> Large Language Model usage levels out the Dunning–Kruger effect.

That's basically my title. I think that's the interesting finding in the study.