This month’s Harvest Summit in Sonoma County opened with a discussion about “incremental versus exponential thinking.”
It featured Hans Peter Brøndmo, the General Manager of the robotics initiative inside Google’s research and development division, known as X; and Marcus Shingles, CEO of the XPRIZE Foundation, which designs competitions to solve some of the world’s biggest challenges.
Here are some of the highlights from the session.
At X, we have a very simple definition of moonshot. It’s the combination of three things.
First, you have to address the huge problem, something that could affect literally billions of people.
The second is you need to come up with a radical solution to addressing that problem. People might’ve been thinking about it for years, decades, or millennia, but we’re looking for radical new ways of thinking about that problem.
The third is that we need breakthrough technology to address the problem, and to solve it with a radical solution. In our particular case, technology’s the lever and the unique part. But if you think of those three areas as circles, the intersection of those, that’s how we qualify a moonshot.
Even more important than the moonshot itself is what we call moonshot thinking. How do you think about problems at the scale of being able to affect the world? How do you think about radical solutions?
…We distinguish between “roof shots” and moonshots. You could say, “If I wanted to get up on the roof of this building, I could come up with a really innovative solution. I might actually technically, be closer to the moon, but I didn’t do anything to get to the moon.”
In order to get to the moon, you have to actually think about the process. It’s not an incremental problem; you have to think 10x. Building a really cool ladder that can get you up to the top of a building, while technically closer to the moon, doesn’t get you there.
The other [thing] is, we use a lot of spaghetti. By that I mean, we have to throw a lot of spaghetti onto the roof to see what actually sticks. We run tremendous numbers of experiments, and very few of them actually turn out to [be] moonshots. We might have really interesting ideas, but we just go through tons and tons of spaghetti to see what finally sticks, and to see what type of early ideas we might be able to develop further to get to the moonshot.
Moonshot thinking is about being more in love with the problem than the solution. Really obsessing about problems. Really obsessing about how to solve problems, and really being in love with that state of mind. Rather than necessarily falling in love with the solution — because the solution’s almost always wrong.
Every organization right now should have some edge organization. An edge organization thinks, and has brand permission, to take moonshots. The reason why it has to be on the edge is because the core business has antibodies in it that smother innovation.
Any organization that’s done innovation well, you don’t create innovation from within the core because of those antibodies. Also…the normal tendency is [that] the core sees a success at the edge, and it brings the edge business back into the core because it’s having success and they feel territorial.[Instead,] the objective is that the edge business pulls the core out to the edge, and eventually your edge becomes your new core business one day.
The other notion is that if you give the freedom to do 10x versus 10 percent, it doesn’t cost 10 times more to do 10x. You don’t need 10 times smarter people to do 10x, so why would you not try to go 10x? Why would you not give some portion of your organization permission to go 10x?
If you do that, you have to create a different cultural environment. You have to have a fail‑fast, fail‑cheap, fail‑forward mentality. You have to create mechanisms that reward failure. Astro [Teller] talks about this at [Google] X, that they actually celebrate when someone fails. Why? Because the culture there is that failure is not the opposite of success. It defines the true parameters of what success could look like.
That notion is very important, because…if you are successful with a project and if you don’t find yourself failing often, you get to a point where you think, “This is great” — the normal culture of big business — “We didn’t fail. We got to success. We got to a solution.” If you didn’t fail, by definition, you didn’t really stretch the bounds to see where things break.
In today’s technology environment, with biotech and nanotech and advanced robotics and 3D printing and artificial intelligence, if you’re not failing once in a while, you can pretty safely assume that you’re going for competitive advantage, you’re not going for disruptive advantage.
We focus on learning. The key is, what do you learn from your failures?
How many people in this room have at some point in their life failed and learned something? [Laughter, most hands go up.] If you can bring that mentality to business, then we can allow a culture where, look, if we’re going to take risks, we’re going to potentially fail. But then what can we learn from that, and how can we learn fast?[We should also allow people to acknowledge that something is failing and move on to something different.] “If I’m really brutally honest about the fact that I’m failing at what I’m doing right now, or part of what I’m doing is failing, there’s an opportunity cost associated with continuing to push and push and push,” and try to prove to others that you’re actually not failing and put up a veneer.
What if instead I could say to you, “You know what? You failed. Congratulations. I’ll give you a bonus, and now you actually get to pick your next project. You get to work on something else that unlocks your potential, and I’ll give you the resources to work on that.”
If you create that kind of culture, then people will be more honest and they’ll be more open to admitting that this isn’t working out. “I get to work on something else.” As a company, as an organization, you get to reallocate those resources towards something that might ultimately generate more of a positive outcome.
