Google X’s Astro Teller says desire for learning should drive innovation

Google X’s Astro Teller: "Failing fast is normal – as children we do that, but we unlearn it as we grow up."

One of the biggest drivers for innovation is having a deep, ingrained desire to learn new and difficult things. This is where the magic happens, says the head of Google X lab, Astro Teller.

Teller – who has led projects such as the self-driving car, Project Loon and Google Glass – was in Brisbane’s River City Labs this week to share his advice on innovation.

When developing a new project, learning should be the top priority, Teller said. Often people create priority lists ordered by the most important task to making the project successful, rather than the task that will force them to learn the most.

“Re-order that list in order of learning. You put the thing at the top that you think you’ll learn the most from doing, however unimportant it is.

“You can’t know what 20 most important things are; you can only know what experiments you can run that will teach you the most.”

One of the best ways to learn is to release a product out into the world in its beautiful imperfect form. This is to avoid determining the end product and how it should be received early in the process, which is what hampers the process of learning and discovering something new, he said.

“If you want to lose, just keep your head down, stay in a conference room that has a whiteboard, order a lot of pizza, and believe that you can simulate the world well enough that if you navel gaze long enough your design will eventually reach perfection.

“If you want to win, make it bad, make it quickly and get it out into the world. It won’t work, but you will learn something interesting when it doesn’t work.”

As much as we all try to predict outcomes, in reality the world is unpredictable and the many ways it can surprise you is key to making innovation happen.

“You can’t simulate what people will do with your product, even though you wish you could predict what they will do. You often can’t even predict what physics will do with the thing that you are making, if you are making a physical thing. The world is just surprising in that way.

“The thing you are positive will break first, if you are lucky has a 20 per cent chance being the thing that actually breaks first. Take it out into the world. That has been our experience at [Google] X over and over again; that is how to learn.

“Run as hard as you can at those learning things,” he said.

In order to get a product out into the world early, he said companies need to embrace failing fast. He said recently at Google X a 25-person project was closed down because the team was able to fail fast and discover a new direction or where they really wanted to go in terms of solving a real world problem.

“This was an unannounced project. We got to a place where it was like this would work if it was for super wealthy people and they would only do it in their basement and they really hated the noise of fans, and other things. We thought ‘that’s not a business’.

“By the time you are torturing yourself to try and make it sound like it’s great, you’re done.”

The team was not disheartened by having decided as a whole to end the project and take a new direction. Instead, they were rewarded for learning it was a dud, in terms of Google’s standards for world innovation, early in the process. This means time and money invested is not wasted by learning this years later.

“If you want people to fail quickly, you can’t just get a sticker that says fail fast and put it on the wall of your startup and think people are going to spontaneously fail fast. If you want people in your groups to fail fast you have to actually thank them when they fail fast.

“Failing fast is normal – as children we do that, but we unlearn it as we grow up. We unlearn it because all of our reinforcement teaches us not to fail fast.”

People also need to feel psychologically safe to embrace the fail fast mentality, Teller said. Being humorous about the rough edges of project puts people in the state of mind that it’s OK to not be successful the first time and to not take perfection too seriously, Teller said of his experience leading teams.

“I would encourage you to think about humour and creativity and how they are linked, and what you can do in your company to encourage people to bring their whole selves to the company. Failing doesn’t happen unless people feel psychologically safe,” he said.

Google focused on building intelligence

Teller gave some insight into where he thinks the future of industry and technology is heading, with the world moving into an era of intelligent machines.

“We’ve been in this process for 50 to 100 years of kind of industrialising the world mechanically.

“The way cars are thought of in terms of their safety, for example, is still fundamentally take the car and smash it into a wall really fast and make sure that the mechanics or the physics of the structure make it so that the person or dummy inside is still safe,” he said.

“We are moving to a world where for cars, and lots of others things… it’s asking the question how can we make this technology more intelligent? How do you make a car so it will never hit a wall [in the first place]?”

Problems that used to be solved by mechanical means are now being solved by embedded intelligence, which is where the future is heading, he said.