Big Technology Podcast How Google DeepMind Operates & Experiments — With Lila Ibrahim and James Manyika
159 snips
Feb 18, 2026 
Lila Ibrahim, COO of Google DeepMind, runs operations for ambitious AI research. James Manyika, SVP for Research, Technology, and Society at Google, translates breakthroughs into real-world impact. They discuss DeepMind’s mission-driven, interdisciplinary model. They cover Google Labs’ revival and experimental product paths. They explore AI in education, global weather and flood forecasting, materials discovery, quantum milestones, and training AI in space.
AI Snips
Chapters
Transcript
Episode notes
Keep Structured Time For Exploration
- Preserve 20%-style exploratory time and accept ideas from across the company into Labs.
- Surface promising internal projects in Labs to test, iterate, and find product-market fit quickly.
From Research To Real-World Impact
- Google emphasizes moving breakthroughs from research into real-world impact rapidly.
- Examples include AlphaFold adoption by millions and flood forecasting covering 150 countries.
Ship To Learn, But Red-Team First
- Treat shipping as a learning mechanism: release, observe real users, and iterate with red-teaming.
- Balance boldness with responsibility; use shipping to gather essential feedback.