
Intelligence Squared Demis Hassabis and Sebastian Mallaby on The Quest for Artificial General Intelligence (Part One)
97 snips
Apr 19, 2026 Sebastian Mallaby, journalist and author who chronicled DeepMind, and Demis Hassabis, DeepMind co-founder and Nobel-winning AI scientist. They discuss Hassabis’s lifelong drive to build powerful AI. They explore DeepMind’s London roots, the Google merger and Gemini, ChatGPT’s shock to the field, and the ‘infinity machine’ idea linking models to huge problem-solving.
AI Snips
Chapters
Books
Transcript
Episode notes
Infinity Machine Solves Huge Search Spaces
- Sebastian calls DeepMind an 'infinity machine' for solving problems with astronomically large search spaces.
- AlphaGo and AlphaFold exemplify this by navigating combinatorial explosions in Go positions and protein conformations respectively.
Plan Back From A Long Term North Star
- Demis recommends planning from a long-term North Star and breaking goals into subgoals, a strategy he credits to chess training.
- He argues this backward planning helped him pursue AGI across decades despite setbacks.
Childhood Motto Fueled Near Exhaustion Effort
- Demis took his father's advice “do your best” to an extreme, interpreting it as pushing to near exhaustion to know he'd done his best.
- That literal, intense drive shaped his approach to competition and research from childhood onward.




