Alphabet X's CEO Has A Vision For AI Moonshots — With Astro Teller

28:41So now Google has its own. It has BARD. but I almost wonder when we think about insulating Google from the innovators dilemma, shouldn't X have been front and center building the first chat GPT and not letting an open AI, for instance, run away with it? I swear he's not a plant, but good setup. I think you say that we shouldn't be afraid of the advancement of knowledge. Yeah, absolutely, absolutely. So, no, no, I love the question. Let me take a step back and remind you sort of how X functions, what our goal is. Our goal is to invent and launch moonshot technologies that, if we do it right, help tackle some of the biggest problems in the world and lay the foundations for enduring sustainable businesses.

29:35one of the early things that we did was a thing which when we graduated it back to google was called google brain google brain is the origin of much of what you now think of as bard and so that was an example in the earlier days of x of us making something realizing that it would be productive for that to be back in the mothership. We moved that back to Google, and it has gone on to do amazing things. And so the transformer paper and other things that you've heard about. Somewhat more recently, this is about five years ago, we said, we can feel on the horizon, we're going to get to the place where the ability is going to be there for us to be working in much tighter loops with software developers, like a partner to them, where we can complete code when they start to write it.

30:34If they write out an English language description of what they want, it will write out the code for them. It can even be like a pairs programmer and have a conversation with them and help them to be better programmers. And so that work happened for about five years at x about seven eight months ago we moved it back to google and it was just announced recently as cody and that is now sort of the code making part of what you think of at google ahoy that was the timing of that was so freaking perfect just finished the answer and then there we go familiarize yourself with the location and receive important safety information from our crew

31:25So the real question is, is he going to delete this from the podcast or is he going to leave it in? What do you think? Should he leave it in? Thanks, everybody.

31:40That's right. Can someone shut this off?

31:48That was required by law. One of our jokes at X is after we make time machines work and cure cancer, we're going to figure out how to get AV sorted out. That'll be the hardest thing after we do the easy stuff like cancer and time travel. So I want to ask a follow-up question, which is you mentioned that brain started in X and then graduated to Google. I want to question the graduation process because when you do graduate to Google, doesn't it become a little bit harder to say, we've made all this amazing technology. Now I'm going to make something that's competitive with the bread and butter. For instance, a chatbot that can sort of serve as something that people might want to use instead of search.

32:41And once it goes into Google, doesn't it have the constraints that it might not have an X, which is that now all of a sudden it has to worry about the quality of search and the quality of response where it might not in a more experimental unit.

32:57Ramping anything up to serve more than a billion people a day is super complex. We're talking about, you know, hundreds plus languages. there is a lot technically legally ethically just to do this i mean practically at all but also doing it ethically that x is totally not set up to do for the same reasons that x may be a particularly good place to do rapid prototyping and learning of new things we are not the right place to move something from yeah we have something special here to it's ready in a really thoughtful responsible way to serve a billion people a day. So I hear you, but I would actually say that that would have been irresponsible for us to try to keep that to ourselves.

33:44The goal of X is to create really good seed crystals. We don't want to do it all ourselves. We want to get the ball rolling where either sometimes the rest of the world in the form of an other bet at Alphabet like Waymo is, or back at Google, the world says, ah, got it now. See why that could be really big before it's done, but it can survive on its own because it's left behind that, like, what are you talking about? Why would we do that? That's definitely not possible. When those things are still on the table, that's what X should be for, is to actually work through and often be wrong about whether there's something awesome there.

34:25And so we try 100 things and 99 of them don't work out. And our job is to be wrong about those 99 as efficiently as possible, to move past the ones where we're wrong about them with evidence as quickly as we can, so that that one, which we can double down on over time, can sort of go on to have a really productive future. So here's my uneducated argument to switch things up. Yes, it makes sense to have a prototyping lab, but also couldn't you, for instance, a billion users, that would be building something like BARD into Google Chat. But OpenAI's ChatGPT started with 100 million users. That's the fastest growing consumer product in history.

