People who are used to doing actual research and who understand the necessity of things like a peer review process understand how dangerous it is to have something that’s straight up confidently wrong 20% of the time, and yet have people who treat it like the ultimate authority on things. These people have likely seen AI be stunningly stupid in fields they know about and can extrapolate from there that it’s probably stunningly stupid about most things.
People who are used to doing actual research and who understand the necessity of things like a peer review process understand how dangerous it is to have something that’s straight up confidently wrong 20% of the time, and yet have people who treat it like the ultimate authority on things. These people have likely seen AI be stunningly stupid in fields they know about and can extrapolate from there that it’s probably stunningly stupid about most things.