I was thinking this morning about how AI is pretty good at starting research projects, but not very good (at least for now) at finishing them.
As a result, it’s become much easier for scholars to accumulate an unmanageable backlog of half-baked, AI-assisted projects... 1/n