The AI discourse is broken right now because people are failing to keep separate what are very separate questions.

(1) Is AI useful—economically/occupationally?

(2) Can AI think?

(3) Is AI an industrial or financial bubble?

(4) In the near future, is AI fundamentally a force for good or bad?

I have seen so many people say that AI can’t think (2) and therefore it’s not useful (1). Or that AI’s unit economics are shaky (3) and therefore it can’t be a consequential technology in the near future (4). These arguments collapse separate questions into simple efforts to declare the technology good/useful v. bad/useless.

It would be so much better to have specific debates rather than to be constantly conflating separate questions in a way that makes everybody dumber about the changes to come or not come.