About this episode

Roberta Raileanu discusses the ingredients for achieving superhuman scientific discovery through AI at RAAIS 2026.

At RAAIS 2026, Google DeepMind's Roberta Raileanu lays out a recipe for superhuman scientific discovery: AI systems that make groundbreaking discoveries across domains faster than people can. She walks through three ingredients - reinforcement learning to discover solutions where progress can be measured, open-ended divergent search to find new problems rather than climb known ones, and meta-learning to speed up discovery on problems no one has posed yet. The through-line: we can search for anything we can measure, but we still cannot measure what makes a discovery good. The bottleneck isn't the search. It's the signal. Chapters: 00:00 - Introduction 00:51 - Defining superhuman scientific discovery 01:40 - The state of play: real progress, real plateau 06:58 - Ingredient one: discovery as reinforcement learning (Move 37, MLGym) 12:51 - Ingredient two: open-ended search and why greatness cannot be planned 18:30 - Rainbow Teaming: quality-diversity in practice 21:07 - Ingredient three: meta-learning the process of discovery (DiscoBench) 25:06 - The recipe, and the missing signal

People in this episode

Host: Nathan Benaich

Guest: Roberta Raileanu

Topics covered

• superhuman scientific discovery
• reinforcement learning
• open-ended search
• meta-learning
• AI systems
• groundbreaking discoveries

Keywords

• AI
• scientific discovery
• reinforcement learning
• open-ended search
• meta-learning
• groundbreaking discoveries
• quality-diversity

Mentioned in this episode

Organizations: Google DeepMind

Books & works: DiscoBench, MLGym

More episodes of Air Street Press

• The case for intelligence beyond language · July 21, 2026 · 7 min
• Europe’s air defense gap · July 16, 2026 · 7 min
• From driving the world to dreaming it · July 14, 2026 · 8 min
• Turning compute into intelligence · July 9, 2026 · 7 min
• Accelerating science and medicine with collaborative agents · July 7, 2026 · 8 min
• Compute scarcity is an engineering problem · June 30, 2026 · 8 min

Explore listener stats, chart rankings, contacts and more on the Air Street Press podcast page.