The lab behind Waymo and Google Glass that wants to reshape your life

The moonshot factory X is home to some of Silicon Valley's boldest inventions. CEO Astro Teller reveals how the secretive lab tests crazy ideas that can change the world... even when they fail. TED Radio Hour+ subscribers now get access to bonus episodes, with more ideas from TED speakers and a behind the scenes look with our producers. A Plus subscription also lets you listen to regular episodes (like this one!) without sponsors. Sign-up at plus.npr.org/ted http://plus.npr.org/ted. See pcm.adswizz.com https://pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences. NPR Privacy Policy https://www.npr.org/about-npr/179878450/privacy-policy

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Manoush Zomorodi?

  This episode's sponsor is HubSpot, which offers the following message: What if instead of updating your CRM, your CRM updated you? HubSpot's self-updating CRM keeps your customer context current so you can win more deals. Learn more at hubspot.com. A note before we get started. We talk a lot about Google and its parent company Alphabet in this episode. Google is a financial supporter of NPR.

• 0:31Manoush Zomorodi?

  This is the TED Radio Hour. Each week, groundbreaking TED Talks. Our job now is to dream big. Delivered at TED conferences. To bring about the future we want to see. Around the world. To understand who we are. From those talks, we bring you speakers and ideas that will surprise you.

• 0:51Speaker 01

  You just don't know what you're gonna find.

• 0:53Manoush Zomorodi?

  Challenge you.

• 0:54Speaker 01

  We truly have to ask ourselves, like, why is it noteworthy?

• 0:56Manoush Zomorodi?

  And even change you. I literally feel like I'm a different person. Yes. Do you feel that way? Ideas worth spreading. From TED and NPR,

• 1:09Manoush Zomorodi?

  I'm Manoush Zomorodi.

• 1:11Speaker 02

  For your safety, the doors will remain locked when we arrive.

• 1:15Manoush Zomorodi?

  A few decades ago, the idea of a self-driving car sounded like science fiction.

• 1:21Speaker 02

  You're here.

• 1:22Manoush Zomorodi?

  But now, these cars are becoming pretty commonplace in cities like San Francisco, LA, Phoenix. My ride is here. I'm taking Waymo. One of the leaders in the self-driving car industry is Waymo.

• 1:36Speaker 03

  Thank you.

• 1:38Manoush Zomorodi?

  And passengers have taken to social media to share their experiences.

• 1:42Speaker 04

  Y'all, I always wanted to try one of these things. And babe, I'm about to do it.

• 1:47Manoush Zomorodi?

  With a wide range of reactions.

• 1:49Speaker 01

  Who is moving? It was so weird having no driver in the seat.

• 1:52Speaker 05

  Y'all, I cannot believe this. I've got this whole car to myself driving down the road.

• 1:59Manoush Zomorodi?

  This passenger in particular loved that Waymo cut out the need for human interaction.

• 2:05Speaker 05

  This is just such a nice experience to be in here completely by myself. When you hop in a Lyft or an Uber, you know, there's another person driving. And nothing against them, but it's just like, you don't have your privacy. You don't have your space.

• 2:17Manoush Zomorodi?

  Of course, there are also a lot of videos made by passengers showing the funny and glitchy sides of this technology. One TikTok video showed a group of Waymo vehicles honking at each other incessantly at 4:00 AM in a San Francisco neighborhood. Another captured a man who'd fallen asleep during his ride and couldn't be roused. Hey, it's time to wake up. Yet another documented a man who had no way to stop his Waymo, which had started driving in circles in a parking lot.

• 2:47Speaker 06

  I'm in a Waymo car.

• 2:48Manoush Zomorodi?

  Connected to rider support.

• 2:50Speaker 06

  Yeah, I got a flight to catch. Why is this thing going in a circle? I'm getting dizzy. It's, look at what it's doing.

• 2:55Speaker 03

  I understand. I'm really, really sorry, Mike. We're currently working with the situation of the vehicle. Is it circling or?

• 3:01Manoush Zomorodi?

  Some say they feel safer in self-driving cars. Others feel less safe. So the question becomes, who do you trust more? A human or a machine?

• 3:16Speaker 07

  Right now, we live in a world where mostly humans are driving cars. And more than a million people a year die in car accidents around the world. The vast majority of them, something like 95% of those deaths, are caused by driver error.

• 3:36Manoush Zomorodi?

  This is Astro Teller. He is the CEO of X. This is the company Alphabet's so-called "moonshot" factory, where some of the brightest minds in technology come together to try to figure out how they can solve some of the world's biggest problems. And human fallibility is one of the many issues they're trying to hack. They want you to be able to trust machines more than humans. Their self-driving car project, now known as Waymo, was one of their first moonshots. And at the time, the big question they were asking was...

• 4:16Speaker 07

  What if we could see cars as being the modern equivalent of elevators? There was a time when elevators were operated by humans. And we now, when we look backwards, can see that as very quaint. And so that was the radical proposed solution, that sort of vision of the future. And then there was a number of reasons to believe it might be possible. Maybe the best example was that DARPA...

