interesting for that kind of environment too, where I would assume if you
can optimize your keystrokes to save you time in any way, it is welcomed.
And if you get used to something like a tiling window manager,
that can certainly do that for repeated actions.
So I could see why there's this desire.

(51:30):
But interacting with other colleagues with Word documents.
This is tricky because if you still live in a world where everybody's using
the desktop application workflow and they're using this reference manager software
that's Windows-specific, that's where you get jammed up, right?
What's liberated Linux users over the last decade has been the migration to

(51:50):
web apps where it's all web workflow.
Doesn't matter what platform you're on.
Yeah.
Back in the day when I had to use some Windows apps like that,
I ended up just running it all under VM. So I just have a Windows VM and I would do it in there.
At that time, I was using VirtualBox, which had the functionality to pop out.
Windows from.
Inside your windows vm so i can't say i would recommend that setup to anyone

(52:14):
sane but you know there are things you could do like that but it would depend
on how much of you know if it just means a single giant vm window and you don't
get any of the benefits that's probably not worth it if it's like you can run
a couple apps in a way where you can still take advantage of your window manager
and linux shortcuts maybe there's something there.
Yeah i was hesitant to recommend the vm route just because that can be painful you.

(52:34):
Also need hardware that's going to do it in a way that isn't awful to work with.
But you could just have a dedicated like windows virtual
desktop and you you know you just run windows full screen on that virtual
desktop when you switch over to it you're in windows and then you switch back and you're
doing all your other work i mean it's doable let us know if you find anything
uh ks and if anybody has any recommendations please send them and we will pass

(52:55):
them along because uh i have been in that spot i've been there and i really
feel feel for you also thanks for boosting and listening to the show.
Nice to have you on board.
No stromo boost in with 25,000 cents.
We got some weird looking stuff and maybe improperly or incomplete base 64.

(53:19):
Did you have to fix his improperly encoded message?
Well, it looked like base 64. So it wasn't totally happy, but it did spit out
hi with a smiley face. So we'll take it.
Hello there. I like that you had to razz him on his improperly.
I mean, it could have been my software processing it too. Who knows?
I mean, you could have just not called him out. I mean, I'm just saying.

(53:40):
It could be my base 64.
Oh, yeah, no. You're walking it back.
He's back there.
Yeah, there you are.
I wasn't trying to blame you, Stromo.
Thank you, Stromo. Appreciate the boost.
Well, there's a mega rove ducks here from Derivation Dingus.
Hey, Chris, I think you nailed it with the point about how early we are on everything
here at Linux Unplugged. We aren't wrong.

(54:01):
The rest of the world is just slow.
At least that's what I'm going to tell myself.
You know what they say, stay humble and stack sats, you lovely bunch of trendsetters.
Oh, thank you, derivation. I've been reflecting on that. I still feel that way.
You know, we're out there a little, we're too ahead of things sometimes.
And it means that we don't necessarily address the mass market we could potentially be addressing.

(54:24):
But hopefully you're all still listening in 2050, the year of the Linux desktop.
One day we'll be right. One day. Mixit, that's how I say it,
comes in with 20,000 sats.
please more unhinged ai audio as you come across it you have no.

(54:44):
Idea what west creates on a weekly basis.
Daily basis let's be honest oh man do you have any recent favorites west there's been some good ones.
There have been.
Yeah yeah i love it when you get the voices that's always my favorite.
Well that british van lifer was pretty decent.
That was that was good that was.
But i'll keep that and try to pick a very good one sometime.

(55:07):
For you there.
Mixip.
There you go.
Doornail 7887 comes in with a row of ducks.
First boost.
Hello. Thank you.
Yeah. Coming in from Podverse. Nice.
Wow. So not only is it their first boost, but then they also set up AlbiHub, I'm imagining.
Probably, yeah.
Well done.
Episode 620 hit home for me. Brent loves building things.

(55:30):
I have 11 ESPs spread across a larger property with a barn, greenhouse,
and a garage, all separated from the house.
My guy.
Zigbee and Z-Wave never really worked, but Wi-Fi solved my problem.
One ESB8266 is a custom thermostat for our greenhouse furnace.
All powered by Home Assistant.
Saves a ton in propane costs.

