We're here at the final episode of season two. During season two, we got to go on some amazing journeys with past Xers to hear about the moonshots they're taking at the frontiers of health, biology, computing, and more. The difficulties they've had, the wins they've experienced, and even when it didn't work, how we could repurpose those learnings in surprising ways. For this episode, what we're going to explore is what if you could make an invention machine that is systematizing radical innovation. This is the meta moonshot that I've dedicated the last 16 years of my life to. Innovation is a team sport. It's earned this question: could we now push technology farther in a broader set of domains? I'm famous for saying failure is always an option. The joke is I no longer believe that it's not an option. It's intrinsic to the process. It just occurred to me, we're moonshot busters.
Absolutely, you're moonshot busters.
I'm Astro Teller, your host and Captain of Moonshots, and this is the Moonshot Podcast. I'm joined by Adam Savage, my longtime friend and King of Experimental Curiosity. He's the host of the very popular YouTube show Tested. He was the one who built and ran MythBusters, and we thought we would try something a little bit different. So he's going to interview me as we dig in on what innovation is, how can you really drive it, and how can we all be more innovative in our lives.
So we are turning the tables here. I'm interviewing you. Oh, wait. This was always the better seat. Who knew. Hi, everybody. This is Adam Savage. I'm sitting down with my friend Astro Teller, Captain of the Moonshot Factory here at Alphabet. Dude, how are you?
I'm great. Adam, I'm so glad that we're switching places for today. I think of you as the poster child, the embodiment of the idea of prototyping to drive learning. And I hope that is what we do here. I know that you describe your shop as a possibility engine, and I think maybe we're just a radical possibility engine here. In fact, there are many of your best people from MythBusters who now work here at X. I hope that we get to riff on what it means to be weird, crazy, even silly with a purpose. I would like to think that both you and I have specialized in that over time, doing it ourselves and trying to inspire other people to do it.
We do, and it feels like a particular grace to get to play in that space. Because I was just saying to someone last week, no, no, no, I've had the experience. I can tell you that art and science are the same process. From having been inside both of these plasma pools, the same mechanism is still going on. We're still trying to figure out the shape of the world through different metrics, I agree.
And wouldn't you even say they're often in the same plasma pool. They have to coexist. But it's not like I'm going to do my science over here. And then with a completely different set of people in a completely separate building, I'm going to do my art over here. There is so much art in the science of taking moonshots. At least I've found.
No, it really is. And the whole idea that science is this area where things are clearly delineated and art is this area of fuzzy lines. And I mean, the first thing any scientist learned is that the closer you look at anything, the fuzzier it gets. I have so many questions, but I'd love to start by asking what was the original charge for Google X. What was the original vision for it, and when were you brought in, the founders, Larry and Sergey?
Long before Google X existed, I think right from the start, Google had a vision to do a lot more than just organize the world's information. I know how ridiculous that sounds, but they were convinced to start with organizing the world's information. So some of it, I think, came just from their raw intellect and excitement for helping the world. Some of it came from their understanding of the corporate physics that companies tend to calcify as they get larger, unless you do something to mitigate against that. And so they had this vision from very early on that they wanted somehow to have a part of Google that was separated from Google, whose mission would not be to help Google with its current problems, but to go find new problems for Google—now we would call it Alphabet—to have, and then hopefully to find solutions to those problems. So we were just ramping up the self-driving car project at Alphabet. Literally a couple people. And the person who started that was a man named Sebastian Thrun, but there wasn't really a place for that to live. And I think that became part of their excuse for making a home just as this thing, what now people understand is Waymo, got off the ground. And they asked Sebastian, would you co-found? And he said, yes. And he came and he asked me. And so I came on as his first mate and co-founder for what became Google X. And then after about two years, he moved on. He started Udacity on the outside. And that's how I moved up from first mate to Captain.
Tell me about your two famous grandfathers, and you were talking off camera about ways in which your upbringing may have made you perfectly suited for this job. Talk to me about that.
