Strategy-Stealing Argument Against AI Dealmaking — LessWrong