(55:51):
Now, I can't seem to get fountain working at all. Email codes just say, that didn't work.
Any ideas? No. That's a very odd, that's a very odd error message.
But they do have a Telegram group. If you're on Telegram, you might pop in and
see if somebody can help you there.
Way to persevere, too. But, you know, bouncing off fountain and you still set a boost.
No kidding. No kidding. Absolutely. gosh dang it this i like you you're a hot

(56:15):
ticket thanks doornail i have fantasized about using home assistant to automate
a greenhouse because it just seems perfect both from the heating and the cooling
and the humidity monitoring you could also do soil sweet.
Plant dashboard yeah.
Yes dude you've.
Also fantasized about using esps to get a custom thermostat in lady jupes.

(56:37):
Oh for sure like to run the furnace yes absolutely one.
Day one day well bhh boosts in 2000 sets,
Hi, guys. I just looked into SniffNet, which was the pick from last episode.
That GUI toolkit is iced. Just figured I'd let you know.

(56:58):
Oh, thanks, PJ.
Yeah, nice to hear from you. Thank you for the boost. Well, PJ's here with 12,140 sats.
Ahead of the curve with Linux stuff, this boost amount is how long ago the last
Arch Challenge was prior to the release of this episode.
Can you figure it out? It's 12,140. 40.

(57:19):
I remember thinking, it's about dang time. Yeah, fair, fair.
Yeah, we're done. That's the crazy thing about our audiences.
They're ahead of us on a lot of this stuff too.
So Wes, did you do the math there?
Well, let's see, we got some hints, huh?
Yeah, the first two digits is years. Ignore the zero. So...
Is that 2013 if we subtract 12?

(57:39):
Yeah, yeah.
Okay, and then ignore the zero.
Oh, how long ago? Oh yeah, that's a tricky one. I was going to say 12.
It might be 2012 it's close.
Could be 2012 too yeah could be.
Yeah 2012 2012 2013 seems like the right range.
What are we doing with the 14 though it's 14 clearly yeah month 14 that makes.
Sense,

(58:00):
14 2012.
Got it all.
Right nailed it let us know how we did pj let us know.
Rotted mood boost in with 10 000 cents ah this is a check-in boost from episode
618 the two-week challenge kickoff we got some live long and prospers and mood
says getting caught up on old episodes a.

(58:23):
Time traveler check in thank you sir appreciate that that's nice see we're getting some data.
That's great yeah.
Pingu98 boosted in another row of ducks,
Keep up the awesomeness. I'm a recovering Linux on Chromebook addict.
Finally got working audio on my 2019 Pixel Go, running Ubuntu 2504.

(58:46):
It's like having a new machine. I could probably have done that ages ago with
Arch, but I'm not quite there yet.
Nice. You know, we all got into some kind of weird thing. I,
instead of doing Chromebooks, got into Netbooks for a while.
Oh, yeah, I had a Netbook.
Thought that was going to be a big thing.
It was a great size.
We did get unity out of that. Thank you, Pingu. Appreciate that.

(59:08):
Marcel's back with 5,000 sats.
Brent, I didn't tell you about these projects last time I saw you,
simply because they didn't exist yet. I'd be happy to talk about them next time
you're in Germany, though, wink, wink.
I can even bring the e-paper screen. I know you love Berlin,
but Bavaria is worth a visit, too.
I'm noting how a lot of these European cities start with bees.

(59:32):
So all the B-Cities must be the good ones, right?
Makes me hungry. Makes me think of breakfast.
Yeah, I did call on Marcel last episode because he mentioned all sorts of really
great projects that the last time we were together, we didn't get a chance to chat about.
So glad to hear you're learning new things and always pushing the envelope.
Sounds like we're on the same path. I like that.
And it is nice to hear from you, Marcel.

(59:56):
Confi, blue sin with 2,500 sats. No message, all sats. thank.
You sir appreciate the value.
But gene bean comes in with uh double ducks oh.
That's 4444 sats.
Oh and this one's directed right at me wes how do
you deal with the fact that esp home in nix is always way behind ah fair i've

(01:00:20):
tried going that route for building updates but it lags behind uh haos and the
associated add-in too much so far for me i've resorted just using a virtual
environment on mac os to do builds so that the version versions match and builds are fast so.
I'd be curious to know actually in practice why you need the latest versions of esp home that's what,

(01:00:43):
But generally, unless you're doing something pretty new, I don't really know if you need it.
Like in practice, I don't know how much it matters. And I'm curious if Gene
Bean has a reason why. You know what I mean? Yeah, right.
How often do you need to update if you're not making a change or whatever?
I don't think it's going to impact you.
I did a little bit of looking. And my first honest answer, Gene,
is I just really haven't used it really in anger yet. So it may be that I run
into exactly this problem.