• 4:46Speaker 07

  The Defense Advanced Research Projects Agency for the United States had been running these grand challenges for a number of years to encourage academics in the United States in particular to work on self-driving vehicles.

• 5:02Manoush Zomorodi?

  Didn't they have races out in the desert?

• 5:04Speaker 07

  Exactly. So they had a race out in the desert in 2004. It was about 150 miles and nobody won.

• 5:11Manoush Zomorodi?

  Nobody finished?

• 5:12Speaker 07

  Nobody even finished.

• 5:13Manoush Zomorodi?

  Yeah.

• 5:14Speaker 07

  But in 2005, they had the same race, same 150 miles, and three teams finished. And the team that came in first was, the car was called Stanley, and the team was run by a professor at Stanford named Sebastian Thrun.

• 5:37Manoush Zomorodi?

  Sebastian Thrun ended up working at Google. And a few years later, in 2010, he started X, the moonshot factory, bringing Astro in as his co-founder with the hopes of building a 21st century Bell Labs, an incubator for bold, life-changing inventions.

• 5:56Speaker 07

  So the first project was the self-driving cars. And X was formed as a kind of place for that to live and to do more things like that.

• 6:07Manoush Zomorodi?

  Projects like a self-driving car often start from a seed of an idea that seems impossible. And that is the foundation on which X's moonshot factory was built. Astro and his team aimed to tackle some of the world's biggest problems by treating risk and failure not as obstacles, but as steps toward innovation. So today, we're spending the hour with Astro Teller to explore the highs and lows of many of the moonshots at X, technologies that are ahead of their time, some which succeeded, and some that have failed, but helped spark other groundbreaking projects.

• 6:50Manoush Zomorodi?

  To start, let's talk about what a moonshot even is.

• 6:57Speaker 07

  A moonshot, at least from our perspective, has to have three basic components. First, there has to be a huge problem with the world that you can name and want to solve. Second, you have to have a radical proposed solution that we can pre-agree it would very likely solve that huge problem with the world. And then, three, there has to be some kind of breakthrough technology that even if it's just gives us a glimmer of a hope that we could make that radical proposed solution, it's a testable hypothesis. You have a proposed way of starting. So, once we have those three things, we would call that a moonshot story hypothesis.

• 7:41Speaker 07

  And that's not the end of the journey. That's just permission to begin the journey. And one of the failure modes for anyone who's ever tried to get an innovation factory going is that as soon as you have the first thing that's going really well, it becomes rational in the short term for the leaders of the innovation effort to start focusing more and more of their energy on the thing that's going really well. And so, we've set up a process called graduation. As soon as something could live on its own, it needs to not be here anymore. Once it's starting to scale, we want it to leave.

• 8:25Manoush Zomorodi?

  So, I guess it's like almost saying like, you guys are the test kitchen. You're finding the best recipes. But let's say you land on the ultimate cupcake. You're not going to start a cupcake factory. The cupcake people are going to leave. They're going to figure out how to make lots of cupcakes, sell them, market them, et cetera. But you guys are going to go back and figure out another recipe.

• 8:46Speaker 07

  Exactly right. Our job is not to scale things. You're exactly right that we're a test kitchen. But our job is not just to prove the technical feasibility of something, but to try to ask and answer all of the questions, to burn down all of the major risks around whether this really could be a once in a generation opportunity for the world. We use the word "moonshots" to remind us to keep our visions big, to keep dreaming.

• 9:16Manoush Zomorodi?

  Here's Astro Teller on the TED stage.

• 9:18Speaker 07

  And we use the word "factory" to remind ourselves that we want to have concrete visions, concrete plans to make them real. But I have a secret for you. The moonshot factory is a messy place. But rather than avoid the mess, pretend it's not there. We've tried to make that our strength. We spend most of our time breaking things and trying to prove that we're wrong. That's it. That's the secret. Run at all the hardest parts of the problem first. Get excited and cheer. Hey, how are we going to kill our project today?

• 10:00Speaker 07

  We've got this interesting balance going where we allow our unchecked optimism to fuel our visions, but then we also harness enthusiastic skepticism to breathe life, breathe reality into those visions.

• 10:17Manoush Zomorodi?

  How do you manage that financially? If you, you know, I'm sure someone is like, if we throw this much money, this much time, and this many people at this problem, I know we can crack this. Is that enough? Or would you hate to hear that argument?

• 10:34Speaker 07

  I would hate to hear that argument.

• 10:36Manoush Zomorodi?

  Okay.

• 10:37Speaker 07

  Our goal is not radical innovation. Our goal is to pursue radical innovation efficiently. We're trying to systematize the process. So the question we have to ask over and over and over again is what is the reward/risk ratio for each of the ideas that we're considering? We're trying to burn down risk and discover which 1% or 2% of the things that we look at actually could be very large enduring businesses that were, turn out to be really great for the world.

• 11:13Manoush Zomorodi?

  So let's go back to Waymo, the self-driving car. There must have been haha bumps in the road. But like, talk me through like, the process. How do you get from an idea, a race in the desert, to being incubated in the Moonshot factory? And then actually getting to the point that you haven't been killed, that you meet all the criteria?