Most of our performance metrics as an organization are wired a way, that if you’re managing a project and you know it’s not going that well and you’re starting to have that sinking feeling in your team, like “This may not work,” what you usually do is you usually ask for more resources and more money and more bodies to try to make it work, because you don’t want to go to your management team and say, “This is not going to happen.”
That means you’re taking a “fail fast” into a “fail long.” That means you’re taking a “fail cheap” into a “fail expensive.” Then the notion of failing forward is, you’re failing without making the same mistake you did the last time you failed. There really is a discipline, and it’s a corporate culture [that] creates good failure versus bold, epic, negative failure.
We have this incredible and unique privilege to manage, which is we have a lot of capital to apply to this. We can make investments in these things that have five, seven, 10 year horizons. The cars — you’ve probably heard of the Google self‑driving car — that’s been seven years now.
We’re solving the hard problem. The easy problem is getting a car to follow lanes. You put a little camera in the front, and call it autopilot. The hard problem is to actually make a car that can drive through the neighborhoods. When the father with a child in a stroller walks across the street, the car knows to stop.
In the unique edge cases, the many, many cases that you would never be able to predict. A big truck passes across, and there’s a big stop sign on the truck. Recognizing that that’s actually not a stop sign, but a moving car. There’s so many edge cases where you basically have to say, “If a human was sitting there, what would the human do, and how would the human react?” Now the car has to do it. That is a moonshot.
That is something that’ll take, to truly create a self‑driving car where you get into the backseat, and you go to sleep, and the car wakes you up when you arrive, that’s a very, very different way of thinking about how we solve big problems, than solving them by incremental solutions, which is what some of the nice luxury cars can do.
There’s seven billion people on the planet. The chances that you have a monopoly on the smartest brains in the world is probably unlikely. …They’re somewhere out there in the crowd, and you have to really start thinking about a crowdsourcing model for talent. It’s going to change the economics of firms. It’s going to change the economics of businesses.
My daughter graduated from college last year, and she’s probably [in] the first graduating class where her options are better to not go into industry and not go into an organization. She will make more money and be more employable by being in the crowd economy.
Within a matter of moments after doing a search on Google for “environmental science oceans crowdsourcing,” she found dozens of sites of someone in the world that was looking for the knowledge in her head and willing to pay her for days, weeks, months, or hours to get that knowledge.
We’ve got three billion people connected to the Internet today, and within seven years it is fact we should have the whole planet online, whether it’s the Google Loon project, whether it’s the Facebook drone project, whether it’s Qualcomm satellite initiative, there are multiple initiatives that will get the whole planet online.
You have to think about the fact that in five to seven years, you have seven billion people connected, not three billion people. That’s four billion new minds that are entering the connectivity. You have to think to yourself that there is genius in the crowd that will get tapped into as a result of that, and you have to consider the fact that that crowd has supercomputers in their pockets.
They have more technology in their pockets than big government or big business has had twenty years ago. That makes a different world. That’s going to make a different world for social impact, which X Prize does because we crowdsource that for coming up with solutions to grand challenges.
But it’s also going to affect how you run business. One of the things I would tell my executive clients before getting into the X Prize role [Shingles was a partner at Deloitte Consulting] is you do not want to be caught flat‑footed five years from now when all of a sudden you wake up and you got seven billion people on the planet connected and five or seven more doublings of the supercomputer in your pocket. That’s going to be a significantly different place.
Virtual reality and augmented reality will actually change the corporate culture and dynamics of how people work and collaborate, even social networking. You should really be planning now for that trajectory of identifying how you do moonshots, but how do you leverage the ecosystems that are out there?
From Tel Aviv to Santiago, Chile, to Brooklyn, to Silicon Valley, you could have Google engineers by day that are working in the crowd for you on nights and weekends. That’s what you’ll find out with those models.
There are two areas that I’m very excited about that X is working on right now. One is energy. We’re making a lot of investigations and big bets on energy. I think we’re going to make a lot of progress on the energy side. It might take more than five years, but we’re on a good trajectory there.
The other one is our artificial intelligence and machine learning, big data. The breakthroughs that we’ll see in AI and ML are going to be remarkable. It’s really an ingredient. It’s not a solution unto itself. It’s something that’s going to pervade everything.
For those of you in food, it’s going to have a huge impact on how we think about food, how we think about health. It’s going to have a huge impact on certainly the work I do in robotics. [We’ve only started adding] true forms of autonomy and intelligence to the world of machines. That’s going to be enormous, whether it’s cars or trucks or forklifts or something else that we might invent in the future.