Normally I would agree, especially if it's some Arxiv paper with two dudes talking about mind reading from fmri (you know who you are..). But Meta AI has shown

Normally I would agree, especially if it's some Arxiv paper with two dudes talking about mind reading from fmri (you know who you are..). But Meta AI has shown that they are very much capable of creating serious models and releasing them, so I remain optimistic this isn't a smoke screen.