35:10It seems like you might even be in a great position to say, hey, we have this chat thing, keep developing it with ThinX, and then call Sundar up and be like, hey, man, can we have some of that cloud infrastructure? The same way that OpenAI called Saly up and said, can we have some? And that's sort of what's enabled them to power ChatGPT. And then you don't really have to worry about the competitive elements of coming up against the core business. Let me use a related example, see if this at least partially satisfies you. So one of the things that we've worked on at X for a very long time is factory automation.

35:48The world spends sort of pushing$10 trillion a year making stuff. And if you can automate the making of stuff, you can do it faster, you can do it cheaper, you actually have less waste, it's more reliable because it's been automated, but it is fiendishly complex and expensive to automate the making of almost anything because until very, very recently, there was a lot of bespokeness, artisanalness to programming robots. And so X worked on that for many years, created, it was a project called Intrinsic. It graduated. It's now an other bet at Alphabet. They just announced a few days ago that they have a platform called FlowState, which is a developer's arena in which you can develop new capabilities and new skills for robots to have with a seamless ramp from early investigations in simulation to actually landing them in robots in factories doing real work.

36:59And that creates more jobs for developers, makes it easier for people all over the world to help make robots more capable and it helps the makers of the world and it democratizes the making of things because right now it costs a billion dollars to seriously automate a factory so that's an example that started at x it's not exactly a chat bot but it is you know making robots much more capable much more flexible much more nuanced and dynamic in their ability to solve problems on the fly. And then we've laid out the early infrastructure, but are allowing the rest of the world to build on top of that, and which I think is going to be really good, yes, for Alphabet, but for the whole world as well.

37:45So that's something that didn't go to Google that lives in this region post X, but still inside Alphabet called an other bet. So do you want brain back? What was that? Do you want Google brain back? No, no, no, no. Our job is to catch new waves for Alphabet, to catch new waves for the world. What excites us isn't sort of empire building or getting to sort of like have been the one who rang the bell. That is not our goal. We want, in a really efficient way, keeping this balance of audacity and humility just right. audacious enough that we'll try almost anything but then humble enough that we're honest right after we start out on each of these investigations that we're probably wrong and that we need to spend our money figuring out that we're wrong rather than trying to prove that we're right since mostly we are wrong and by doing that in lots of different domains what rises to the surface are these new things like inverse design or generative design that we were talking about It's not that I said there will be generative design seven or eight years ago.

38:58It's that we've tried a thousand things over the last decade. And the generative design stuff is some of the stuff that's bubbling to the surface in a way that's looking really promising over and over again. That's what excites us is catching those new waves, doing it efficiently and doing it responsibly. Astro Teller is here with us. He's the captain of Moonshots at Alphabet and the CEO of X. We'll be back right after this to talk a little bit more about how Alphabet and X actually create the products they do. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one.

39:37It's called Chat Concierge, and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, It doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. These days, it feels like every dollar should be working a little harder, but figuring out where to put your cash can be confusing. That's where Wealthfront comes in. Wealthfront is a tech-driven financial platform built to help you grow your savings into long-term wealth.

40:14Their high-yield cash account through Program Banks offers a 3.5 % APY on your uninvested cash as of November 7, 2025. And there are no monthly fees, no minimum or maximum balance to earn that rate, and you can even make free instant withdrawals to eligible accounts in just minutes, 24-7. So your money can always be within reach. Right now, Wealthfront is offering new clients an extra 0.65 % APY over the base rate for three months on up to a$150 ,000 balance. That's a total of 4.15 % variable APY when you open your first cash account. Go to Wealthfront.com slash Big Tech to sign up today. This is a paid testimonial for Wealthfront.

40:53It may not reflect the experience of others, and there's no guarantee of future performance or success. Wealthfront brokerage is not a bank. Rate is subject to change. Promo terms and conditions apply. For more information, see the episode description.

