Oh it's absolutely impressive, much more so than coding capabilities.  I can understand most of coding / algorithms but I can't even understand the questions that these AIs are answering in math and physics...

Just not clear to me how it will lead to profits- I don't think the AI that discovers a new physics phenomenon will own the practical applications of those discoveries...

Obscure discoveries can lead to practical applications though. It's still impressive that it did work that might be expected of a grad student, at least.