@cabel A pretty fundamental problem with LLMs — they are unavoidably literal. Absolutely take every word at face value because it’s impossible for them to do anything else.