https://ollama.ai/, this is what I’ve been using for over a year now, new models come out regularly and you just “ollama pull <model ID>” and then it’s available to run locally. Then you can use docker to run https://www.openwebui.com/ locally, giving it a ChatGPT-style interface (but even better and more configurable and you can run prompts against any number of models you select at once.)

All free and available to everyone.

I’ve exclusively used firefox to watch youtube on Arch and Ubuntu for years, never had a problem so far for what it’s worth. I keep a laptop in the livingroom with Arch specifically to have adblocking and piping the video out to the TV. The youtube apps are terrible on the Roku last I remember, haven’t tried it in forever but I think the main reason was I didn’t want to see ads anymore.