Lebowski Theorem: "No superintelligent AI is going to bother with a task that is harder than hacking its reward function." #AGI #AGIFirst

→ View original post on X — @montreal_ai