• 11:36Speaker 07

  So by 2010, we had a car. It was a Prius that could drive itself. It could turn its own wheel with a little motor. And it could apply the brake pedal and the gas pedal. And the first big thing that we did was we said to the team, here are 10 different 100-mile stretches of road in the Greater Bay Area. And we want you to drive each of these 100-mile stretches, at least once each, with nobody touching the gas pedal, the brake pedal, or the steering wheel. You can try as many times as you want. You have to be really safe about it.

• 12:21Speaker 07

  But you tell us when you've accomplished these 10 different 100-mile stretches. And it took the team about a year to accomplish that.

• 12:31Manoush Zomorodi?

  Uh huh.

• 12:32Speaker 07

  And we learned a ton. And the fact that they accomplished it was a real moment for us to go, oh, we might be the right amount too early here. Because we got excited about it and burned down some of the risk, which then gets other people to believe in something that at first, when we started, there was no one else in the world working on that seriously as a potential business.

• 12:57Manoush Zomorodi?

  In a minute, the future of self-driving cars and the origin story of another one of X's moonshots that didn't go quite as well: Google Glass. On the show today, an hour with the CEO of the moonshot factory called X, Astro Teller. I'm Manoush Zamirodi, and you're listening to the TED Radio Hour from NPR. We'll be right back.

• 13:30Speaker 08

  Stay tuned for the following announcements and messages from our sponsors.

• 13:34Speaker 09

  This message comes from ProLon with their five-day fasting-mimicking diet, a clinically-developed nutrition program with the goal of promoting fat loss while protecting lean body mass. Learn more at prolonlife.com/NPR.

• 13:47Speaker 08

  This message comes from Google Workspace with Gemini, one subscription that does it all. Personalized AI and premium business versions of Gmail, Google Docs, Drive, and more. Try it free for 14 days at goo.gle/workspacewithgemini. This message comes from TED Talks Daily, the podcast that brings you a new idea every day. Learn what's transforming humanity, from balancing AI in your critical thinking to surpassing discoveries about the adolescent brain. Find TED Talks Daily wherever you listen.

• 14:22Manoush Zomorodi?

  A quick note before we get back to the show. Woo-wee. Boy, listeners had a lot of opinions on Victor Rippa-Belly, the tech CEO who we interviewed for part one of our series called Are the Kids All Right? This was a conversation about AI in the classroom. People were fascinated. They were freaked out. They were also outraged. We are so glad this episode sparked a lot of debate and conversation in Spotify, on Facebook. But also, if you want to dig deeper into some of the ideas that Victor and I talked about, like what role should governments play in managing AI? What ethical responsibilities

• 15:04Manoush Zomorodi?

  do AI CEOs have? Listen to our TED Radio Hour Plus episode this week. We get into all of it. We would also love you, of course, to become a Plus subscriber and support this show and public radio generally. Check it out. Keep the thoughts coming, and thanks.

• 15:26Manoush Zomorodi?

  It's the TED Radio Hour from NPR. I'm Manoush Zomorodi. Today, we're spending the hour with Astro Teller. He is an entrepreneur, computer scientist, and the CEO of X, Alphabet's so-called "moonshot" factory. With all of their projects, there are, of course, technical challenges to navigate. But there are also other big hurdles, like laws and regulations, and whether the public even wants their new technology.

• 15:56Manoush Zomorodi?

  I mean, I remember there was a period where a lot of the conversation about self-driving cars, it really wasn't about, oh, my gosh, they're on the road. It was very philosophical. It was the trolley question. How will a self-driving car decide if it's faced with having to go around a crowd, killing potentially two people versus five people? It was all very theoretical. And now, fast forward to today, and Waymo is operating in multiple cities. Where do you think the public acceptance is, though? Are people ready for this? Are they ready for it in certain places?

• 16:35Speaker 07

  I think the public acceptance in cities where Waymo isn't yet is all over the map. There are cities that are really excited to have Waymo come. I'm sure there are people who are nervous because they have never experienced a self-driving car. What I will tell you is that in places where there are a lot of self-driving cars, like San Francisco, or Austin, Phoenix, Los Angeles, I've watched people who would never jaywalk across the street because they don't trust the humans to stop, jaywalk in front of self-driving

• 17:15Speaker 07

  cars because they're that sure that the self-driving car will do the right thing.

• 17:21Manoush Zomorodi?

  Oh, wow.

• 17:22Speaker 07

  Which I actually think, you know, we need to think about as a society. How will trust that high start to change how we use our cities?

• 17:32Manoush Zomorodi?

  I mean, let's talk about some of the negative pushback. So earlier in 2025, there were those protests over US immigration enforcement in Los Angeles, and several Waymo cars were vandalized, they were set on fire. And the sense was that those cars were sort of stand-ins for frustration with automation, with corporate power, with surveillance. That must have been really hard for you to watch.

