Podcast Interview: Waymo Founder Sebastian Thrun

Episode Summary

In 2009, Sebastian Thrun had just sold a company to Google when Larry Page came to his cubicle and proposed building a self-driving car company for ordinary people in cities. Sebastian, fresh off winning the DARPA Grand Challenge and building the first self-driving car for the military, told Page flatly that it could not be done and was too dangerous. Page kept coming back for several days, eventually asking Sebastian to simply write down the technical reasons it was impossible so Page could explain the decision to Sergey Brin and Eric Schmidt. Sebastian could not find one. That failure to object became the birth of Project Chauffeur, which later became Waymo.

Sebastian is candid that the same expertise that made him credible almost stopped Waymo from existing, and he carried that lesson into how he built the team. He hired an automotive expert early on who insisted the whole codebase be rewritten to match an existing industry norm, and Sebastian fired him for it. He staffed the project instead with former students and postdocs he already trusted, choosing people for their willingness to learn over their credentials.

On decision-making itself, Sebastian estimates that roughly eighty percent of the decisions made while building something new turn out to be wrong, and that even successful companies would rebuild themselves five times faster with the benefit of hindsight. His answer is a decision system built around ownership rather than command: whoever owns a decision has to bring the right people into the room, reach a consensus, then document and freeze the decision so it cannot be quietly reopened. Once a decision is delegated further down, the original owner becomes an advisor rather than a boss. He credits a conversation with Jeff Bezos, who called himself Amazon’s “chief slowdown officer” because the next seven quarters were already locked in and he could only really influence the eighth quarter and beyond, with reshaping how he thinks about a leader’s actual span of control.

Sebastian also describes two small personal habits he built to fight fear and guilt, which he considers the two emotions most likely to stall innovation. On what he calls “no-grump Tuesday,” he forces himself to say yes to ideas rather than immediately picking them apart. On “no-complaint Thursday,” he treats problems as his own responsibility rather than someone else’s fault, even when they clearly are not. He points to both practices, along with a decades-long habit of rereading Dale Carnegie’s “How to Win Friends and Influence People,” as the actual mechanics behind staying open to ideas that look impossible at first.

What You’ll Learn

Why Sebastian deliberately chooses projects he’s bad at rather than good at, even though it means failing one of his own three tests for meaningful work.

Sebastian compares today’s narrow use of AI inside existing businesses to an early phase of mobile technology, arguing the real transformation comes once AI combines with things like crypto, cybersecurity, and robotics.

A team member once told Sebastian bluntly, “you destroyed my aspirations, I’m going to leave you, I hate you,” and Sebastian’s response was simply to accept it as fair.

Attending a conference at 21 and meeting scientists like Geoff Hinton before they were famous is what convinced Sebastian to become an academic in the first place.

Why Sebastian credits a book by Zynga founder Mark Pincus with the idea that instincts about a big goal are usually right, even when the specific steps to get there are usually wrong.

Leading an organization, Sebastian says, is a completely different kind of engineering than building an aircraft, since every person in it is working for both the company and themselves.

About Sebastian Thrun

Sebastian Thrun is the founder of Waymo, the current market leader in autonomous driving in the United States, which grew out of a Google self-driving car project he helped start in 2009 as Project Chauffeur. He grew up in West Germany and earned his PhD in computer science, drawn to the field because it let him fully understand and build systems in a way other disciplines could not. Before Waymo, Sebastian led the Stanford team that won the DARPA Grand Challenge, building the first self-driving car for the U.S. military. He has spent his career founding and leading teams built around machine learning rather than conventional programming.

Related Episodes

Episode 166: Software Pioneer Mitch Kapor
Mitch describes building Lotus at a time when nearly everyone in computing doubted personal computers would matter, the same kind of expert skepticism Sebastian had to overcome to start Waymo.

Episode 48: Waze CEO Noam Bardin
Noam describes betting on the iPhone over the dominant Blackberry platform with incomplete information, a concrete example of the incomplete-information decisions Sebastian’s ownership system is built to handle.

Episode 346: Techstars Co-Founder and CEO David Cohen
David hires first for motivation over credentials at Techstars, the same principle Sebastian applies staffing Waymo with people willing to learn over people with the right resume.

Episode 213: PEAK6 Co-Founder Jenny Just
Jenny built a trading firm on the principle that most individual trades will be wrong, the same acceptance of a high error rate that shapes how Sebastian’s teams treat their own decisions.

Episode 349: California Pizza Kitchen Co-Founder Rick Rosenfield
Rick built a company culture around firing a hire who violated its values in the first two months, a parallel test of culture to the one Sebastian describes on the original Waymo team.