You can actually just ask me directly for a brief statement of a thing I think, if that's not cheating? I herewith state: The danger from sufficiently advanced AI is that it is very good at predicting the world, and very good at making choices that steer the world toward some

→ View original post on X — @esyudkowsky,