Funny coincidence, this is also a great way to get ChatGPT to produce correct solutions when it produces incorrect stuff on its first try.