(01:01:05):
I do suspect I may, I'll probably start pretty small and may not need to update
them a ton once they are working. So we'll see how much that matters.
So, I mean, you're talking about setting up pretty basic, like,
IR blaster, temperature sensor type stuff.
Yeah, that's where I'm starting anyway. And I've been surprised how well it has worked.
I did look, and right now, or at least as of whenever this boost came in the
other day, it was up to date in terms of the ESPHome package in Nix.

(01:01:29):
And it looks like the last few updates are pretty much just the real good kind
of Nix packages updates, where it's basically just changing a version,
the URLs all stay the same, the dependencies stay the same, and you just need
a new hash and a new version string.
and those are really pretty easy if I did need to,
hack that in to myself to get a new version I'd be willing to do that or even
just clone the package or whatever so may do that with an escape hatch but I

(01:01:52):
think right now this is homework for me to do some more ESP work this summer
so I can report back on if this is actually a problem in practice and if.
Gene has any examples I'd love to know.
Totally yeah I'm curious more about your setup there Gene he.
Also had something we should check out which looks like it's some Hyperland
setups that are kind of like ready to
go for like different And it's like Jakoolit, I think is how you say it.
Okay, yeah.

(01:02:12):
Put a link to that in the show notes.
I've already played with the Fedora one and plan to try the Ubuntu one today.
I mean, that's neat. Cool.
A passionate Linux enthusiast. Okay. And tech hobbyist.
Neat, but sharing cool setups. Yeah. Fedora Hyperland, Hyperland Ta,
Debian Hyperland, NixOS Hyperland's in here.
Yeah.
Okay.
Yeah.
Cool. Thank you, Gene.

(01:02:32):
Well, Magnolia Mayhem sent us, not quite a row of ducks. It's 3,333 sets.
all caps it says go podcasting.
Yeah go podcasting thank you go podcasting nice to hear from you mayhem appreciate
that that came in this morning as uh the live stream was getting set up so uh

(01:02:54):
thank you that's a fresh boost fabian came in with 2 500 sats,
Also, a fresh boost came in on our live stream. I disagree that Google isn't crawling more.
Every website I manage is getting hit by bots, including Google bots,
way more than I've seen in the last 15 years, especially in a faceted search page. A lot harder.

(01:03:18):
I have also been seeing a lot of noise about the bots intensity increasing.
We were talking about in the pre-show, like there's this number they have now
of something like, what did they say, 1500 bots scrapes to every one user visit?
Yeah, and I think maybe, because I think they were trying to say more that the
visits were going down, not necessarily anything about the scraping traffic.

(01:03:40):
But the scraping traffic is also going up.
That would make sense.
So it's a two, what they're getting at is that the problem is that a lot of
times information is either provided directly by Google or an AI assistant now,
so you don't need to go to the website.
And they're scraping more often. And of course, people want to have more current
information. So that's probably another reason why the scrapers are turned up. It is a problem.
I don't know if, like, it could be maybe we level off.

(01:04:05):
Like, it's really hard to say where this goes. If people have other insights
on this or if you've noticed traffic, your project, I'd really like to hear that.
Yeah, I wonder if we'll see more, right, like Cloudflare's trying to do more,
like, hey, if you paid a scrape.
And we've seen, like, the Anubis project you all talked about on Self-Hosted coming in.
Right.
So.
Well, thank you, everybody, who supported the show with a boost.
You helped make Episode 623 directly possible with your support.

(01:04:27):
shout out to all 24 of you who just streamed those sats
as you listen you just sat back and sent some value collectively you
stack 63,227 sats
just by streaming them when you combine that with our boosters we stacked a
total of 226,810 sats this week not too bad not too bad would love to see that

(01:04:50):
number creep up a little bit but we still appreciate all the support it really
is one of our favorite segments.
We get the interaction with you, new ideas come into the show,
and the messages are extremely motivating for us.
And on top of that, the entire thing is built on open source and open standards.
There's no middleman, there's no Stripe, there's no PayPal taking a cut.
It just goes from your podcaster wallet to our nodes.

