Big Technology Podcast Google DeepMind CEO Demis Hassabis: AI's Next Breakthroughs, AGI Timeline, Google's AI Glasses Bet
441 snips
Jan 21, 2026 
Demis Hassabis, CEO of Google DeepMind, is a pioneering AI researcher known for leading breakthroughs like AlphaGo. In this conversation, he discusses the current state of AI and the potential for future breakthroughs, including a five to ten-year timeline for achieving true AGI. Hassabis emphasizes the importance of continual learning, memory, and hybrid systems in advancing AI. He also shares insights on Google’s ambitious AI glasses project and the ethical considerations surrounding AI and advertising. It’s a thought-provoking look at the future of intelligence.
AI Snips
Chapters
Transcript
Episode notes
Room Left In Current Models
- Existing architectures and data still have significant headroom through engineering and tweaks.
- DeepMind plans to pursue both scaling foundation models and blue-sky new architectures in parallel.
Key Missing Capabilities Before AGI
- Continual learning, better memory, and efficient long contexts are likely required for AGI.
- Large foundation models will remain a key component even if additional breakthroughs are needed.
AlphaZero As A Continual Learning Example
- DeepMind has achieved continual learning in narrow domains like AlphaZero and AlphaGo Zero.
- Those methods worked in games but scaling them to messy real-world problems remains unsolved.