This was a bad paper and an embarrassment to RAND. They argue that humans have secured a bunch of specific attack avenues, therefore, something vastly smarter than humanity can be secured along every dimension.

→ View original post on X — @esyudkowsky,