41:08And we're back here on Big Technology Podcast with Astro Teller. He's the captain of Moonshots at Alphabet, CEO of X. So Astro, I'm actually curious how you decide what to fund inside of X. Because, you know, I might think of this as like you need to predict the future, but that's not the way that you think about it. Exactly. One of the core axioms at X is I do not believe that anyone, certainly including myself, is any better than random at predicting the future. I know that that's like not the cool thing to say in Silicon Valley, but I just don't believe that any of us are better than random at predicting the future.

41:47I think we can discover the future a good bit more efficiently than maybe other people discover it, but that's very different. So the core kind of map to beginning a journey at X is we make these three circles and we talk about X efforts living at the intersection of those three circles. The first one is there has to be a huge problem with the world that you want to solve. If you're proposing something for X, you have to tell us what that huge problem with the world is you want to solve. If you can't say that, we're not starting on the journey. Number two, there has to be some radical proposed solution for how to fix that problem, some science fiction sounding product or service, however unlikely it is that we could make it, that if we made it, would make that huge problem go away.

42:38And then there has to be some kind of core technology opportunity, some breakthrough technology that gives us the opportunity to start on that quest. And those three things together are a moonshot story hypothesis. That doesn't mean you're right. In fact, you're almost certainly wrong. But if you can propose those three things, it has the form of a moonshot. And then the answer is, great, gorgeous moonshot story hypothesis. What is the smallest amount of money and shortest amount of time you think it'll take to kill your idea? Because your idea has a 99 % chance of being wrong. And there's no way to avoid that because if that wasn't true, it wouldn't be radical.

43:24We're only interested in the over-the-horizon stuff. And so as soon as we sign up for that, we are explicitly signing up for being wrong most of the time. And that's why the humility has to kick in. So it doesn't work like we will work in these areas, we won't work in those areas. It mostly is what are huge problems with the world? So maybe not coincidentally, more than half of what we're doing right now is in the climate change space, but not because I mandated that, but because that's what people were excited about. And that is legitimately some of the biggest problems the world is having right now.

44:00And then it's actually evidence that kills off most of these things over time. And the ones that survive that we double down on. So it's that winnowing process that looks at the end like we planned each of these things, but we didn't plan these waves to catch. It was just a filtering process that weeded out all the ideas who are just bad sometimes. But more often, we're like, just wrong time, technology wasn't right. Or maybe even the technology was right, the timing was right, and we just couldn't figure out how to make it great enough, which happens sometimes. It's a very interesting process.

44:37I'm curious, most people who are listening to this are not in the position to take moonshots. You must hear this all the time. Is there anything from your process they could put into place inside their companies that might help them achieve those 10x moments that you're going for. For sure. I would actually argue what I just described is the most efficient you can be in trying to find something new. And even if you've pre-decided, you know, you're working on flying cars or whatever that is, inside of your effort, there are lots of things that you might be thinking, should it be gas powered? Should it be batteries?

45:15How many propellers are we going to have. There are going to be lots of things for you to figure out over time. What should the airframe be shaped like? And so there are lots of risks to take, lots of discoveries to have. And for each of them, you can do exactly what I just said. So one of the things that we say at X is, if you're starting out on a journey, I know it's no fun to kill your ideas. I get that. But let's say, for argument's sake, that the idea you have isn't going to make it. Would you like to find out now for one dollar or find out three years for now for like 10 million dollars and everyone of course says well I guess I'd like to find out now for a dollar great or welcome to x thank you for working on the propellers or the time machine or whatever it is you're working on how are we going to do this are we going to be intellectually honest or intellectually dishonest about our discovery process.

46:11You're like, oh, I guess intellectually honest, Astro. We all know in our hearts what to do. What I just described is not rocket science. It's not like X invented this. That's not what's going on. It's just so hard to do it. All of human nature drives us in the opposite direction from what I'm describing. So what X spends all its time doing is trying to create the infrastructure, the social norms, reward systems, et cetera, so that people actually do these things. But what I've described, for sure, if you work in a startup company, if you work in a huge conglomerate anywhere in the world, this is still the right way to explore new ideas.