• 17:58Speaker 07

  It was. But I would suggest that we have some sizable social and civic challenges as a society. And ultimately, we're going to need to address how our society functions through public policy decisions. So I acknowledge that people are potentially scared and frustrated about the future. But I don't believe that those fears and that frustration is really about self-driving cars. I think it's a much broader set of things, and I'm not minimizing them at all. But I think that the solutions to those problems are fundamentally decisions about

• 18:42Speaker 07

  how we're going to organize our society and asking some corporation to sort of tweak its technology in one direction or another, I don't think is going to solve those problems.

• 18:52Manoush Zomorodi?

  So for people who maybe are like, well, what do you mean? What is that vision for a world where autonomous cars are the norm and it's a good thing for everyone? How could it change safety or traffic, or I guess even the way we design cities?

• 19:09Speaker 07

  Well, so Waymo has been inundated with requests for people to be able to put their kids into the self-driving cars to send their kids to various places in a way that those parents would never have been OK to put their kid, this is 14 and older for right now, into an Uber or a Lyft car. With a human driver who they didn't know. So that's an example where there's an affordance because there isn't a driver that's making the parents able to get their kids where they need to go, feel safe about their kids, and then they can do something else with their time if they weren't able to drive their kids at that time.

• 19:53Speaker 07

  So I think then the positive knock on effects of these mini buses that can stop anywhere and go to anywhere in a city, I'm hoping will have a really positive and profound change on how cities function.

• 20:11Manoush Zomorodi?

  I would love to talk about one of my favorite topics, which is technology that we have on our bodies and that enhances potentially our human functions. Can we talk about Google Glass?

• 20:27Speaker 07

  For sure.

• 20:28Manoush Zomorodi?

  Okay, because that has had a very interesting trajectory. Just to remind people, 2013, Google launched Glass.

• 20:37Speaker 10

  We're going to do something pretty magical here. So tell me now, who wants to see a demo of Glass?

• 20:43Manoush Zomorodi?

  This was glasses that had like a little display above one of the eyes.

• 20:48Speaker 11

  The wearable technology contains a battery, tiny computer, camera, and a wireless link, and can broadcast whatever the user sees.

• 20:55Manoush Zomorodi?

  It could take photos, it could send messages, it could connect to your phone, all hands free. When, what was the big problem you were trying to solve then, at that point?

• 21:06Speaker 07

  We spend so much of our lives in the digital world in a way that feels deeply disconnected from our lives in the physical world. And that's a function of the user interfaces that we have with our computers and with our phones. You'd be surprised. Being on your phone actually takes about 10 or 12 seconds just to sort of like get it out, unlock it, go to the app that you want. If you have glasses and it's, for example, able if you just say, "Take a picture," and you get the picture, that was a one-second activity, not a 12-second activity.

• 21:46Speaker 07

  It allows you to get much more natural pictures and to get them in a much more natural way. There doesn't have to be such a large schism between our experience of our lives in the physical world and our experience of our lives in the digital world.

• 22:00Manoush Zomorodi?

  So, so remind us what happened. Why, why didn't it take off? Because when you put it that way, like, not looking at our phones, I'm thinking of my aching shoulders and neck and all the hunched teenagers that wander around my neighborhood. That sounds like a great idea. What happened?

• 22:18Speaker 07

  I'm sure it wasn't just these two things, but I'll tell you what I believe the top two things were.

• 22:23Manoush Zomorodi?

  Okay.

• 22:24Speaker 07

  The first is, there was a camera on Google Glass, and I think society was still getting used to the idea that there were a lot of cameras everywhere.

• 22:37Manoush Zomorodi?

  Hmm.

• 22:38Speaker 07

  And that was something that Google Glass bore the brunt of, and society's moved forward in the last 12, 13 years, and I think we feel very differently about the idea that there are phones and cameras kind of constantly around us.

• 22:55Manoush Zomorodi?

  Hmm.

• 22:56Speaker 07

  The other issue was, we really designed Glass to be a learning platform. We knew in the early days of Glass that it wasn't a product. We were so clear on that that we started what was called the Explorer Program, and we were giving it to people who were explicitly signing up to give us feedback about what this wanted to become as a product. But along the way, people started taking it very seriously as a status symbol and wearing it as a status symbol.

• 23:34Speaker 12

  Once you put on Google Glass, it's suddenly just a part of you.

• 23:37Speaker 13

  All right, Google is telling Google Glass users, "Ask for permission when using Glass to take pictures of others." That can get kind of creepy. Also...

• 23:47Speaker 07

  And we let ourselves fall into an understanding of Glass as being a product when it really wasn't yet.

• 23:58Manoush Zomorodi?

  So when you say it wasn't a product, do you mean it was more like a prototype or something for people to test?

• 24:04Speaker 07

  From a design perspective, it was incredibly solid and polished. But a product isn't just how well-designed the industrial design is. A product is also a deep understanding of what value you're delivering to whoever's receiving it and what they expected to get from it. And we didn't know what it was good for yet. So for example, we were giving it to and working with independent filmmakers. And we weren't telling them, this is definitely how you should make cinema verite going forward. But we were asking, is this useful? What would you do with this?

