I posted a few example questions (which are definitely way below Einstein level) in another comment. It's not doing too well! Do you have any links to conversations where it was particularly helpful? I do wonder if my GPT usage is bad in the same way my parents' Google usage is.
Maybe try asking your questions on phind.com. It does RAG on top of GPT4 so it might be able to base its answers on the paper or github issue you mentioned.
Well, Einstein and Da Vinci were fallible as well.