Great take. But I think when autonomous agents become good enough, intelligence is certainly possible. Especially when those agents start to interact with the real world.