(01:05:13):
And each of us are set up in a split directly, including the podcast app and the podcast index.
Thank you, everybody who supports with a boost. You can do it with Fountain
FM. That's the easiest way to get started.
but there are other paths like Podverse and AlbiHub if you like to do the self-hosted
thing and of course you can use the Podcast Index if you have AlbiHub too so
there's some other options there.
Thank you everyone and thank you to our members.

(01:05:43):
Now the show actually begins, because this week our pick is a doozy.
It's a brand new project that's only about a week old.
It was started on kind of a lark, I believe by Jack Dorsey, and now has been
picked up by a rather large community.
The project is called BitChat, and it is Bluetooth Mesh Chat with IRC vibes.

(01:06:05):
It is decentralized peer-to-peer messaging that works over a Bluetooth mesh.
No internet required, no servers, no phone numbers. You can do private chats.
You can do public group chats.
It is very early days, so we're not necessarily counting on the security,
but it has a lot of great security ideas implemented.
We'll want to see how it goes. It is an open source project.

(01:06:27):
But what's really insanely neat is that it is decentralized mesh networking,
not like MeshTastic, which is using low-powered radio.
This is just using the Bluetooth in your phone. So it's a much shorter range
thing, but you could think of, you know, at an office, at a group event,
during a protest, you could actually really see different scenarios where being
able to stand up mesh group chat very quickly could be really valuable.

(01:06:51):
I'm blanking. In a past episode, I tried one I used on the plane.
It might have used the Wi-Fi, maybe use Bluetooth, but it was not nearly as slick or easy.
Besides the permissions step, which is standard these days, it really just worked pretty quick.
I mean, you and I popped right on. Yeah, you updated yours, and then we were

(01:07:11):
able to immediately chat.
And it has such, such awesome IRC vibes. And I've really been missing IRC recently.
You do a slash, and you get this beautiful menu of all your different commands.
You can set up a new channel. You can hug someone. You can slap someone with a trout.
A lot of the classic IRC stuff.
it has emergency wipe too you can triple tap and clear all the data,

(01:07:35):
ios and android versions exist they
use lz4 for message compression it has
really good adaptive battery modes and they really worked on optimizing the
networking it's really fun and early but each device acts as both a client and
a peripheral that does the transmission it does automatic peer discovery it

(01:07:55):
has stored and forward offline message delivery so wes was able to DM me a message
even when I had the client closed...
I mean, it's pretty neat.
That's so great.
Right? And, again, needs to be tested early days.
Don't go using it to exchange a whole bunch of your most precious secrets.
But they are trying to use really good encryption technology here, too.
Okay. And, yeah, it does look like, what, there's Android and iOS?

(01:08:18):
It was iOS first, and then they ported it to Android, I think?
So there's, like, a Swift and then Kotlin version.
Yeah, it's basically two different projects, right?
At least the Android one is MIT-licensed. I have not checked on the iOS one.
I saw TechCrunch did a really lazy takedown. The headline is,
quote, Jack Dorsey says his, quote, secure new BitChat app has not been tested for security.

(01:08:38):
I mean, like, the dude wrote this on a Saturday and a Sunday.
And then, like, on a Tuesday, they're giving him a hard time for not having it audited.
Right? And Dorsey's created a white paper, which is what people are following
to do the different implementations.
So it's really lazy and cheap to take a shot at something that's, like, seven days old.
But that's the entire pitch of the TechCrunch article. So, you know,

(01:09:01):
this is, of course, going to get parroted by anybody that...
Hey, you clicked.
Well, it got me, yeah. It's just, it's very clicky.
IOS version is under the unlicense.
Okay.
Yeah, which is basically straight to public domain, as far as I can see,
which is really interesting.
Yeah. I think it's a little harder to get the iOS version because you have to
go through TestFlight at the moment and there's a 10,000 max.

(01:09:24):
Right.
But, you know, with Android, and especially if you use Obtainium,
you just point at the GitHub repo and it just installs the APK.
It's really nice. It's so fast, and the battery life seems to be very minimal
because I've had it installed for days and I haven't noticed any major impact.
I could also see this being just a great chat with the wife and kids around the house.

