AI lab

Had a quick squiz at it, and if it meets your needs, I’d just wait, unless you want to get into the lower levels of things (and run linux), llama-swap is just an inference server, you’ll need something like Open WebUI for chat as well etc.

Not sure about AI lab (although I also use podman, prefer Llama-swap), but pretty much everything uses llama-cpp under the hood, which usually takes a day or three to setup for a new architecture. Although I seem to recall them being ready for Qwen3.5 day one due to collaboration.

I find giving it a week or so for the dust to settle (even if it ‘works’, best parameters, quantization bugs etc take a while to shake out) unless there’s a huge motivation.

Also benches are more like guidelines than actual rules, best to do your own on your own use cases.

Crank it up to 11 12, there’s profit to be made.

ETA Another likely scenario:

George: Hey, Ron, you know that maintenance we’ve been putting off because line must go up ?

Ron: Yeah…

George: Wouldn’t it be terrible if something happened and the government had to pay for it to be fixed real fast…

Ron: Heh heh heh.

The best punchline would be if the government nationalized it as well due to negligence of critical infrastructure…

Would also send a great message to the power companies about line maintenance. But I’m (probably) dreaming, I’ll send it out into the aether anyway.

Last time I looked the laws (long time now, but I think I’d have heard) from the Vietman war haven’t been repealed, so lottery by birthdate of male citizens again?

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.

Hmm, haven’t had to fire up Unbound for a while, and Technitium looks interesting (any opinions?) doesn’t effect me yet, but it’s good to have options. The internet finds a way, but ordering VPNs and DNS resolvers is a new thing I think. Wonder how well it will take.

Fair cop, shall look at pipepipe, wait, that’s a phone app, I like my vidja on a larger screen, You do you, I’m good :)

Kudos to you then I guess. Nice offsite backup you’re not paying for, love it. Shit, getting payed for, brilliant !

Solid. I seem to have a ~90% overlap, copying both this and the movies to my obsidian, shall do some diff / grepping later… There’s gems in there, thanks.

Sympathy for the no adblocker (moron?, but with taste), must burn.

freetube

90% of the time it works every time (OK, maybe 80), but when it does it’s great. I just let my rss queue backlog until it does. (Not on them, YT is actively rugpulling coz they hate them, which makes me like them more, shame about the electron though). Grayjay is closed, and likely questionable, but a useful backup.

Yup, long as it’s copyleft (GPL) open source, I don’t care if it’s microslop paying.

That said, watch out for a new wave of EEE (embrace, extend, extinguish) using unmaintainable AI code, and be ready to fork.

Valve, not so much.

These things are for the nasty little surprises, more relevant for the proprietaries, but not useless on linux. They’re for the unknown unknowns, and that’s a good thing. The next supply chain vuln might bite you, and opensnitch might let you know.

Or Obtainium.

Seeing as you have a backup of your contacts (don’t you??), why not just delete it…

Don’t forget thinking…

More like 89.91… they dream of four nines.

Yup, want privacy, host your own models. Perplexica Vane + ollama (or Lemonade, better for AMD) is pretty damn easy to stand up and works well I find (never used perplexity) and you get references to real web pages to double check for hallucinations.

Nice thing about deep research is you can ask a question and just come back later, so latency is less of an issue, even CPU + RAM can do the job.

Given the current state of the internet it’s heading towards a necessity to play Spy vs Spy AI Slop vs AI Slop. Might as well control it yourself.

FWIW I find Qwen 3.5 27B a great model if you can fit it (and I use the perplexity embedding model that they open weighted pplx-embed).

Gonna file this under blaming the individual for systemic failure while things like the DMCA (making reverse engineering interoperability illegal among other things) and life + 50 copyright are in place and enforced worldwide by US trade deals.

Here’s hoping Trump’s crapping on Pax Americana gives a once in a generation (or two) chance to get out from under.

Well, that inspires confidence.

The Apollo 11 Guidance Computer had 4KB RAM and 72KB ROM, but the code was provably correct. How the mighty have fallen.

Didn’t the Apollos use (Hassleblad 70mm 3900x3900) film. This is a downgrade. (but it’s streaming, hur, hur). Hopefully there’s some better cameras as well.