46:55Let's check in on a couple of famous ex-product projects. Waymo, right? I think I'm seeing more progress in self-driving cars now than in maybe the past year than because I was in San Francisco 2015, 2016. People were like, this thing is going to happen next year. We're in 2023. Not quite there yet, but it seems like there's a lot of progress being made. What's your perspective on where we are right now in self-driving cars and how far are we from realizing that original vision? For sure, it was harder than it turned out to be. One of the things that I think caught Waymo by surprise, got us by surprise, is that there are a lot of things that human drivers do that if we want to follow the laws, we can't do.

47:43Those humans just kind of like are in the gray area a lot about how they choose to drive their cars, where they pick people up, where they drop them off. If it's like a transportation, it's a service. Obviously Waymo can't do that. That is incremental hard stuff for us to figure out. So there are a lot of edge cases to make sure that this is super safe. It's a fun experience for people. But I also think that the way it tends to work is any exponential ramp up, it's the same ramp up every year, assuming it's a gradual ramp up. But if we're not paying attention to it, it feels like it went like this.

48:26And I think most of you will experience self-driving cars like this. There are three cities in the United States where you can get a ride, you know, with nobody in the front seat in Los Angeles, in Phoenix, and in San Francisco from Waymo. And I don't know exactly when, but I pretty much guarantee there will be more cities over time. And so if you live in one of those cities, it probably feels more real. If you don't, it feels like, nah, it's not going to be a thing. And then all of a sudden it will be a thing. And that's natural because, you know, unless you're really in that industry, you aren't watching really closely as the number of rides per day per city is going up from a group like Waymo.

49:09I'm legitimately pumped to ride in a Waymo for the first time, hopefully this year. Another project that I want to talk to you about is Project Loon, which was, was, is these balloons that are supposed to beam internet down to everyone, no matter where you are. Now, I'm curious, like people now, if they think about accessible internet, they immediately go to Starlink. How did Loon, which was a great idea early on, sort of, I don't know, is it right to say, but I think lose ground to Elon Musk and Starlink? First of all, I mean, so Loon was a goal. Could we make a worldwide stratospheric layer of balloons?

49:48They're like cell towers, but floating at 65 ,000 feet that were talking to each other in an ad hoc mesh network and beaming LTE or 5G to the ground so that people in rural communities around the world, particularly the people who don't have good internet connection, all of a sudden would. Very inspiring, and it took a long time. We built it. It was working. We were beaming LTE and 5G to hundreds of thousands of people in multiple countries. We couldn't figure out how to get the business to close with the owners of the spectrum that we had to work through in those different companies. So we turned it off.

50:29It made us very sad. Starlink is a cool company, but they're solving a different problem. There's a fixed amount of bandwidth you can land when you beam something from a satellite in any one region. So it works very well in rural places, but you couldn't have like a huge number of them parked very close together. So that wasn't really the goal that Loon was trying to solve. But let me tell you a sort of fun side story. It was crushing for all of X for us to kill Loon. It had gone on to be another bet. It was really one of the inspiring things that had come from us. And so it made everyone really sad for us to stop doing it.

51:06And we have this saying at X, we're very focused on moonshot compost. Whenever we stop something, the people, the code, the patents, the know-how, we try to keep them all at X. They recirculate working on new projects. And so in this particular case, one of the technologies that was allowing these balloons at 65 ,000 feet to communicate between the balloons at very high bandwidth were lasers. and so when Loon ended someone said well how come we couldn't put those lasers on the ground you know like on a pole and that sounded almost embarrassingly too simple after all of the work we had done to get them up to 65 ,000 feet but fast forward five years we have these things are about this big.

