Comment by toddmorey

I worry more about the human behavior predictions than the artificial intelligence predictions:

"OpenBrain’s alignment team26 is careful enough to wonder whether these victories are deep or shallow. Does the fully-trained model have some kind of robust commitment to always being honest?"

This is a capitalist arms race. No one will move carefully.