Right? The smartest dogs can mimick toddlers intelligence, but we aren't looking to hire dogs in to anything. An AI might know it's scope decently, but is still flush with mistakes. I don't want those mistakes near anything of critical importance, just like I want intelligence making decisions.