I run podman containers on my bazzite machines, basically you convert a docker-compose file to a .container file, here’s a bunch of examples, nextcloud is there, drop it in ~/.config/containers and run systemctl daemon-reload and it’s now a systemd unit that you start stop etc like any other. Updates are with podman autoupdate.

You can use podlet to convert docker-compose files (90% it works, otherwise it gets you 90% of the way there). It’s basically the fedora (/redhat) way to run containers.

I have no idea where you got it not being recommended (but adding to the main image sure is discouraged), and it’s certainly better than adding a vm for containers, which pretty much defeats the purpose of containers (to run using your main kernel, but contained).

I’ve been running my arr stack (with gluetun in a pod) etc this way for years now, very trouble free. Here’s a immich example.

It’s a bit of a learning curve, but it pays off.

Given how long they’re viable for (obsolete GPUs in 2 years or less, high likelihood of needing significant retooling for the next gen) you can’t mothball them. Given how few exist in any way other than on paper there will be little to bulldoze.

On the plus side counties are scrambling to decouple their economies from the US and hopefully it’s just the trickle before the flood, so when the US economy tanks from the AI bubble pop it’ll have less effect on the rest of the world.

Just repeating what I read elsewhere, the reporting has been somewhat muddy.

Nope, not top, but apparently not awful (seems to be her husband coding). Some interesting things happening currently with memory systems for local AI which seem to truly enhance (make smarter) smaller thinking models, meaning you can use more, and more coherent context.

As with most of this stuff I’m waiting for a bit more maturity before looking deeply at it, but there’s definitely some excitement at the moment, and it has the potential to make models that fit on say a 16GB video card capable of many more use cases than previously.

Legal thieves, and their kin.

Can’t imagine how boring the world would be without stuff I don’t understand.

Don’t forget thinking…

Been using an quadlet podman arr stack for a year or two, pretty damn bulletproof once set up, easier to read, rootless, SELinux enabled, systemd controlled, update with podman auto-update. Worth the time to learn.

podlet can help you hit the ground running. It can create Quadlet files out of Podman commands or even (Docker) Compose files. 90% of the time it works every time ;}, but even the oopses get you most of the way there.

My arr stack is set up in a pod which means they all have their own gluetun network and come up as one, but you can just use Network=container:gluetun in container files.

Yeah, feel the same, curated RSS feeds also contribute significantly.

The amount of BS flying daily in the ‘media’ and with AI slop makes every day April Fool’s anyway. A constructed environment means the BS detector doesn’t have to be on 24/7, which becomes tiresome.

more people use their servers exclusively for personal entertainment than I expected.

Uh-huh, think of it like jigsaw puzzles…

That said, I prioritize ease of maintenance and simplicity, still wouldn’t expect my family to pick it up in any reasonable amount of time, nor have the motivation, more’s the pity.

I’ve moved to podman (quadlet) containers mostly, easy to read and edit, secure (mostly userspace), systemctl integration, autoupdate. I’ve done my distrohopping, fedora (in my case bazzite immutable) isn’t going anywhere, does everything I need. I run fairly lean, but have a bunch of stuff that can be spun up at a whim that I don’t use daily. It’s entertaining without being a burden, and useful stuff just happens.

Honestly, ssh and btop cover most of my monitoring needs, serious stuff gets a notify-send to my laptop. I’ve tried the web gui stuff and I don’t look at it enough to justify it, I’m not a sysad monitoring hundreds of computers, it’s just a hobby.

Absolutely correct, and likely the eventual outcome (tool not replacement), because LLMs are not, and very likely never will even be close to AGI (let alone techbro dreams of ASI). Code smell is a thing, and LLMs reek. It is possible to learn from codegen, but that sure as rut is not it’s purpose. mostly an excuse to drive down wages to make the rich richer.

Somewhere along the line we lost the virtue of do one thing well. Do it right, move on, do something else right. If you need complexity plug some right things together.

Now is the time of jack of all trades, master of none, stuff it into a phone whether appropriate or not. Get the job done as cheap as possible with the same generic crap. Do you want your airplane pilot using only a touchscreen ? Why should it be different for a spaceship, a car, an audio engineer ?

A steering wheel, once learned, is a wonderfully flexible interface that meshes with human intuition, throw rock left, rock go left, we have brain circuits devoted to this stuff, it works. Tactile interfaces can and should speak to our evolutionary core. Let’s take a step back, economic imperative be damned.

Written from 500g rugged 1cm+ thick brick

With a battery that lasts a week, horses for courses…

Pinchflat

Thanks, looks like it could work, bit more than I want, and it’ll need integrating, but it is what it is. I’ll see if it’s worth the pain to transition.

Of course there’s this…

MonkCanatella @sh.itjust.works English

I just have a somewhat annoying yt dlp script that runs on a schedule on my computer that downloads all my subscriptions. I tried everything else including pinchflat and it wasn’t really much of an improvement tbh.

TheLeadenSea
@sh.itjust.works
English

Could you share that script please? I’d like to set up something similar
    MonkCanatella
    @sh.itjust.works
    English

    Give me a bit, I’ll set up a gist with the details!
        TheLeadenSea
        @sh.itjust.works
        English

        Thanks!

followed by crickets. Such is life.

So, cool and all, but I’ve been using freshrss with libredirect and freetube for this for ages, and it’s great when freetube works, which with the cat and mouse with youtube is less often than I’d like (kudos to freetube, great project).

Honestly what I’d really like is something to download rss yt vid to a cache with yt-dlp (which breaks much less), grab the sponsorblock and pipe to something light and distraction free like mpv or vlc. Clear the cache after a day or whatever.

Anyone seen anything like this ?

Nice, if android even cares though. Dunno, what’s the state of android AV these days?

And iOS I suppose, the glasses aren’t attached to a laptop most likely. And does the app auto-open QR codes? Enquiring minds want to know, because this is a solid idea if it can be made to work.

For say a keypass db you don’t need even that, Just sshd gets you rsync on your computer with cron or systemd timer / service… Personally I just use an old version of Syncthing-Fork though, security implications for local network are minimal.

Sounds good.

Sensible, limiting scope and knowing your limits are wisdoms all too many lack. Having something that fills the text/sound/vidya needs and is easy to spin up will find uses, doesn’t need to do everything.

To the end of being easy to spin up, which is likely to attract other developers in time if that’s something you’d like, consider wrapping it up in a docker container. It’s not that hard, basically follow your own instructions in a special docker build format.

Some things to think about.

Even ZFS now let’s you add a new drive to an array, and the sweet spot for $/TB is ~16-20TB at the moment, so maybe think about 2 or 3x16TB and add more later (also less power).

Consider manufacturer recertified (not refurbished) server drives from serverpartdeals.com or your local equivalent, after all RAID is there to let you survive a disk failure, it’s treated me well, and lets you avoid SMR drives.

You can mix drives of different sizes if you use Unraid or roll your own with mergerfs+snapraid (+OpenMediaVault perhaps). I do the latter, it’s a bit of a setup, but has the advantage that drives are just drives and you can use the working ones while rebuilding the array and you can recover accidental deletions (for a while), which brings me to ‘RAID is not a backup’.

For true data safety you should have an offline backup (i.e. drives that live disconnected from your computer except during backup, safe from lightning, accidental deletion etc.) and eventually an offsite copy.

Personally I think the AI bought all our drives from WD is likely BS (seems lightly supported) to goose their product prices, so hopefully it’ll blow over, but prices seldom go down, inflation catches up. Sigh.