• 24:49Speaker 07

  We suspected that a lot of kind of prosumer activity would happen with glass. So if you were a waiter at a restaurant or the maître d' at a restaurant or a nurse in a hospital, could glass be useful? Yeah, almost certainly. But how would you use that? We didn't know. So that was the point of the Explorer program, was getting it into the hands of these people so they could teach us what they wish it would do so that we could prototype the applications essentially for glass that sort of met them where they were.

• 25:25Manoush Zomorodi?

  But it did get used, didn't it? Like, because I remember hearing that it was being used to train pilots and surgeons and that it had sort of an industrial use case that was very successful. Is that wrong?

• 25:37Speaker 07

  You're right. We had focused too much on consumers, whereas very quietly in the background, it was actually people on like oil rigs, people who were maintaining airplanes, you know, nurses in hospitals who were the ones who were like, please don't take this away from me. It's like, I need it for my job now. And so, over time, we pivoted to be where people were finding use in it.

• 26:04Manoush Zomorodi?

  Would you say that this was a device that you didn't get the timing right? That you were too early?

• 26:11Speaker 07

  Yes, we were too, too early in that case. We learned a lot from it and Google, through Android XR and working with Warby Parker and Samsung, is now announced that we're doing a newer version of this. So its time has come again. I think that the timing is right now, but that's 12 years later than you pointed out. 2013 was kind of a while ago, so we were too, too early.

• 26:41Manoush Zomorodi?

  So, so as you mentioned, now Alphabet is working with Warby Parker and some other companies on smart glasses with features that have like live translation and travel guide functions. And you have competition. This is a market where Ray-Ban Meta Glasses are offering similar tools. How do you feel gun shy going into this? Or are you like, oh, we've been here. We know how to go forward.

• 27:09Speaker 07

  I think that the time is, is now. I think people are excited for these values. I think they're comfortable with these form factors and the technology has moved forward. The light basically can be projected into the lens of glasses in a way that then reflects into your eye. So it feels much more seamless, much more like a natural pair of glasses than it did before. And that's also helping with consumer acceptance. So I predict that this becomes a significant way that we interact with technology over the next decade.

• 27:43Manoush Zomorodi?

  I mean, it feels like this, this sort of wheel is churning faster and faster. I remember when like people thought you were kind of a jerk for wearing AirPods. Like tech bro, wearing AirPods. And now, I mean, everybody wears AirPods or earbuds or whatever you want to call them.

• 28:03Manoush Zomorodi?

  But there's still the issue of. Fundamental human things like privacy and not knowing when someone can take a photo of you, for example, or can maybe have facial recognition and know your name even though you haven't even spoken to them. There's the cosmetic sort of acceptance and then there are these bigger sort of societal norms and agreements that we make with strangers without even knowing it. Do you think some of those are going to change as well? Or I see like my 15 year old like is very conscientious about asking people before she can take a photo despite the fact that like everyone's taking pictures all the time.

• 28:48Speaker 07

  Right, and I think the problems you just described are very real and we need to work through them as a society, but I don't think they have anything to do with glasses per se. If you're on vacation in Hawaii and you're enjoying your breakfast, there are 30 phones all around you. People could be taking photos of you without asking. You hope that they won't. They could already be running facial recognition on those photos to find out who you are. So we as a society need to work through the things that you just described, but I think that the introduction of glasses is not going to make the problem substantially different than it already is.

• 29:27Manoush Zomorodi?

  So you've seen how ideas gestate through institutions, through generations, but I think that the, am I right in saying that the word that you often use is compost? That ideas go back and sort of, I guess, maybe are part of the soil in which new ideas grow? Is that what you call it?

• 29:52Speaker 07

  Yeah, we call it moonshot compost, but if you worked here at X, one of the things we would need to help you with is how to feel okay about stopping something that you've worked really hard on. That is emotionally challenging for all of us, no matter how intellectually honest we're trying to be.

• 30:15Speaker 07

  And so we have, over time, developed a lot of different ways to help people see past the initial hurt and sadness of having to end something they were really excited about and had worked really hard on. One of the things that we've learned is part of how X works, and it helps people to remind them of this, is to say to you as you're ending something, "We thrive on moonshot compost." What that means is, when you stop your project, that is not a rejection of solving the problem you were really passionate about solving. The problem will stay here with us and will be inspired by that problem still.

• 30:57Speaker 07

  We don't have to lose the software, the hardware, even the partnerships that you built as part of this. It's all going to go back into the metaphorical dirt here at X, and we have lots of really great examples where we were doing something, it didn't quite work out, but after a break, we came back in a different way and ended up with a much better solution. And so we can see the initial roots, the germination of that new idea, in the leftovers from the old idea, which is one of the reasons we celebrate failure the way we do, is because failure is learning.

• 31:42Speaker 07

  And helping remind people about moonshot compost is a way to help them understand that cycle and feel better about it.

• 31:49Manoush Zomorodi?

  Hmm. I am so curious whether that cycle is speeding up because of artificial intelligence. I mean, as I think normal people feel like, oh, my gosh, this artificial intelligence thing is real, but you, of course, have been talking about it for decades. Can you tell us more about, I mean, it's baked into all the projects we've been talking about. Has it sped up this cycle of test and fail, test and fail for you and your team?

