Its not that bad but the most time you spend on installation is opting out of services. This takes up like 75% of the install time :D

Joplin works well for me.

Opera one has messengers integrated in the sidebar. Really nice concept but PWA and masto lemmy etc are missing as an option.

I tried quay.io/go-skynet/local-ai but my Server lacks the Cpu instruction set for it.

In short, there have been key signing issues multiple times in the past that. That were the most notable problems from the user side.

Running three instances with KDE that are unproblematic.

Its a collection of models you can download. It acts as a simple gui entry point into the llm world. Great to test different stuff.

Yunohost might be a good point to start? Havent tried it yet but it seems to be a fast and safebway to host.

Hmm on desktop at least not working for me except maybe per gestures?

Installed lemmy and elk.zone as pwa via vivaldi. Lemmy only misses the back button.

Future AIs might be trained manually over centuries by groups of experts.