Have any LLMs solved any of the big (or even lesser known) unanswered problems in math, physics, computer science?
It may appear that they are solving novel problems but given the size of their training set they have probably seen them. There are very few questions a person can come up with that haven't already been asked and answered somewhere.