Each of those has fixed rules, and each took lots of effort from human researchers to hand-craft and hand-train the models. Not that it isn't impressive, but it's still fundamentally a different thing than general intelligence. I don't think it's just a matter of doing more of the same.
