One of the weirder moments in our AI future is having the LLM say it can't do something, and for me to say "yes you can", and then it does it.