March 24th, 2015

SXSW 2015: Google[x] Moonshots with Astro Teller 

Astro Teller, “Captain of Moonshots” and director of Google[X], the tech giant’s semi-secret innovation lab, revealed the group’s approach and hard-earned lessons from developing technologies that make the world a different and – hopefully - better place during his SXSW keynote presentation, “Moonshots & Reality.

Google[x], also known as Google’s “moonshot factory”, began in 2010 with the development of Google’s self-driving car. While not every project since its launch has been an obvious success – see Google Glass – the “bumps and scrapes” along that way are part of the group’s process because, according to Teller, “If we want to make a ton of progress, we need to make a bunch of mistakes.

Moonshots Astrotellersxsw 1

Embracing failure quickly and the lessons you incur along the way is part of the group’s approach to solving some of mankind’s biggest problems, in other words, taking moonshots. Kennedy’s vision statement in 1961 to put a man on the moon was the original moonshot, not (only) because it was a success but because it showed the value of dreaming BIG. Big dreams mean big challenges, but they also motivate people, stoking their drive and passion to succeed.

But it’s not all big dreams for a group that still needs to deliver real solutions and products. Teller explains, “By saying we’re taking moonshots, we mean we’re going to go after something that’s ten times better rather than incremental, 10% kind of progress. And it also captures the risk and long-term nature of what we’re trying to do. By saying it’s a factory, we’re reminding ourselves that we have to have real impact. We should take on research-level risks but ultimately we’re developing products and services for the real world. And it also means we have to continue to create real value so Google will continue to support us.”

Moonshotsthinking X 2

The picture above shows how the group approaches moonshots, serving as a blueprint for whether they should move forward with an idea. But the team’s approach has always been to embrace failure, to run at all the hardest parts of the problem first and as fast as possible. Teller explains, “The only way to make progress is to make a ton of mistakes, to go out and find and even create negative experiences that help us learn and get better.”

Teller discussed several ongoing Google[x] projects, sharing their approach to developing ideas, their failures and the resulting lessons learned along the way:

Self-Driving Car: An autonomous car to improve people’s lives by making it safer, easier, and more enjoyable to get around.

Googleselfdrivingcar2 3

Problem(s) to solve:

If a car could drive with greater safety than when people drive in those same places, there are over a million lives a year that could be saved worldwide. Plus there’s over a trillion dollars of wasted time per year we could collectively get back if we didn’t have to pay attention while traveling in a car.


The team had to figure out the 10,000 things they’d have to do to make a car drive itself and see what unusual real world situations the cars would face. Pretty good and pretty safe would not be good enough. The cars may face situations the team had not thought of, but it would need to learn to handle any situation that might occur.

When a new version of the software is produced, it has to prove itself in tens of thousands of situations in a simulator using real world data. What one car learns or is challenged by in the real world can then be transferred to all the cars and future versions of the software, so each lesson only has to be learned once so that every rider can get the benefit from the learning moment.

Googleselfdrivingcar 4

Lesson(s) Learned:

In 2012, the team gave drivers access to cars using the self-driving software for their commutes to work on the freeway. They gained a valuable insight to drivers’ behaviors: people do really stupid things when they’re behind the wheel and it only gets worse when they think “the car’s got it covered.”

They quickly came to the conclusion the car had to always be able to handle the situation without relying on a human for backup. The best way to make that clear was to design a car with no steering wheel , a car that could drive itself from point A to point B, at the push of a button.

The project has succeeded in learning and adapting the technology, so much so that the cars are having to travel further to find next situation to learn from.

Google Glass: The smart phone becomes wearable in the form of an optical head-mounted display in a mission to produce a mass-market ubiquitous computer.

Googleglass 5

Problem(s) to solve:

People have been envisioning how our physical and digital lives will merge through the use of smart glasses in sci-fi for decades. Knowing how to convert that into a product that can be made today and will really work for people is a very different matter and that is exactly why the Glass Explorer program was created.


The Glass Explorer program gave a select and diverse group of people an early version of the device to test.  The program pushed the team to find a wide range of near term applications and uses like firefighting, surgery, cooking or even learning to play the guitar.

Googleglass2 6

Lesson(s) Learned:

The areas that needed technical improvements were quickly obvious. For example, the need for longer battery life was a major obstacle.

But the biggest lesson the team learned came from the “bad decision” to allow, and sometimes even encourage, too much attention for the program. Instead of people seeing the Explorer devices as learning devices, Glass began to be talked about as if it were a fully baked consumer product. The device was being judged and evaluated in a very different context than intended. Instead of being viewed as an early prototype to learn and improve on, people just wanted the product to be better right away, leading to some understandably disappointed Explorers.
Despite rumours, Google’s Eric Schmidt recently clarified that Google Glass has not been scrapped, but instead has been moved out of the Google[x] program and is not its own standalone unit “to make it ready for users.”

Project Loon: A network of balloons traveling on the edge of space, designed to give everyone around the globe access to the Internet, help fill coverage gaps, and bring people back online after disasters.

