Hassabis: AI boosts engineer productivity 3-5x, closed-loop self-improvement is the risky phase

Google DeepMind CEO Demis Hassabis outlined two phases of recursive self-improvement (RSI): today AI boosts engineer productivity 3-5x with humans still in the loop, while the next closed-loop phase—where AI improves itself autonomously—is the most dangerous. Original video of his remarks has surfaced.

2026-11-06 ~ 2026-11-06 · 2 related posts

• Hassabis: AI now boosts engineers 3-5x, but closed-loop self-improvement is the risky next phase — haider1 · 2026-11-06
• Original video: Hassabis on RSI phases, from 3-5x engineer boost to closed-loop self-improvement — haider1 · 2026-11-06