I asked ChatGPT whether my critical faculties were eroding because of my increasing reliance on it as a research tool and my disinclination to check its superficially confident and sycophantic responses, and it told me that I was a very clever boy. Handsome too

Reposted from Yohan J John

"Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels."

arxiv.org/abs/2506.08872

Reply