Basically the local models don’t (and can’t) contain the full knowledge of the universe.

BUT they can call tools pretty well and if you give the harness the capability to search Wikipedia for example, it becomes a lot smarter

Qwen3.5 and Gemma4 are the best ones for tool calling that don’t need massive amounts of memory

Nothing you can run with affordable hardware. The SOTA stuff requires hundreds of gigabytes of memory - and not RAM, GPU memory.

But you can try with stuff like gpt-oss or qwen coder

I do run Jellyfin too, it’s pretty handy at getting metadata for stuff locally so I have something outside of Plex’s internal databases.

And IIRC I have used Jellyfin in the past to watch some more esoteric stuff that Plex (and its official clients) couldn’t play.

But there’s no way in hell I’m opening it to the outside world.

I paid 79€ or something for Plex Lifetime well over a decade ago and it’s by far the best bang for buck I’ve ever got.

It’s also dead simple to share it with friends and acquaintances safely. I know how to manage VPN tunnels and all that shit. I’m not setting that up for my aunt who lives 4 hours away. We did manage to get Plex running on her TV over the phone though.

It’s literally a Raspberry pi 3B+ and a USB hard drive in a plastic storage box at my parents house 😅

A second offsite NAS (my old one) with the same capacity for the larger files

Backblaze B2 and a Hezner storage box for Really Important stuff.

My point was that the frequency of updates doesn’t correlate with quality at all

The huntarr project released a new docker image 3 times a day…

I’m not self hosting email but my rule is that no email gets to be in the inbox except for VERY rare exceptions

When an email lands in my inbox, I immediately make a rule that labels it correctly and moves it the fuck away from my inbox.

This way I can have notifications on for inbox emails and they’ll either be important or a new sender whose next email will end up labeled and NOT in my inbox

Big +1 for second hand corporate mini PCs

They’re cheaper and better in every way than the Pi

Only get the Pi if you need a specific HAT or GPIO. And even then get a zero.

What’s the difference between this and say FreshRSS?

Azure was down? Did anyone notice? 😆

Even if you have a dynamic IP it’s trivial to set up automatic DNS updates with a good provider that has an API to do it.

It’s all explained here: https://valetudo.cloud/pages/general/why-not-valetudo.html

The dev has a specific vision and that’s it. If you don’t like it you can use something else.

Just like the dev of Calibre/Kitty. He does things a certain way because that’s the way he likes it. It’s a stupid and shit way, but Calibre has no real competition so I use it 🙂

I tried it but Copyparty worked better, it has a massive community suddenly and tons of cool features that mostly stay out of the way unless you enable them

This is the only correct answer

This will get downvoted to oblivion because this is Lemmy:

Get a Mac Mini. Any M-series model with 32GB of memory will run local models at decent speeds and will be cheaper than just a 5xxx series GPU

And it’ll fit your cool rack 😀

I bought a Venta LW25 and couldn’t be happier. Simple and functional, good old German engineering

Yea, it’s been broken for weeks. You can’t search for new artists at all.

It got sorta fixed a while ago and now I can search for artists, but adding them doesn’t work. Their metadata server crapped out months ago and they’ve been fixing it ever since, should be done soonish. Or not.