What generic AI gets wrong with this prompt
“I love traveling and trying new cuisines” — ask a model for a fun fact and you get the statistical center of all fun facts, which is by definition the least fun one. The box rewards surprise, and surprise is the property the model’s training systematically removes.
Ask your AI — “Rate my fact’s frequency: out of a hundred MBA applicants, how many could plausibly write the same sentence? What detail would drop that number to one?”
It sneaks the resume back in
“Fun fact: I once automated a reporting pipeline that saved 200 hours a year” — generic AI converts fun into achievement because impressing is its default register for application text. Booth built this box specifically to escape your professional self. Answering it with work reads as not being able to stop.
Ask your AI — “Is my fact something I’d tell a hiring manager or a dinner table? If it belongs in the interview, what would I actually say at the dinner table?”
It compresses the color out
The specifics are the fun — the exact number, the ridiculous name. Generic AI trims those to save characters, keeping the category and cutting the texture, because summarization is its instinct. Without expertise, it does not know how to guide you to reflect on the meaning behind what you are sharing.
Ask your AI — “Which concrete detail makes my fact mine — a number, a name, a place? If the answer got summarized, what got lost?”
It invents quirk you’ll have to own
Ask generic AI to punch the fact up and it will happily add color that never happened — a competition you didn’t win, a habit you don’t have. Every embellished character is a liability, because this is the box an interviewer brings up to break the ice.
Ask your AI — “Compare my answer to what I originally told you. List every detail you added that I didn’t say.”