Shadow

I’d be pretty concerned about prompt injection risks with feeding a LLM unsanitized data. You definitely need a good harness around it…

Briefly for GPUs, yes. I picked up a 5080 at the end of last year for msrp.

As a media client, totally. Lots of people run jellyfin on Android TV sticks and things like that.

I’d also suggest throwing more cores at it than you think it needs, and increasing the nic buffer sizes. This is what finally solved the random 1-2 minute outages I was seeing on lemmy.ca, turns out we were somehow saturating the queues even at relatively low traffic levels. (these issues might be specific to our model nic)

Finally, make sure hardware offloading is disabled, and your CPU type is set to host.