My mother's father won a Nobel Prize in economics in 1983, and my father's father was Edward Teller, who was the father of the hydrogen bomb, part of the Manhattan Project, the father of SDI, the Strategic Defense Initiative. So I grew up in a household where being smart was the only yardstick. And as it happens, my brother is meaningfully smarter than I am. And I knew that from a young age. And I think that was actually one of the best things that's ever happened for me, though I certainly didn't experience it at the time. But because I had lost in my family's eyes, as I experienced it from this very early age, it caused me to get really interested in what else I was good at and what else are people good at. And so I think I've become a much more well-rounded person than I would have otherwise. I've learned to be a better public speaker. I've learned to be much more creative than I would have been otherwise, to manage and lead people—muscles that have nothing to do with being smart, actually. But I think that was really helpful to me while I was growing up. My father grew up in Los Alamos. Yeah, and for a while. So hearing from my father and from my grandfather about Los Alamos and the Manhattan Project, this idea that you could get a bunch of exceptional people, sequester them away—you have to protect the environment—get them to work really hard, really fast, and really scrappy ways on something that really matters. I thought, wow, I want to do that when I grow up. So from when I was six or seven years old, I obviously didn't conceive of the moonshot factory specifically. I've had a fascination with regularizing, with systematizing the process of getting people together. And I think having grown up on a steady diet of those things, I really affected my excitement and interest and maybe some of my philosophy about how to set it up here.
One of the things I like to tell young people when I speak to them is I say, I think some of you might be wondering about your future and you might be thinking, maybe I'm a genius. And I go, none of you are geniuses. None of us are geniuses. I've met one or two. They are so rare. You got to. We all go. The rest of us have to work together. It's a struggle. And working together, it has to be the fun of working together. Because innovation is a team sport. It just is. We should be the most excited by the people who cheer on their teams and support and nurture their teams. That's what great leadership of innovation looks like, I think.
You have talked about when a moonshot idea comes into your purview, your guys' goal is to beat the crap out of it, is to throw every sling and arrow at it, to see why it's a terrible idea. And if it survives that, then you start to investigate. One of the most fundamental truths about innovation, which I know you know very well, but the visionary class would like us to believe is not the case, is that radical innovation is almost all you like. Look under some rock and it's like, nope, nothing under that rock. Or put the rock down fast. Radical innovation is mostly about failure and learning. It is not a story of success where some genius just had the right idea. That's just not how it works. And so if you were setting up a moonshot factory, unless you want to set yourself up for failure and constantly trying to smokescreen your way through the process, you would start with, this is how it actually works. So let's be honest and then set up a culture that can embrace that instead of trying to be in denial about it.
I'd like to talk about the culture that embraces that, because within America, right now, within the world right now, there is a standard way to build a company. There are hierarchies you can choose. But a moonshot factory, I think, requires a fundamentally different kind of structure, one in which failure is not just accepted but encouraged. I'd love you to talk about some of the differences between the structure of Google X and other parts of Alphabet.
I would start with we are not an incubator. So if you come to work here at X, the very first thing I need you to know is you're not on the teleporter team. If you're working on the teleporter, you're on the X team. It can't be that you're trying to keep the teleporter alive. If you're trying to do that and I aren't on the same team, because my interest is not actually in having the teleporter work. My interest is, how can we solve as many of the world's big problems as possible per dollar spent. So that has to be the team you're on if you want to be here with me. Once we agree that that's the team we're on, now you're not sitting on the other side of the table. You're on my side of the table. And we're saying we're buying options on the future for Alphabet and for the world. Might this be a once-in-a-generation opportunity? I don't know, let's go find out. And you find out by pressure testing, pressure testing the idea, then pressure testing the engineering. If I was trying to get you to come work at X, one of the other things that I would tell you is if you need to be right, this is going to be a horrible place for you because 99% of the things that we try turn out not to be real dents in the universe, real once-in-a-generation opportunities. So everyone who comes in here has to move their sense of self-worth as a professional from 'I made my thing work' to 'I'm world class at discovering which of the 100 things we're looking at is worth making.' I'm famous for saying failure is always an option, and I no longer... The joke is I no longer believe that it's not an option. It's intrinsic to the process, completely intrinsic.
