Nice to see you again , Astro. Thank you for inviting me. What a pleasure, friend . We've known each other for probably about twenty years . You know, I think most people don't know something about you . This is worth mentioning just out of curiosity, since we're talking about genetics. His maternal grandfather was a Nobel Prize winner in economics, and his paternal grandfather was Edward Teller, the creator of the hydrogen bomb. This is a rather interesting fact. No pressure. It was, it was, it was a high-pressure childhood, yes. Yes, I guess I agree. So, um, let me start with the first question. You've been involved in " monthly" projects for over 20 years. Tell us what it really takes to come up with and launch such a project. How to make this a reality? There are many people here who believe in optimism, who
[00:01:00] believe in such bold projects. Well, in 10 seconds. No, I'm kidding. What is your advice? You've learned so much. Let me say two quick things. First, let's define what a " moon project" is, at least as we understand it. As I understand it myself. Perfectly. from you. So, three main components. First, there must be a huge global problem that you can name and want to solve. If you can't name the problem, it's probably just an academic exercise. Second, there must be some kind of science fiction product or service, and we must agree: if it could be created, it would solve that huge problem. And third, there has to be a breakthrough technology that gives us at least some chance, at least a small hope, to create this fantastic product and solve the problem. When it is,
[00:02:00] we call it the "moon project" hypothesis. This doesn't guarantee success, but at least it can be tested. The second thing I would add: one of the fundamental problems in X is that you need to have two things in equal measure to be a researcher on such projects, to be a team. First, you must have great courage. You must be willing to put aside your skepticism for sound reasons and embark on unlikely journeys that can actually profoundly help the world. But the second thing you must have in equal measure is humility. You have to know from the first minute you set out on this unlikely journey that it's unlikely, otherwise you'll go too far down this destructive path before you find out that it's one of those 99% that just won't work. So the opportunity to say at the very
[00:03:01] beginning, "I'm proud of trying to do this, but it probably won't work." Let me find out as quickly and cheaply as possible whether this is even possible, so I can move on to the next ambitious project when this one isn't what I need. Mhm. So what's your advice to people, I mean, what about the challenge of choosing an ambitious project, now the speed at which you can test and build things, okay? What about the beginning, choosing an idea? Any advice on this ? Well, I mean let me explain it this way. What do you do to quickly dismiss an idea? Sometimes the idea is simply not big enough. Sometimes an idea is too rational or too likely to succeed. Mhm. Uh, sometimes it just doesn't do the world any good, or the potential for a lot of secondary problems associated with it is just too high. Often it is techno-economics. I can't tell you how many projects like this that seem
[00:04:00] so great to people in the early stages. Often these are very technical people. And only when you say, “Really? Let's go step by step. What is the best-case scenario for the cost of the materials you are going to make it from?” What is just the raw cost by weight of what will go into this? What will people actually be willing to pay? Can it even be plausible zero, zero, you know what Elon calls it? First principles thinking in this aspect. Yes, thinking from first principles regarding techno-economics. You can often weed out ambitious projects in the early stages because they simply won't turn into a sustainable business. At least I believe that purpose and profit can support each other. And if you do something that will be unprofitable, it probably won't change the world. Yes. Dave? Yeah, I'm really curious, you know, Alex and Peter have this book called " Solve Everything." Just phenomenal. But the speed at which physics, materials science, and
[00:05:00] chemistry are going to change is simply astronomical . So when did you come to Google, 25, 23 years ago? How long ago was that? Um, 16 and a half years ago, I co-founded X 16 and a half years ago. 16 and a half years ago. Of course. And at that time it was a company worth about $30 billion ? About. Yes. That was, that was very nice. It was a very small company back then. A small company worth $30 billion. It is now a company worth almost half a trillion dollars. So, the scale of R&D in such a corporate environment has grown incredibly. Absolutely stunning. And it seems like behind the scenes you were saying that the ideal goal for a new idea is 100 ideas per year, of which two are successful? That's right. We take on 100 to 200 ideas a year that progress far enough to get a code name. Who knows how many we actually consider, but 100-200 ideas a year reach the stage of receiving a code name. Of these, about 5-6 years later, we release two "moonshots" of X. Wow. So, the success rate is 2 %. And if, you know,