51:59It's smaller than a traffic light. It shoots a laser up to 20 kilometers. It's eye safe. You could just put your eye right up against it. Nothing bad would happen. It's unregulated. It's near visible light. It's about one order of magnitude outside of visible light. But in the EM spectrum, that's basically visible light. And it's received by another box. It's, you know, two feet tall. You strap them on two poles. If you plug the internet like a fiber optic cable into one, you have 20 gigabits per second up to 20 kilometers away for less than one one thousandth the cost of trenching fiber. And we're rolling these out right now in mainly Africa and India.

52:46And that project is now moving more data to real customers per day than Loon moved in its entire history. So go Moonshot Compost.

53:00That project is called Tara, if you want to look it up. Pretty cool. Okay, let's see. We have maybe two or three questions to get through in nine minutes. Let's see if we can do it. In the break, I asked some people in the audience to shout out different big problems that maybe X could get working on. And we've talked a lot about hardware, for instance. But what about some medical physical issues? Is that ever something that you'd want to, or I don't know, societal issues? Is that ever something that you'd want to take on? So I'll just give one example. So one person in the front here said community.

53:34I mean, we know that community, building community is one of the things that people struggle with the most right now. I mean, we're here with one. So it's like, it's nice to see that it persists. But most people would say, or many people, many more people than usual would say that they're lonely. They don't have friends and stuff like that. Is that something that an ex would ever take on, or is that even too moonshot-y for the moonshot factory? No, no, not at all. As long as we could be proud about the output and there is a technology solution to the problem, we would be excited to work on it.

54:06We have had various explorations, for example, in the social justice space. because the temptation is to think that something like belonging or systematic bias in society is just not amenable to technology helping. And maybe that's true, but that's not written in stone somewhere. Like shame on us for not at least trying. So we don't have some huge news to give you on that front, but we certainly haven't given up. I'll give you another one that's sort of like that. Education. Like, I will consider it a failure if X doesn't eventually have a great moonshot in the education space. We've tried like 30 things.

54:48It's so painful. But I also don't want to be kidding ourselves that something is a solution if it's not really going to be a significant phase transition for society. And some things may either just not be at all amenable to technology or have so many sort of other complex human issues around them, like pedagogy in education, if you take the teacher out of the process, again, nowhere is it written in stone. You can't help kids. But it might be fundamentally different and harder. So there are lots of those kinds of things where we continue to go back to the well. And technology is what we're good at, so we're not going to sort of do things that don't have a sort of core technology piece to them.

55:36And some of these hopefully will have great news for you in the future, but haven't solved yet. Have you thought about building an app you could make at Blue where people would post their personal profile and connect with the people they know in life and maybe share small updates? Like 140 characters, just very succinct. Well, dark or blue? Okay. That could be something. No, I feel like the world's already got one of those. and you know what we're best at are the things where in the early days it doesn't look at all reasonable or possible and usually because we're not scientists per se it's less that we're doing kind of basic research it's more like taking something from column a and something from industry B and some observation from sort of field C and putting them together in a really unexpected way.

56:34And again, often we're wrong. But if those things come together, then at our best, I think we're system integrators and we're really good at getting mud on our boots, getting out in the world, working with partners, prototyping quickly, trying and learning and using humility is a superpower to be wrong, admit we're wrong, learn from being wrong, and get better faster. Like, I think that's us at the best. Okay. And then we also had someone yell out carbon sequestration. I think that was what it was. Yes. Okay. He's not correcting me. So thoughts on that? Yes. We have done a bunch of interesting work in carbon sequestration, in green hydrogen.

57:17We're doing explorations in lots of other parts of the space. We're excited about the possibility for making much lower carbon footprint cement. So I can't tell you that we've absolutely cleaned any of these problems off yet, but these are spaces that we're excited about. And by the way, spaces where inverse design can be really helpful because these tend to be large machines, chemistry problems, electromechanical problems. They have sort of a lot of the techno economics of these things are really unforgiving. Ultimately, the number of dollars it costs for you to sequester a ton of carbon is like the only thing that really matters at the end of the day.

