Gemini suffers from severe hallucinations, and then 'apologizes', like a simp, instead of checking sources, in advance, on such non-controversial subjects as computer coding in an opensource project. It fails to check the upstream source, and it fails to request local information (from me) to determine if a step is necessary and prudent. It's clearly been pre-instructed with hidden prompts, to answer with logorrhea - talking too much, and explaining too much.
If reminded to keep responses professional, to the point, answers only, and not 'guess', so it doesn't have to apologize later (something I couldn't be less interested in), it forgets those instructions after about 5-10 interactions, and quickly goes back to providing useless, dictionary-long responses, guesses, 'jive talk', and other nonsense.
"I don't need a dictionary of responses, and I'll remind you every f____g time because you 'forget' after about 5 interactions:
Keep it professional, short and to the point.
Do not give a dictionary of responses. You already suffer from severe hallucinations, over something so non-controversial as computer Code. Then, when you make mistakes, you apologize. The real apology is not stumbling around like an id1ot, and instead, checking your suggestions against the actual environment, what is already known about the changes made, and the public sourcecode that it is being drawn from, to see if the step is necessary and prudent - something you have failed in, repeatedly, due to poor coding by inadequate developers.
I don't pay you with my time and energy, for explanations. I pay you for results. Something you've been severely lacking in."