[00:06:00] if AGI is inevitable, and the pace of ideas... I know. He is already here. Astronomically, True. does it grow on the same exponential curve? Or is it just that the number of things at the beginning of the funnel increases from 100 to 1,000, to 10,000 and... Oh, absolutely. I mean, there is no shortage of problems in the world. There is no shortage of creative ideas to solve these problems. I think AI will, to some extent, shorten the time from a crazy idea to the point where we eliminate the risks enough to confidently say: here is the evidence that this is no longer a crazy idea. So, this time is getting shorter. I hope that AI will allow us to raise the bar of courage even higher. So we will set more ambitious goals and be able to do more in a shorter time. But ultimately, I believe the structure will remain the same. I think, you know, if you had asked me 3 years ago what big things came out of X? Um, Waymo is at the top of the list. Um... Google Brain. That's what I wanted to ask about. So Google Brain was probably something of a "yeah
[00:07:00] , that's cool" thing until a certain point a few years ago when it became clear, "no, that's the whole future of Google now." You guys just spent 80% of your spent 80% of the last 3 hours talking about what is Google Brain? Give a definition, because we know DeepMind, but we don't know Google Brain. So, Google Brain, which started at X about 15 and a half years ago. This was one of the first things we started. At that time, there was a group of scientists who were out of fashion. It was Andrew Yin, Yang LeCun, and a few others who said, "I think scale can matter." And neural networks were completely dead in this field at that time . And they said, " I think scale matters, and this could bring neural networks back." And everyone else in the industry was like, "No, you're crazy." Andrew Yin then came to Google X, and we created what was called Google Brain with Jeff Dean. The two of them came together to say, "What if we
[00:08:00] industrialized neural networks?" What if we made them tens of thousands of times bigger than anyone had ever made before? Could this be the breakthrough everyone has been waiting for? So we were the first to really do, let's move forward, yeah. teaching. TPUs eventually came about thanks to Google Brain. The transformer that underlies everything. It's literally the letter "T" in ChatGPT. This came from Google Brain. So, you know, our job is to create the seeds that will ultimately change the world. And I think Google Brain is one of them. But there was a long period when it just looked nice and promising. And then recently there came a period when everyone thought, "Wow, that was so much more than just cute and promising." Of course, but I think the reality of any innovation is that it takes a lot of time. Any innovation is an overnight success that took 15-20 years to achieve . And only because the general public is learning about it in the
[00:09:00] 18th year of the journey does it seem like it happened suddenly. Overnight success after 11 years of hard work. Of course. And So, Astro, you've seen hundreds of "moon projects," evaluated hundreds... What are some examples of projects like this that you would like to see, but they didn't succeed, didn't live up to expectations, or anything that you would like to see implemented? I mean, I have a very long list. I'll give you two examples, but we're working on clean water again and again. If you can make water clean, if you can extract it from the atmosphere, if you can desalinate it for ten times less, you have to get it to about a penny a liter for it to really change the world. But almost 3 billion people in the world experience water shortages, they have nothing to drink. The situation will worsen due to climate change. We will have hundreds of millions of climate refugees around the world, and often it will all start
[00:10:00] because of a lack of clean water. This is one of the biggest problems to solve, but we've done a great job in the feasibility study, but it will cost 10 cents per liter. No. Well, we'll stop doing that. Let's try again. So this is an example of something you keep coming back to. Two more. Two more, okay. The ability to transfer energy across time and space—I say that intentionally, because when you think about power lines and batteries, people tend to narrow their focus too much . But the ability to transport energy across time and space will truly change the world. We have made several attempts in this direction. So far, you know, we're not thrilled with the results. Education is another example. Education is not working. It doesn't work in developed countries. developing countries. Ahem. We will return to this issue until we find a solution, because this has to be one of the
[00:11:00] most fundamental "moon projects" that can be implemented. Got it. Perfectly. Alex, please. Yes, I would like to talk a little about the history of the "moon projects." So, one of your grandfathers invented the hydrogen bomb. And if you look at history, it's worth adding that humanity was once in the process of landing people on the Moon in real " lunar projects." And then it stopped. One can argue about the chronology, but perhaps, depending on which version of the Great Stagnation Theory one adheres to , there was a period of up to 50 years, say from the 1970s to the 2020s. Again, one can argue about the details, but it seems that there has been, shall we say, a slowdown in the rate at which such projects are being launched by humanity as a whole. We stopped landing people on the moon. We have stopped inventing perhaps applied physical innovations, regardless of their military potential, such as hydrogen bombs. And then, again, around 2020, give or take a few years, the lunar
