Really? My experience with GPT is more the description I add the better the results. I presume this is because it has a longer prompt to attend upon, I think the whole idea of focusing on keywords/ concise sentences is a very “search engine” paradigm and language models do better the more you describe your question

Details are fine. But think about how this thing works. It does not think about your request. It comes up with the most probable answer. There is some tuning to imply self reflection, but that’s mostly fake. When you say “Do X, but if you can’t do X, do Y”, you may very well encourage the model to do Y without any qualitative assessment over whether it could actually do X.

Same for questions where you ask “is X good or bad? And why?”. It answers good or bad before it comes up with the reasons. That’s very plausibly ok, but it’s different from how people imagine it works and thinks.