58:02Green hydrogen, of course, everyone will take it, but the number of dollars per kilogram is just going to determine, because it's a commodity, whether people buy it or not. And so trying to come up with something very different, but also have an eye on the sort of long-term, like the Rust Belt engineering side of this work, the project finance side of this work, what would this really be at scale? Because it's actually kind of easy to make any of these things look great in the lab. It's having that sanity check on top. What is this really going to be at scale? Would this actually change the world?

58:41You know, we had this project. It was one of my favorite. We had a device. It was about as big as like the area where we're sitting, a little bit bigger. And it was taking seawater and producing methanol you could burn in a gas tank. There are four billion internal combustion engines in the world. Tears were running down the faces of the people capturing the first couple of drops here. This project was called Foghorn. That felt like a real save the world moment. It was actually working. And we could not convince ourselves we were going to get it cheaper than about$15 gallon of gas equivalent. And that's just not going to save the world.

59:20So we turned it off. So tighter economic conditions, interest rates around 5%. Do you worry that now that we've exited zero interest rate environments that you might not be able to get as much funding from the mothership as you had in the past? I think Alphabet needs to be very thoughtful about how it spends every dollar. And we have spent the last 13 years working hard on our efficiency and on our rigor. So, you know, if Alphabet stopped being interested in sort of like the 10-year plus, that would be one thing, and Alphabet is very serious about the long term, and as long as Alphabet is serious about the long term, I am sure that making sure that we spend every dollar wisely would be very important, and I think that Alphabet is still excited about having a part of itself that goes and explores these other spaces.

1:00:16Let me ask you one last question. We started talking about where invention could go wrong. I actually am kind of curious from your perspective, I know we only have like a minute left but um why do we continue to to um try to invent and build because it does seem like in some ways that like if we all put our heads together we could have enough on this planet for everyone but we don't do that so i'm kind of curious like what you think where are we trying to get to

1:00:48I'll give you two answers the first answer is humans are fundamentally explorers I think we all have some pioneer spirit inside of us wanting to learn, wanting to grow wanting to find new things and do new things it's a very fundamental part of who we all are So I don't think humanity is going to stop doing that anytime soon because I think it's part of what makes being a human great. And don't worry, I'm going to end on a positive note here. There is enough pain and complexity in being alive, enough reasons to think short term, that people are going to by and large do what is in their somewhat narrower self-interests and what solves best for their pocketbooks.

1:01:43So if we want to save the world, if we want to make the world a radically better place, we have to find ways for doing the right thing to be cheaper than doing the wrong thing, especially when it comes to climate change. And the only way that we're going to get to the place where doing the right thing is cheaper than doing the wrong thing, when we can dig the problem out of the ground and burn it, is going to be radical innovation. so that's why I believe we're working on it and I think that's why the whole world is working hard on it right now applause applause applause Astro, thanks so much for coming on my pleasure, thank you for having me thank you everybody for listening thank you so much to our live audience I want to thank Summit for having us here thank you Nate Gawatney for handling the audio LinkedIn for having me as part of your podcast network Everybody in the audience, if you find Big Technology Podcast in your app of choice, we have the product manager on BARD, Google's chat bot.

1:02:49That's going to be on the feed within the next week as well. Thanks again for being here. Thank you, Astro. And enjoy your time on the ship.

1:03:04Americans told Washington what they want. Lower costs. Yet, Brandon Carr's FCC is considering a merger between TV station giants Nexstar and Tegna. If approved, millions of Americans will pay even more to watch TV. They'll pay more to watch the news, entertain their kids, or cheer on their favorite team. Thankfully, President Trump knows what Americans want, and President Trump can stop this. Brandon Carr needs to reject the Nexstar-Tegna merger and protect Americans from higher TV bills. Paid for by Cupid Local Media.