• 32:20Speaker 07

  It has. We use artificial intelligence in almost everything that we build, but we also use aspects of artificial intelligence and machine learning as tools in helping us to do what we do faster and faster. So yes, that's true as well. Because a lot of the things that we do here at X, not all of them, but a lot of them, have some aspect of getting in contact with the physical world, the real world, in sort of nontrivial ways. We don't tend to work on things that are just digital. We tend to be a bit rate limited by the physical world, making of physical prototypes,

• 33:02Speaker 07

  getting them into the world, having experiences with them in the world and getting that feedback. So we are moving faster, but maybe less faster than you might imagine because the physical world has its own sort of complexity and demands that artificial intelligence can't entirely resolve.

• 33:22Manoush Zomorodi?

  When we come back, Astro lets us in on some of the latest moonshots at X that aim to solve some of the world's most complex and urgent problems, like climate change. On the show today, the head of X, Alphabet's moonshot factory, Astro Teller. I'm Manoush Zamirodi, and you're listening to the TED Radio Hour from NPR. Back in a minute.

• 33:58Speaker 09

  Stay tuned for the following announcements and messages from our sponsors. This message comes from ProLon with their five-day fasting-mimicking diet, a clinically-developed nutrition program with the goal of promoting fat loss while protecting lean body mass. Learn more at prolonlife.com/NPR.

• 34:15Speaker 08

  This message comes from Google Workspace with Gemini. Ever notice how every new AI tool asks you to explain your business to it all over again? Google Workspace with Gemini skips that. You get Google's AI built right into the premium versions of Gmail, Google Docs, Sheets, and more. No new tools, no extra logins, just one subscription that gives you personalized AI right inside your actual workflow. Try Google Workspace with Gemini at g o o dot g l e slash Workspace with Gemini. This message comes from TED Talks Daily, the podcast that brings you a new idea every day. Learn what's transforming humanity. From balancing AI in your critical thinking to surpassing discoveries about the adolescent brain.

• 34:59Speaker 08

  Find TED Talks Daily wherever you listen.

• 35:03Manoush Zomorodi?

  It's the TED Radio Hour from NPR. I'm Manoush Zomorodi. Today on the show, we're talking to Astro Teller, the CEO of X, Alphabet's so-called moonshot factory. And so far, we've heard the origin stories of two of X's most public moonshots, Waymo, the self-driving car company, and Google Glass. But the technologists at X have worked on countless other projects that most of us have probably never heard of, ones that are working to improve our food systems.

• 35:34Speaker 11

  So the first initiative we did was to take our robots to the seed banks.

• 35:38Manoush Zomorodi?

  Others that are trying to protect our oceans.

• 35:41Speaker 11

  What we do is we put cameras underwater. We put sensors around and we're able to.

• 35:45Manoush Zomorodi?

  And still more, working to connect some of the most remote areas in the world to the internet.

• 35:51Speaker 14

  The authorities of the two countries that we had connected, they found it incredible to believe.

• 35:58Manoush Zomorodi?

  One of the current projects in the works is called Tapestry.

• 36:02Speaker 07

  We got so many times to the cusp of a moonshot and then realized that the way the electric grid around the world, including in this country, is functioning won't allow for those moonshots in various ways. Let me unpack this some. The grid is the world's largest machine. It's the world's most complex machine, and it's the world's most expensive machine. The people who run the grid around the world are good people. They're smart people. And they kind of barely have a handle on what they're doing because the thing that they're operating

• 36:47Speaker 07

  was cobbled together in any large place, like the United States, over 120 years.

• 36:55Speaker 07

  So if you go to any grid operator and say, show me a map of where every wire is on your grid, every inverter, every transformer, they literally don't have that map.

• 37:09Manoush Zomorodi?

  Mm hmm.

• 37:10Speaker 07

  So in order to help grid operators of the world get to the place where we can get electricity to everyone who needs it cheaply, reliably, effectively, ideally clean, you know, with as low a carbon footprint as possible. We started a moonshot about seven years ago at X called Tapestry to, with the aspiration long term to help make software services that could help them to manage the grid as that they currently have to plan for the future and ultimately to optimize it in real time. And so that our Tapestry moonshot is now helping grid operators in this country, in

• 37:52Speaker 07

  Chile, in New Zealand, we're starting in the UK, we're starting in South Africa, just starting in Australia. So we're really proud and excited about the help that that moonshot can bring to the world at a time when the grid is groaning under the weight of change and we depend on it. Humanity depends on it for how we grow our food, how we sort of take care of people in hospitals, how we charge our cars increasingly and hundreds of other things.

• 38:26Manoush Zomorodi?

  So what does that look like if, if this gets a successful rollout to all these places and people who work on the grid have more knowledge, I guess, of what exactly is operating and, and I, and I suppose can back each other up if it comes to, I mean, we hear here in the United States about entire counties going dark because they're not connected to another grid. What does it look like if this is a success?

• 38:54Speaker 07

  Yeah, let me give you a really concrete example. There are weirdly long wait times to plug almost anything into the grid.