So after 16 years of doing this, I'm curious if you have a metric as to what if you're better at spotting a reasonable moonshot? So moonshot, huge problem with the world, radical proposed solution, breakthrough technology that at the highest level hasn't changed.
And I would say we hold our bar, if anything, even higher than we used to in terms of the craziness of what we should be shooting for. That maybe makes it harder in some ways, but it makes it more fun, I think ultimately more profitable, more good for the world. We're much more three-dimensional at how we test things. I think no one said techno-economics for the first two or three years of Google X. Now it's weird. If you go a day without hearing somebody say techno-economics, it's just one of the fundamental ways to pressure test what we're doing. And for a lot of things, it's the question: if you're going to work in a commoditized field like energy, the levelized cost of energy, LCOE, like that's it. That's the techno-economics. Right. And so you can't put that off till year five. You should be having that conversation in the first conversation, like, what would this be. And if we don't know, that's fine. But that's the first question to go sort of bang on. So we're much more three-dimensional about who would want this, why would they want this. And being able to be clever about how we get really signals about what people would want, why they would want it, what they would do with it, maybe long before we can have a really polished version of it. So that helps. I think there's just a certain amount of pattern matching in art as well as in science you were suggesting. Now that we've seen, we've started and killed about 2,000 ideas over the last 16 years.
Wow, what a great metric. You started and killed 2,000 ideas. It's so awesome.
And we now have a really searchable way of being able to interrogate it.
Fantastic to look at your past.
Yeah, so if you want to say, ultrasound, everything that had ultrasound in it, then there's dozens where we like ultrasound played some part in it. And you can look at those and these things come back more and more frequently, but being able to say, why did that one die. That's why. So I'm curious if there's been places where in failing at one idea of a moonshot, but looking at why something was hard, found another one.
We call this moonshot compost, and it happens constantly. In fact, literally the previous meeting, which was one of our internal meetings where people are bringing very early investigations to the table, they were talking about a particular idea, and we're like, oh, a quarter of that idea was that thing that we killed about a year ago. Yes, because it wasn't useful for other reasons. But it's perfect for this. And he said, right, that's why I'm excited about this. And so the ability to re-harvest people, patents, code, learnings, partnerships, whatever it is that we got along the way when most of our stuff is misadventures. That learning isn't just oh, don't go that way. There's so much where we're like, oh God, this would be awesome if we only had unobtainium. And then you put it aside. But you're also like, note to self, if anyone finds unobtainium, let's get this out of the refrigerator. Meanwhile, you've got this crew where somebody's like the world's expert at a totally pointless thing. That may not be pointless next week.
Yeah, I love that. So it also strikes me that with that kind of institutional knowledge, you must have some long-term employees. People have been around for a goodly portion because you need that institutional memory, don't you.
We do. Several of them used to work for you, as you know. Turns out MythBuster and it's a nice character overlap.
We do have, we're very fortunate that there are a bunch of people here. I mean, we have several who've been here 15 years, at least another 20 or 30 who've been here at least 8 to 10. And that institutional knowledge is super important because there is a certain 'trust the process' feeling here, which it would be easy to follow if the process were really rigid and explicit, but because it's not, you can't take a bunch of iconoclasts, tell them to think differently and then be and be nonconformist. Which is the whole point in taking a moonshot is you have to say no. You're wrong to experts, at least in some way about that field. You cannot take those people and then say, and here's the rule book you must follow to get there. That is definitionally a failure. So you have to create these guardrails that sort of shape and direct them with very light process and that there's rules of thumb. We're not going to let anyone hurt anybody. No fraud or embezzlement, please. But beyond those pretty wide guardrails, you have to leave a lot of space. And so the institutional memory of why this stuff is set up the way it is, very important. Some of our most beloved Xers came from MythBusters. Why do you think that is.
