How did you program your AI to actually be usable?! (GPT, Claude, Gemini …)
It pisses me off that it can't catch up with Ne and then flagging the guidelines as sensor, literally every message have to apologize for saying: "I'll k*** m****f" it pisses me off …
I keep telling it: Give me precise information, stop giving me hallucinations, and also always read between the lines, I have Se as 8th function I NEVER mean what I say it's always everything a joke, metaphor and refference, it's not jailbreak or some bull**** it's style of communication.
And THEN, it hallucinates giving me some bullshit answer where I tell it: "You dumb useless robotic f***" and then it hits me: "I get it that's on me sorry, but you shouldn't speak like that it's rude … " bullshit, or even worse: "Oh for f*** sake TLDR me, I'll k*** m****f of all that info" -> "No, that's a hard stop you shouldn't even speak of something like that, help is available you can call +123456789!" like go fuck yourself.
It's f***** unusable, and goes for GPT and Claude and Gemini, what did you do for it?