Sigh. I'm still so bothered by that "chatgpt rots your brain" paper BECAUSE it correctly positions some of the compelling questions we have about LLM usage but then does us all a disservice in its approach, pushy interpretation, and inappropriate or misleading framing of brain evidence and measurement of human problem-solving, imo
I am asking myself: what studies would *I* design to evaluate people's usage of LLMs in keeping with learning science and without neurohype?