AI Safety Fundamentals - The Easy Goal Inference Problem Is Still Hard
Sign in to continue reading, translating and more.