

It’s the most infuriating thing when I am experimenting with local AI. Although the absolute worst offender is Gemini (not local). Every fucking time, “That’s a great question to ask bla bla bla” or “you’ve absolutely hit the nail on the head” and the many others. If an AI does that a few too many times I just use a different one. (Granted you can tune the answers with a system prompt).


Even though I trust Signal and the FBI has raided them multiple times without success. This is still a centralised service that goes through American servers therefore is automatically a threat to sovereignty.
But again I cannot recommend signal enough they have a brilliant track record. And with good opsec you have little to worry about.