Some AI models get more accurate at maths if you ask them to respond as if they are a Star Trek character, ML engineers say::Researchers asking a chatbot to optimize its own prompts found it was best at solving grade-school math when acting like it was on Star Trek.
Are you sure?!