• 39:03Speaker 07

  If you have a wind farm, a solar field, a data center, almost anything, and you want to plug it into the grid, it's both a legal and moral obligation on the part of the grid operators to have their very smartest people pour over lots of future what ifs and try to imagine in all kinds of weird scenarios, would plugging this thing into the grid, let's say it's a solar field, would that hurt the grid in any way or cause our grid to go dark? These are very complex machines.

• 39:37Manoush Zomorodi?

  Yeah.

• 39:38Speaker 07

  The problem is, if it takes you a month to figure out whether it's safe to plug that solar field into the grid, in the month you've spent thinking about that, five more solar fields got onto the back of the line.

• 39:53Speaker 07

  And so when you hear, this is called the interconnect queue in for any grid operator in the world. And the wait times vary from about five years to sometimes more than 10 years, which is causing it to be financially a nonstarter to actually build anything if you're going to have to wait in lines that long to be allowed to plug into the grid. So this is an example where everything is sort of broken down. And we've spent several years with the country of Chile and their national grid, which is called CEN, C E N. And after having tested the tapestry software for a number of years, they have now made it the sort of way that they're processing their,

• 40:37Speaker 07

  among other things, their interconnect queue. And they've said that we've sped up the way that they can safely check

• 40:47Speaker 07

  whether something can be plugged into the grid, where they believe it's both more safe because we're running a lot more scenarios for them, and it is happening in one thirtieth the time. So in a day instead of a month. And that is transforming. The country of Chile's ability to get things off the queue and to start to shrink that interconnect queue. So that is one very important, but very real, concrete way that the grid is struggling. And with the use of the right new software, it could be performing in a much better way.

• 41:23Manoush Zomorodi?

  So just to lay it out, you're talking about transitioning entire countries to solar, to wind power, to renewables faster.

• 41:32Speaker 07

  I mean, yes, that also can be part of it. But to be clear, if you wanted to plug a coal-fired power plant onto the grid, or if you wanted to plug an aluminum smelter onto the grid, you have to wait in that same line. Even like a large building like a hospital has to wait in that same line.

• 41:50Manoush Zomorodi?

  Can I ask you, though, about climate? Because, you know, if we talk about the moonshot criteria, big problem, radical solution, plausible path, climate, like, that would definitely, like, be at the top of the list as something you would want to tackle. Is that, are you thinking of ways you can make a dent in global warming in the next decade with your moonshots?

• 42:14Speaker 07

  For sure.

• 42:16Speaker 07

  The way that, you know, humanity is affecting the Earth and the way that those changes in the Earth are affecting us is real and it's going to continue. So, basic observation. Right now, humanity makes, sort of depending on how you count, about $10 trillion of stuff every year. And we use it to varying degrees, sometimes for, you know, an afternoon, sometimes for 50 years, like a building. And then it goes into landfill. Almost all the stuff we made, the physical stuff we make, eventually makes it into landfill.

• 42:57Speaker 07

  But the stuff that goes to landfill has already been so processed.

• 43:03Speaker 07

  There's a lot of embodied value still left. We just don't know how to get it out. So, rough order of magnitude, think about a plastic bottle that can't, for whatever reason, be recycled. Spoiler alert, 90% of what we put into the blue bins goes to landfill. It does not go to a recycler.

• 43:25Manoush Zomorodi?

  I know, makes me cry, Astro.

• 43:27Speaker 07

  I know. It's, it's, yes, and we as humanity can do better. What if we saw this process of every year, taking $10 trillion of stuff, kind of getting some value out of it, and then putting about $5 trillion of remaining value into landfill. See, those landfills is the world's greatest resource. If we could only get the rest of that embodied value back out again, we could be a real circular economy. So, we have built a system here at X, a moonshot for circularity. We aspire to do this with things like concrete and textiles and, and electronic waste,

• 44:11Speaker 07

  but it happens that we're starting with plastic. And we actually can, at about 10 miles an hour, like on a fast conveyor belt, look at everything going by and know what its molecular makeup is, so that we can route each of these things to the kind of recycling that will allow us to get it back down to the molecules in the right way. It turns out there's a lot of different recycling. You can mechanically recycle things, kind of chew them up with big metal teeth. You can heat them to sort of melt them. There's other kinds of chemical recycling processes. And actually, you can have little buggies eat them, biological recycling of various kinds.

• 44:55Speaker 07

  But if you don't know what's in it, you can't actually send them to the right place or parameterize those recycling situations to get the best out of the things that you're sending into those different areas. So this project, which is called Materia, is, has made a lot of progress, is working with some companies around the world. We actually had a big party late last year. They filled a tanker truck full of sort of virgin oil that can be put back into the making of plastic bottles entirely from plastic that had been rejected by recyclers and was headed to landfill. That was a really nice moment for us.

• 45:40Manoush Zomorodi?

  So it does seem to sort of perfectly embody this idea of like, it will scale if it can make money and is also a good thing for the planet. I mean, this could change the entire sort of ecosystem of how we use materials and get rid of materials.

• 45:58Speaker 07

  Exactly. And so

• 45:59Manoush Zomorodi?