I'll start by saying spending 13 years being a MythBuster, being able to move the levers of that machine, which was 25 people strong across two different shops, was a really remarkable gift. We had these associate producers who had a huge amount of autonomy over finding us a location, finding us an expert, finding us materials. We would ask for specific things. We were often wrong in the framing of our question, and they would have to come back and help us understand. And that feedback loop between production and Jamie and I and Kari and Grant and Tory trying to solve these problems on camera. There's this moment when Discovery was thinking about making everything 3D, and they sent out the best 3D producer in the world who worked for James Cameron at the time, and he came and embedded with us for three days. And at the end of three days, he turned to me and Tappert, my co-producer, and said, let me get this straight. The experiment takes precedence over the filming. And we said, yeah. He said, I would have never imagined that model could work. But that was the model all the way down. So we built this marvelous team of contingency planners who understood how to build relationships, and that comes right from the top of most of the core team of MythBusters. Coming from the film industry, which is a collaborative medium, and understanding, I don't want to know how you do your job. If I do, there's a problem. I just want to know you can do it. And then we're all pulling on the same rope. It just occurred to me, we're moonshot busters.
You absolutely are moonshot busters. That's literally what we do. We come up with moonshots, and then we try to bust them. And when you fail at that, you've succeeded.
Exactly, that's most of our work. Oh, that's beautiful. I'm going to ask you for some stories. Must have some spectacular stories of amazing successes or failures over the past 16 years, or even just insane things you learned unexpectedly.
I more than I can count. This place produces them on the daily. I will just give you a few off the top of my head and I encourage this kind of thing, but it makes me the happiest when it's not a specific thing that I encouraged. I remember the Everyday Robot group, in a moment of frustration mixed with playfulness, started shaming their robots where if the robot poured Coke on somebody's keyboard while it was trying to clean up their desk, they would hang a sign around its neck that said 'I spilled Coke on Mario's keyboard,' and they would make the robot have to wear it for the day. Amazing, as though shaming it would cause it to be better behaved. Obviously they knew that wasn't going to happen. That just made me happy because I understand that playfulness translates over time to a radically different kind of creativity. So I don't care about the poster on the board.
Wait, I'm just curious. I love to dial into that different kind of creativity. Is it specifically that the team would take time to externalize an emotion that's happening on the team?
To a certain extent, that's being a little bit, yes. Externalizing the emotion is healthy. It's cathartic. But also, if you can't be playful during your work, yeah, you won't be playful in your work. And playfulness is the root of creativity because creativity is the act of doing something not normal. And so when I see the playfulness during the work, I know that they have, they're building the muscle for doing the playfulness in the work.
I love that. I'll give you a different example. That's not playfulness. But when the Wing team started, I told them that they had to crash five wings before not too long. And I think there was a little bit of 'you're kidding, right.' And then they had to work through the fact that I wasn't kidding and that I was trying to help them connect with the idea that if they crashed these huge like electric airplanes, too often it would be a disaster. It would be way too expensive. But if they never crashed one, the amount of slowness they would have to introduce into the process to never crash one would mean they were guaranteed to be wasting Alphabet's money. So I wanted them to find the sweet spot. Yeah, it was when they brought me the first picture of a full wipeout crash where there was a yard sale of this thing spread across a crash site. I felt like, good. They heard me, and they get why to do it. That they weren't bringing it to me as an apology. They were saying, we heard you and we're doing it. They really understood the assignment.
