He tries 1,000 ideas a decade & kills nearly all—that's how X built Waymo & Google Brain. | Astro Teller, CEO of Moonshots at X (Alphabet)
Summary
Astro Teller, leader of Alphabet’s X Moonshot Factory, discusses systematizing radical innovation, emphasizing rapid risk reduction over progress. He explains how X tests 200 ideas yearly by identifying the riskiest "monkey" first and using pre-set kill criteria to fail fast. This approach prioritizes learning and intellectual honesty, allowing teams to pivot quickly without existential attachment to ensure only viable moonshots like Waymo graduate.
Topics discussed
Introduction and show premise
Astro Teller's background and X's mission
The X model: Card counting vs. roulette
Psychology of failure and idea filtering
The innovation funnel and team dynamics
Sourcing ideas and avoiding top-down control
Case study: Waymo and self-driving cars
Testable hypotheses and risk reduction
The monkey on a pedestal analogy
Graduation and independent company formation
Kill criteria and project termination
Remote work vs. in-person collaboration
Moonshot market fit vs. product market fit
Final advice on psychology and sharing the show
Listen ad-free on Castria