(01:09:44):
Like, why do we need to be doing anything over the Internet?
And one of the things that I've always been looking for is some sort of offline...
We could be camping without internet chat. And I was thinking NextCloud chat,
but that's a lot of, it's a lot of overhead I have to run.
These are just really simple apps that we could just use for just quick messaging
back and forth when like one of us is inside and one of us is outside or we

(01:10:07):
just want to send each other notes.
Yeah, and it's just so easy to get going. And the idea of a little Bluetooth
app, Bluetooth mesh is so cool.
I did look, Briar was the other app that I'd tried previously.
It does Wi-Fi and Bluetooth, but it's a little more involved.
This would be really fun for us to use at conferences whenever we meet up with
the JB community, just so we can all have a shared experience.

(01:10:29):
And I know we did that with MeshTastic. At scale.
Well, we did that at scale.
Yeah, we did it a couple times. It was fun, but not everybody has a MeshTastic
device, etc. But everybody has a Bluetooth on their phone.
That's the thing.
Where in the giant expo hall are you guys?
Well, and you could have channels, so you could see how that could be really useful.
And the thing that's nice is, A, no account or phone numbers required.

(01:10:51):
so people don't have to go through a sign-up process that's always finicky when
you're trying to do anything like that at scale. And there's no server.
So you don't have to worry about setting up a server that people can get to
from the event Wi-Fi. It's none of that. It's just a Bluetooth mesh.
But the other thing that I think is fun and awesome to see is that this was
essentially a vibe-coded idea with a white paper,

(01:11:13):
and it's gone from that to a community building out both iOS and Android versions
and creating something real here.
in like seven days. So it's a perfect example of how sometimes open source software
can just catch fire and really take off. And maybe it doesn't go anywhere.
Maybe it ends up having security flaws. But in the meantime,
it's a lot of fun to check it out. And like Wes says, it's not the distribution.

(01:11:37):
It's the fun you have composing the distribution along the way.
And troubleshooting strange, esoteric.
Anyways, link to BitChat, both for iOS and Android, and some of the details
about how it works in their encryption, which looks like they're really taking
a good shot at making this a private communications tool as well.
But you just kind of want to wait and see how that works out in actuality.

(01:12:03):
All right, that's it for us. Don't forget, we'd love to hear your experiences
with immutable distributions. If you've tried it, if you're willing to try it,
if you bounced off of it, any of that, boost in and let us know.
It's one way you can support episode 624 and get your message read on the show.
Anything above 2,000 sats does get read.
Now, if you want more shows, we're wrapping up and you're wishing we weren't
done yet, you're probably a great candidate for the bootleg.

(01:12:24):
That's at linuxunplugged.com slash membership.
Right now, the bootleg's clocking in at like two hours and 35 minutes with lots
of extra content in there.
And then last but not least, we should mention, Wes, we should try to mention
every episode that we have both transcripts and chapters.
That's right. So you can find what we've talked about. I just,

(01:12:46):
I got an email from an individual that says, gosh, I wish you'd put the names
of everything you talk about in the show notes.
Well, one next best thing to that could be you check the transcript.
We try to link everything we can at linuxunplugged.com slash 623 or whatever
the episode number is. but the next best bet would be the transcript.
Yep.
And those chapters mean you can go right to the segment. So if you want to replay

(01:13:07):
the bit chat segment, you sure can.
But then the real power move is to join us live. Make it a Linux Tuesday on
a Sunday. Come on over to jblive.tv.
That's right. JBLive.tv at 10 a.m. Pacific, 1 p.m.
Eastern. That's where our virtual lug gets together in our mumble room,

(01:13:27):
our matrix chat chats, along with us and helps us title the show.
It just gives that live vibe. You know? Just that live vibe that we like so much.
Yeah, if you come, you can hang in a certain Chris Lasta as a little DJ and
sometimes you had some great picks this morning, let me say.
Oh, thank you. Thank you. Yeah, you got to get a little early for that.
All right, well, that's everything we talked about. Links at linuxunplugged.com.

(01:13:48):
So that's everything right there. That's it.
Also, you can check out notes.jupiterbroadcasting.com if you want to do some backlog searching.
notes.jupiterbroadcasting.com. Thanks so much for joining us on this week's
episode. We'll see you next Tuesday, as in Sunday!

All Episodes

Episode Transcript

Popular Podcasts

Stuff You Should Know

The Joe Rogan Experience

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}623: 50 Days of Blue

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}Stuff You Should Know

The Joe Rogan Experience

Dateline NBC

All Episodes

623: 50 Days of Blue

Stuff You Should Know