Before You Start a Project, Do Your Best to Kill It
Astro Teller's innovation tip? Fail fast. Here's how he cultivates and rewards intellectually honest failures, and helps his team get comfortable with the idea.
Hertz Fellow Astro Teller is now the Captain of Moonshots (aka CEO) at X, formerly Google X, where he oversees X's moonshot factory for building magical, audacious ideas that through science and technology can be brought to reality. Additionally, Astro is co-founder and director of Cerebellum Capital, Inc, a hedge fund management firm whose investments are continuously designed, executed, and improved by a software system based on techniques from statistical machine learning.
Previously, Astro was the co-founder and CEO of BodyMedia, Inc, a leading wearable body monitoring company that was sold to Jawbone in 2013. Prior to starting BodyMedia, Dr. Teller was co-founder, chairman, and CEO of Sandbox Advanced Development, an advanced development technology company. Before his tenure as a business executive, Dr. Teller taught at Stanford University and was an engineer and researcher for Phoenix Laser Technologies, Stanford's Center for Integrated Systems, and The Carnegie Group Incorporated.
As a respected scientist and seasoned entrepreneur, Teller has successfully created and grown five companies and holds numerous U.S. patents related to his work in hardware and software technology. Dr. Teller's work in science, literature, art, and business has appeared in international media from the New York Times to CNN to NPR's All Things Considered. Teller regularly gives invited talks for national and international technology, government, and business forums on the subject of the future of intelligent technology.
In addition, Astro is a writer. In 1997 Random House published his first fiction novel, Exegesis, which went on to critical acclaim and publication around the world in many other languages. He has since published a second novel, sold a movie to Paramount, placed many op-ed pieces and journal articles, and his third book, this one a non-fiction work called Sacred Cows about our society's confusions about marriage and divorce, was published in July of 2014.
For his undergrad work, Dr. Teller received a BS in computer science from Stanford University, and continued at Stanford to obtain an MS in symbolic and heuristic computation. Afterwards, Teller matriculated into Carnegie Mellon University, with the support of the Fannie and John Hertz Foundation Fellowship to pursue a PhD in artificial intelligence.
Astro Teller: Failure, seen properly, is just a recognition of fast learning. So I have training wheels, which I try to put on this concept within X, to try to help everyone get comfortable with the idea. Here's an example. You're working on a project. It either has an Achilles’ heel or it doesn't. Would we like to discover that Achilles’ heel, if it exists, now or later? If you say later you've just signed up to be intellectually dishonest. You're not really going to sign up for that. As soon as you acknowledge that should it have an Achilles’ heel we'd like to know about that now rather than later, then it becomes the right way for us to behave – to check to see if it has an Achilles’ heel now, right now. If it has the Achilles’ heel we'll end the project and we'll go find something better to do. And if it doesn't have an Achilles’ heel we can move forward with that increased level of excitement and certainty that we're on the right track.
Let's pretend that you are setting out to make a time machine at X and you have two choices. One, to cause a pebble or a mouse to move backwards one minute in time to actually cause some small amount of time travel, or to design the user interface for the cockpit of the time machine. I think we can all agree in this particular example that working on the user interface for the cockpit, once you get it done it will look like you've made progress but you haven't you've made motion. We can all sit there, look at your cockpit and say the chances that we succeed are exactly the same as they were before. We've not actually made any progress because all of the risk was on the part you chose not to do. What you should have done first was the other one.
Our shorthand for this at X is we joke #MonkeyFirst. Similar to the time machine, if you're trying to get a monkey to stand on a pedestal ten feet high and recite Shakespeare monologues and you have a choice between training the monkey first and building the pedestal, if you build the pedestal first when your boss walks by he's like, “Hey nice pedestal!” And then you feel good. You just did something useful, you just got a little bit of attaboy. That's why people do that. But you've utterly wasted your company's money if you build the pedestal first because all of the hard part is getting the monkey to recite Shakespeare. If you can get the monkey to recite Shakespeare we can always build the pedestal afterwards. But if you can't, thank goodness we didn't spend a moment or a penny building what turned out to be a useless pedestal.
Now, in order to get a culture at X to do that we need to reward people when they do it. So when someone raises their hand and says, “My project looks really good but I've discovered I don't think it's going to work out because of this subtle but important issue,” and those ten people now need to find a new jobs. If we say, “Sucks to be you,” no one at X will ever again raise their hand and say that. But if we say to them, “Hey, that was awesome. What incredible intellectual honesty you just displayed and good critical thinking to have discovered this. We're going to work super hard to find all of you new jobs somewhere at X or at Alphabet. We're going to give you all bonuses for having ended your project so thoughtfully. Take long vacations and then when you're done come back and join another project here and get going again with extra levels of passion because you're now working on something that you don't secretly know is going to fail at some point in the future, which is how you would have felt if you had stuck on your old project.”
People are sad when they have to end their projects. They don't end of their projects lightly. But because they've come to understand that ending their project is just being intellectually honest, it's driving for high efficiency, which is why we use the word factory and tie it to moonshots because we're trying to systematize innovation. We're trying to drive to high efficiency innovation. And that can best be done by being honest when you're on the wrong track. And that is failing fast.
Astro Teller is a Hertz Foundation Fellow and recipient of the prestigious Hertz Foundation Grant for graduate study in the applications of the physical, biological and engineering sciences. Teller is now the CEO, aka ‘Captain of Moonshots’, at the innovation factory simply called X (formerly Google X). In this video, he illuminates a critical difference: when undertaking a project, do you want to feel you’ve accomplished something, or do you actuallywant to accomplish it? With the support of the Fannie and John Hertz Foundation, he pursued a PhD in artificial intelligence at Carnegie Mellon University.
The Hertz Foundation mission is to provide unique financial and fellowship support to the nation's most remarkable PhD students in the hard sciences. Hertz Fellowships are among the most prestigious in the world, and the foundation has invested over $200 million in Hertz Fellows since 1963 (present value) and supported over 1,100 brilliant and creative young scientists, who have gone on to become Nobel laureates, high-ranking military personnel, astronauts, inventors, Silicon Valley leaders, and tenured university professors. For more information, visit hertzfoundation.org.
Firefighters in California are still struggling to contain several wildfires nearly one week after they broke out.
- Hundreds of people are still missing after three wildfires spread across Northern and Southern California last week.
- 48 of the 50 deaths occurred after the Camp Fire blazed through the town of Paradise, north of Sacramento.
- On Tuesday night, a fourth wildfire broke out, though it's mostly contained.
We know the dangers of too little sleep. Now for the other side of the story.
- Western University researchers found that sleeping over eight hours per night results in cognitive decline.
- Oversleepers suffer similar difficulties on certain cognitive tests as those who sleep under seven hours.
- Not all the news is bad: One night of oversleeping results in a cognitive boost.
Protected animals are feared to be headed for the black market.
SMARTER FASTER trademarks owned by The Big Think, Inc. All rights reserved.