How many people do you think actually bother rephrasing LLM outputs? Do you personally ever copy paste a full chunk of a response?