can a computer be complex enough to comprehend its own complexity?