Astro Teller on How to Handle Failure (With an Infinite Budget)

, , Comments Off on Astro Teller on How to Handle Failure (With an Infinite Budget)

Failure is always disappointing, but two things can help take the sting away: Dispassion and money. Lots of money. As the head of X, Alphabet’s so-called moonshot factory, Astro Teller has plenty of both. Funded by parent company Alphabet to the tune of billions each year, X is dedicated to ambitious and sometimes wacky research and development projects. And, according to Teller, researchers working on these projects—as exciting as they may be—have to be disciplined about their enthusiasm for a project.

“Running the experiments is itself hard,” Teller said onstage Friday during the WIRED25 Festival in San Francisco. “But after the experiment, I need you to—however passionate you were while working on the project—we now need to step back and dispassionately assess. We just got some data from the world: What does this mean? Should we keep doing this project?” If a X-oogler loves their solution too much, Teller says, that blinds them to the true solution.

That sounds like an easy enough approach, but it has to be coupled that special X touch: A long time horizon,and a willingness to give employees the space and time to pound away at hard ideas and truly learn. Add those together, and you have the secret to X’s success, says Teller.

Well, decidedly mixed success. Teller told the Wall Street Journal last year that the Alphabet arm has about a failure rate of about fifty-fifty. Foghorn, an attempt to turn seawater into a carbon-neutral fuel, died after two years of work in January 2016, after X-ers decided the idea wasn’t economically feasible. But some X projects do “graduate” and become Alphabet spin-off companies, like self-driving vehicle company Waymo (currently valued at $105 billion)the drone company Wing; and Loon, an attempt to use huge, floating balloons to provide internet connectivity to all.

One of X’s most exciting in-the-works bets couples AI with agricultural production—it’s making a goofy little blue plant buggy that roams fields, using data and machine learning to evaluate specific crops and help farmers manage their growth.

On Friday, Teller cited one moment of dispassion during the buggy’s development process: X engineers needed some way to move aside crop leaves so they could get a good look at a fruit’s development. Maybe they could just attach a leaf blower to the underside? Nope. “A leaf blower kicks up a huge amount of dust and pisses off the farmers,” Teller said. That rational assessment was an example of X’s typical post-project dispassion, writ small. “The dispassion is happening on a weekly basis.” Good luck, little plant buggy.