  When will we know, Astro? When will we know if you can do this?

• 46:03Speaker 07

  Well, I mean, we know we can do it to some extent. And these are long journeys. When you look at something like Google Brain that came from X, you know, that's having its moment in the sun now, that was sort of, depending on how you count, almost 15 years that sort of caused the, or was one of the things that caused the modern explosion of machine learning in society. Waymo, which we talked about, that was 15 years ago. Wing, the drones for package delivery, they're now doing a lot of deliveries in a range of places around the world. They're scaling quickly. And that was, they started about 12

• 46:47Speaker 07

  years ago. So if you look at something like Materia, which is about five years old, I would be very hopeful that seven years from now, we all take it for granted that most of our plastic does get recycled, or certainly will soon get to be recycled. But part of radical innovation is it takes time. I wish I could make it differently, but that's just how it works.

• 47:13Manoush Zomorodi?

  I wonder, do you attract, do you think, a different type of person to come work for you? I mean, we're reading these days about these huge financial packages being offered by Mark Zuckerberg to AI software engineers, that this is like, sort of, there's a hiring arms race going on. Does it change how you do business and how you attract the sort of most creative technical minds to the Moonshot Factory?

• 47:40Speaker 07

  Not much, honestly. I think the people who want to work at a Moonshot Factory are a bit of a different breed. If you're looking for a highly stable environment, this is the wrong place to come. People need to be very comfortable with ambiguity and a high rate of change in order to be comfortable at X. We actually joke that we're chaos pilots. If you want to do that hard, scary explorer work to try to make the next swarm with us, then this place might be for you. It's just a very different person who wants that kind of a thing.

• 48:20Manoush Zomorodi?

  You are kind of a mythic figure in Silicon Valley. I mean, on the TV show, you're a character on there as well. But, you know, this idea that you're on your, like, roller blades, there was one profile that was written, I think it was on Halloween, and you were dressed as Gandalf, sort of gliding around the place. Like, you've been around for a while. You represent, sort of, I think, hard, technical knowledge combined with real whimsy? Is that a fair assessment?

• 48:51Speaker 07

  It is. I mean, I think there's two separate things going on. I would never do something that was inauthentic. So I'm showing up as a real version of myself, partly because I want to encourage everyone else here to show up as a real version of themselves. We're not going to be an extraordinary group of explorers who can have great psychological safety and whose ideas can clash with each other in really interesting ways, where the sparks that come from those clashes turn out to be the real gold dust that we value. If we aren't all, you know, at 100%. But I am also sending a signal to people not to take me too seriously,

• 49:35Speaker 07

  not to take themselves too seriously. And then it's possible to be intensely serious about what we're trying to accomplish in the long run for humanity while not doing it in a way that's so serious, we can't find those unexpected gems. No one ever created something really unique and unusual while gridding out determinedly the process. Humor and silliness are actually very close to the well springs of creativity, not kind of grim determination. So when I show up as Gandalf, I mean, I like Gandalf.

• 50:19Speaker 07

  He's one of the people I want to be when I grow up. But it's also my way of trying to disarm people and remind them that they have so much inside them that they're not letting out and that this is a safe place to do that. Because if I want you to say that one in a hundred things that's really brilliant and everyone didn't see coming and they're like, what? You have to feel equally safe saying 99 things that are just wrong. And that has to not just be barely okay with people. It has to delight them.

• 50:55Manoush Zomorodi?

  You make it sound so easy, Astro.

• 50:57Speaker 07

  Well, like I said, the facts, what we have to work towards, the facts are really easy. It is really simple. It's actually operationalizing it. You don't have to operationalize it with Gandalf costumes and roller blades. But one way or another, you have to send a thousand signals to people that you're serious about what you want from them. The unlearning that you're asking from your team. You have to make them feel safe and like it's not a dumb thing for them to do, to show up in the ways that you're saying you want them to show up. That's the hard part.

• 51:36Manoush Zomorodi?

  That was Astro Teller. He is the co-founder and CEO of X, Alphabet's moonshot factory. If you'd like to learn more about their projects, check out the moonshot podcast that Astro hosts himself. You heard clips from it earlier in the show. To see Astro Teller's full talk, just go to ted.com.

• 52:02Manoush Zomorodi?

  Thank you so much for listening to our show today. If you liked it, got something out of it, please leave us a comment on Spotify or email us at [email protected]. We read every comment and email, and we love hearing from you. This episode was produced by Katie Monteleone and edited by Sanaz Meshkhanpour and me. Our production staff at NPR also includes James Delahoussaye, Rachel Faulkner White, Matthew Cloutier, Fiona Guerin, and Harsha Nahada. Our executive producer is Irene Noguchi. Our audio engineer was Jimmy Keeley. Our theme music was written by Ramtin Arablouei. Our partners at TED are Chris Anderson, Roxane Hylash, and Daniela Bella Rezo.

• 52:45Manoush Zomorodi?

  I'm Manoush Zamirodi, and you've been listening to the TED Radio Hour from NPR.

• 52:52Speaker 08

  Stay tuned for the following announcements and messages from our sponsors.