Nikita BorisovThis feels SO MUCH like a qual exam with a student who didn't study enough, complete with (1) vague generic answer, (2) trying to stay generic, (3) when pressed, coming up with an obviously incorrect solution on the spot, and (4) steadfastly claiming that your solution works
Mihai Christodorescu@nikita Paper reviewing is about to get super interesting. Frankly as a reviewer I'd love to have a tool that annotates paper sections with an estimate of having been generated by an LLM (". 923 probability of ChatGPT output from prompt 'blah blah blah'").