Loon Illustration Internet 7

Problem(s) to solve:

The goal for the project is to bring Internet connectivity to the other four billion people on the planet that currently have little or no connection to the digital world.


A network of balloons is sent into the stratosphere, between 60,000 and 80,000 feet up in the air, well above the weather and well above where airplanes fly. Each of these balloons acts like a cell-tower in the sky that can talk directly to phones on the ground and to other balloons around it. The team has found ways to make the balloons rise and fall enough (about 10,000 feet) so that the balloons can pick different wind speeds and directions and use that to sail the winds and have some influence over where they will be in an hour or in a day.

Projectlook Nz 8

Lesson(s) Learned:

Loon massively underestimated the difficulty of keeping balloons aloft for an extended period of time by a factor of 10 or 100. When Loon was first tested in New Zealand in June 2013, they were keeping some balloons up for a few days at a time, but more often, it was just for a few hours. They had assumed it shouldn’t be that hard to make super-pressure (non-stretchy) balloons that would stay up for more than three months at a time. It was only after trying and failing to make much progress on this for two to three quarters that it became clear this was going to be a much bigger learning process.

The problem was that the balloons were usually inspected on the ground where everything would seem finem but then at 60,000 to 80,000 feet, it would spring a slow leak. The leaks would only appear in extreme conditions that were difficult to recreate on the ground like when the balloon was at 2% atmospheric pressure, extreme temperature swings between day and night of around 150 degrees Celsius or in high shear winds.

Loon1 9

The team carried out a range of tests to discover how the leaks may have appeared.  They had to first form hypotheses about why the leaks were happening, figure out design changes to the balloon, and then fly balloons with and without the design change to run controlled experiments and see what happened. Since the leaks didn’t always happen, it was a very painful, slow way to find out if the design changes had helped or not.

One experiment the team ran was to see if the fluffiness of the socks worn by the techs building the balloons affected the likelihood that the balloons later had a leak. And yes, it turned out that fluffy socks help since the techs have to walk around on the balloon material as they’re building it. In fact, to control for how they walked around on the material, they had them do a line dance together, first all wearing thin socks and then all wearing the fluffy ones!

Now the balloons stay up for six months at a time, well beyond the three months they think is needed for a viable service.


Project Wing: An approach to delivery using drones to transport goods more quickly, safely and efficiently.

Projectwing 10

Problem(s) to solve:

There is a huge amount of friction left in how we move things around the world. If much of the remaining cost, safety issues, noise, and emissions could be removed from deliveries, while making them take minutes instead of hours, there would be many benefits for both businesses and consumers, but also for disaster relief and emergency situations.


When Project Wing started, the first and most obvious question was “Can an off-the-shelf vehicle be used for this service?” If there was such a vehicle, the team could then focus on the software and sensor issues and move through the learning a lot faster. Sadly, no existing vehicle was even close enough to meeting their needs in speed, payload size, and efficiency.

Projectwing 11

Lesson(s) Learned:

The next question was “What sort of vertical takeoff and landing vehicle style was needed?” They picked the tail-sitter style where the vehicle sits up on its haunches when it’s on the ground, lifts off straight up into the air using rotors like a helicopter, falls forward into a plane-like position for flight and then leans back into hover mode at the destination.

While this approach meant that the vehicle's structure is mechanically simple, it makes it more difficult from a control systems perspective. It initially seemed like a smart trade-off based on the team’s skill-set and because software is improving faster than hardware in most domains.

Projectwing Fieldtest 12

Unfortunately, the tail-sitter was not the right choice because it doesn’t hover well in higher winds and it sloshes the cargo around when it leans back and forth. Even though most of the team realized this about a year into the project, they were resistant to giving up on it because they wanted to get out into the world quickly, and going back to the drawing board would delay doing what is one of the central mantras at Google[x], “Get out into the world and start racking up high quality real-world experiences and learning.”

Sergey [Google founder Sergey Brin] gave the team a deadline of five months to get it out into the world and do some real deliveries to non-Googlers. This had two effects. The first was that it caused the team to double down on the tail-sitter design since there was no time to make anything else work. This may seem like the wrong decision but it did get the project out into the world and there were learnings from the experience. The team launched the test in Queensland, Australia in August 2014.

As soon as the team came back from Australia, they were freed up, with no impending deadline, to do what many of them had wanted to do for more than a year by that time, which was to move away from the tail-sitter design. Even before the Australia test, the team had taken another hard look at whether there was any off-the-shelf vehicle that could work, and having decided again that it still didn’t exist, they were prototyping a new kind of vehicle for a few months in the background.

After returning from Australia, they’ve been hard at work on a new vehicle, its control systems and sensors, and the ways it will provide the service. Google[x] plans on sharing the new vehicle and an update on the project sometime later this year.

Solve for X

You can discover more moonshots, propose an idea for a future moonshot or even get involved in a project at Solve for X, a Google-supported forum for the world to share, collaborate, and celebrate moonshot thinking.

Imagethink Moonshots 13