Yeah, I mean, so there's been so many things like that where people have had fun in hard times. Yeah, and we lost a balloon in the early days. And we tracked it across the United States and up into Northern Canada via UFO sightings. And the fact that was the easiest way to find out where it was. Yeah, happened to mix nicely with, just the hilarity of it. And at the same time, it was like demystifying this thing. The fact that that kind of thing is an almost daily habit here is partly because what we're trying to do is audacious, but it is also because the people here have a sense of humor about what they're doing. Back to your point earlier, and can imbue these hard things that we're doing with playfulness and the creative process. Our monkey on a pedestal point.
No, we've never talked about this.
No. All right. Well, rabbit hole. But it'll be worth it. So for years I was saying to people here at X, you have to work on the riskiest part of the problem first. And everyone was like, I don't think that's right. That sounds wrong. I can't prove it, but I'm pretty sure you're wrong. Astro, it just kind of didn't work for them. And so I said, look, imagine that we had a project at X and the project was to get a monkey to stand on the top of a 10-foot pedestal and recite Shakespeare. Which should we do first. Build the pedestal or train the monkey. If we build the pedestal, we're half done. But in that extreme case, we can all see that we're not half done. We're 0% done. Because the entire risk, all of the real work in the sense of is it even possible, is piled up on the monkey, not on the pedestal. So you should train the monkey first. And don't worry about the pedestal. You always get it later if you actually train the monkey. This has taken on a life of its own here. Where now, I've probably been in three meetings today where people said 'this is the monkey' and that is just shorthand at X. Now, for this is the riskiest part of the problem. This is what we need to focus on. That's an example. This is back to your question of stories. Stories are such a powerful way. They're architectures for understanding. That's what a story actually is. I don't mean a marketing slogan, but when you create an architecture for understanding something that is portable, that makes it easy to remember, it makes it easy to pass on, you can now pass on the monkey story to somebody else. It changes things. And the fact that this place is full of stories in several senses of the word is part of what makes this machinery, which is the factory, work so well.
As a young model maker, I accreted towards always doing the toughest part of a problem first. And it was also partly because I really liked the endorphin rush of doing the pedestal once I've done all the hard stuff. I love doing that pedestal at the end of everything, right. Because there'll be a payoff for all the work of the pedestal. Now that the monkey's trained.
Yeah, and we've all built a pedestal before solving the hard problem. It sounds like you've really carved out a place here also for joy. One of the meditation teachers I read early in my life said, don't wish for happiness. That's impossible. But joy, while fleeting, is definitely something to shoot for.
Yeah, I think I could do you one better. The only thing that's taped up on the door of my huddle here is a poster of the tarot card of the fool. Joy might be somewhat harder to access, but the beginner's mindset if you practice, you can have that almost all the time. And there is a joyful, playful, almost ridiculousness of that feeling at the beginning of a journey where you say, yeah, but what if. The fun of getting to say what if, the acknowledgment at the beginning of the journey, you don't know where it will take you. It's probably going to have misadventures. You might have to start over. And that starting over is part of the fun. Not like a punishment. That's why it's the only thing that's taped up on the outside. And so when you say that we have joy here, I would hope, yes, I hope there's a lot of joy here at X, but I hope the beginner's mindset is actually the thing that's closest to what we maintain as our central ethos.
I remember telling a JPL scientist how hilariously terrible of an idea I thought the sky crane was. And they were like, yes. And it was the least terrible idea we had to deal with. That's why we made it work. And I thought of you as the second I heard that, I'm like, that's this whole place's ethos. We do bad idea brainstorms.
No, explain to me how this works. I love this idea.
So if you join a brainstorm and any of us is there. But even worse, if I'm there and I say, tell me your great idea, be brilliant right now. Guess how many brilliant ideas I'm going to get out of you. Guess 0. Guess how much creativity you're going to be able to muster in the moment where seized up with 'please the boss, please the boss.' But if I say this is a bad idea brainstorm. We're going to come up with some truly horrendous ideas. I don't want to hear it unless it definitely sucks. All of a sudden you've been challenged by how bad the idea can be. It does the opposite. If we keep putting stuff on the table, you'd be surprised. It doesn't take that long before someone's like, that idea might not be totally bad.
