Uploaded November 2026 | Updated November 2026, 3 hours ago
What happens when AI learns to improve itself with no human in the loop? Google DeepMind CEO Demis Hassabis tells Reid Hoffman why a “closed” self-improvement loop could be the fastest and most dangerous road to AGI, and why the result might be an alien intelligence we can’t understand.
From the Masters of Scale Summit 2026.
What happens when AI learns to improve itself with no human in the loop? Google DeepMind CEO Demis Hassabis tells Reid Hoffman why a “closed” self-improvement loop could be the fastest and most dangerous road to AGI, and why the result might be an alien intelligence we can’t understand.
From the Masters of Scale Summit 2026.










