The Switch's Brian Fung spoke with Astro Teller about the big ideas behind X's projects. (Daron Taylor/The Washington Post)

From balloon-based Internet to self-driving cars, Astro Teller and his team have produced some of the world's most ambitious technology products. Teller oversees X — formerly Google X — which seeks to solve enormous societal problems, such as what to do about climate change. Earlier this week, Teller sat down with The Washington Post to talk about the business of big ideas. The following transcript has been edited for length and clarity.

So Astro, tell me about some of the projects you're most excited about — the big challenges you're trying to solve.

Generally, we think of moonshots this way: The first is, there has to be a huge problem with the world that we want to solve. The second is that there has to be some science-fiction sounding product or service that, if we could make it, however unlikely that is, it would actually make that problem go away. And then there has to be some technology breakthrough that gives us some faith that we could actually at least get started on trying to build that product or service.

So one of the things we do is start from basic technologies like new materials, or machine learning, or something like that, and try to go up that stack, toward the problems. What problems could these new technologies help with? Sometimes, we start with the problems. We look at global warming, or food production. We look at cybersecurity. We look at anything that is a big problem in the world and then say, “What could we add to the ways in which people around the world are already trying to solve these problems?” And when they meet in the middle, that's the opportunity for a moonshot.

How do you decide that a project just isn't going to work out? Where do you draw the line?

It really depends. When we say it's not going to work out, it's for very different reasons. It happens usually very early on in the process. Sometimes it breaks a law of physics. "Nope, this just isn't going to happen." Sometimes it's possible, but you'd have to make it out of such expensive materials that there's no way you could make it cost-effective. Sometimes there's no reason to think that it won't work, but it would take so much money just to get the very first piece of information about whether we're on the right track or not, that it's not appealing to us.

Is Silicon Valley doing enough to grapple with some of the social implications of its innovations? There's been a lot of talk about how the sharing economy or other types of disruption are changing the world — and not all of it has been good.

I think there are clearly responsibilities that any technology developer has for at least understanding the externalities of the technology being developed. But sometimes those externalities are the kinds of things that one ought to take responsibility for as a technologist. And sometimes those externalities are really better dealt with by the public — by public policymakers, regulators, et cetera — not by the technology companies themselves.

What's something that nobody's talking about that is going to be really important and influential, or what's a problem no one's trying to solve right now?

Clean water is an interesting example. There are huge chunks of the world in which access to clean water is sort of the rate-limiting piece of civilization, and there is not a sufficiently cheap, reliable way to filter that water. It takes a lot of energy, all the ways to do it right now just take a lot of energy. And the fixed costs for the machines to clean the water is also pretty sizable, ways that take somewhat less energy require very expensive, relatively speaking, expensive filters. So that's an example of something lots of people care about, lots of smart technologists have taken a run at, but so far there really isn't a solution that has just made that a problem of the past.

What's one thing that you guys do as an organization to help keep ideas flowing? Who at the organization comes up with ideas in the first place, and what do you do to facilitate that process?

Everyone is qualified to come up with ideas. If you go to a kindergarten and you ask kids to have some ideas, every single child in the room will have ideas for you. Crazy ideas. The ability for us to have ideas and to put them out in front of our peers is lost as we grow up. It's beaten out of us. It's not something that you need to find rarefied people who are good at doing it.

The trick is to make them feel safe, so that they can put their ideas in front of their peers and not feel like they're going to get laughed at. And then you have to weed through those ideas in a thoughtful way.

So the best-case scenario — and I've seen this happen many times at X — is someone brings up an idea. I don't know, they say, “What if we could harvest the power in an avalanche?” Okay, that's probably not going to work, but the first answer that they have to get is, “What a gorgeous idea!” Because if you can ask that question, and 100 questions like that, one of them is going to be right. And that has the form of a brilliant idea, even if it doesn't happen to be a brilliant idea. High-five! And everybody's excited for you, you feel really good right now. Now I say: “All right. How can we kill this idea? How can we, as fast and as efficiently as possible, satisfy ourselves that this is not the 1 in 100, it's the 99 in 100, so we can get onto your next idea.

Sort of a process by elimination.

Yes. And we sit there at the whiteboard, and we come up with five different reasons why it's really not going to be practical. Can we force avalanches to happen on time? No, not really. Can you restock the snow or the rocks up high and do it again? No, not really. Okay, can you move the equipment that's going to harvest all the kinetic energy in this stuff when it's falling around, to catch all the different avalanches? No, that's going to take way too much energy. Can you put it at the bottom of every mountain in the world? No, that's probably not going to work. After half an hour, you and I have satisfied ourselves that it's not going to work, we cross it off the list, and I give you another high-five. You just got credit from me, both for coming up with the idea and killing the idea. And if you work at a place like that, before you know it, everybody's coming up with ideas.