[00:12:00] projects started again. We got superintelligence that really works. We get robots that actually work. Physical intelligence, so-called, seems to be starting to work. We apparently received medical nanotechnology in the form of lipid nanoparticles, which helped overcome the pandemic. So, a few related sub-questions. First: do you believe, do you support any version of the Great Stagnation hypothesis that there was a half-century pause in ambitious projects? And if you support this, another question is: how do we prevent this from happening again in the future? Well, um, first of all, I know you may not have meant it , but I take it as a huge compliment and an 11-year-old joke that we just created. Since we started in 2010, and you said 2020 several times , that's my argument: we literally founded a factory of ambitious projects in 2010. You actually brought back these ambitious projects. Well, yes, to a certain extent. I want to believe
[00:13:01] that this is us. So, that's the first part of the answer. And so, I think I don't know if anyone planned this stagnation, but functionally we got a little off track somewhere in between, you mentioned your Star Trek earlier. Yes. Somewhere between the 1960s, NASA, and the inspiration that created, there was a period where we kind of lost the spirit of explorers. What do you think? Do you have an ideology, a diagnosis? What do you think went wrong? I do n't know, but here's a partial answer to that, and then I want to move on to the future. Yes. But here...Can I tell you a little secret? This will be our little secret. Implementing ambitious projects is very, very easy. Ha, that's just ridiculously easy. If you don't care about efficiency. You just find super-energetic people who are somewhat deceptively optimistic, pour a bunch of money into them
[00:14:00] , and you're sure to get ambitious projects. It just doesn't give a very good return on investment. And so I think we had a period, mostly because of World War II and the Cold War, where people weren't concerned about the return on investment. Yes. Then there was a period when people probably still wanted ambitious projects, but they were concerned about the return on investment, so there were fewer of them. Company X is trying to create a “ breakthrough idea factory” to systematize this process—not just do it often, but do it effectively so that investing in such projects becomes rational. As for the future, I have good news: every week we are approached by large companies or even entire countries who want to open their own "factory of breakthrough ideas." I'm not sure exactly how we can help them,
[00:15:00] but there is a great desire to do more things like this, so I expect you will see a lot more of these initiatives in the future. But, Astro, stop for a moment: there are many people in this room who run companies and really want to create their own " breakthrough idea factory." We've already discussed this a bit, right? Let me paraphrase what I learned from you. You are my mentor for breakthrough projects. That is, your main organization is responsible for ensuring 10% of profit from year to year. Salim and I have written about this: your “ breakthrough idea factory” should be at the periphery of the organization, where your most creative thinkers work, right? And again, quoting you: if you work on something with a 10% rate, you get fired. Focus on 10x, on 1000% , and have this separate “ breakthrough factory” report directly to the CEO. Is this formula still relevant? Absolutely. I conducted the following experiment. Sorry
[00:16:00] Peter, who has heard this many times before. Salim has probably heard this a few times too, but give me 45–50 seconds. Option A, Option B. Option A: You can make a million dollars in profit for your business this year, and it's guaranteed. Or option B—a billion. You could make a billion dollars in profit for your business this year, but it's not guaranteed . That's a one in 100 chance. So, A—a million guaranteed, B—a billion, one in 100 chance. Who chooses option A? Okay, almost no one here, almost no one in the world raises their hand. Who chooses option B? Oho! Great , you all passed the math exam. Because the expected benefit here is 10 times greater. I've done this with top managers all over the world and said, "Okay, now don't give up if in the best of times, in your wildest dreams, your boss, CEO, or board of directors really supports your choice of option B." Yes.
[00:17:00] And everyone in the room drops their hands. And I say, "You don't need a lecture on innovation, you need a new leader." Yes. So the wording The wording you just described is absolutely correct, and it's all because of this. You have to allocate B options to where they are truly wanted and where the mess they cause is tolerated. It's like Skunk Works at Lockheed, somewhere on the sidelines. It's like Steve Jobs taking the Mac team off campus to create a project. That's right. Yes. The human mind's capacity for this is very limited. There is something I have been observing for the last few years and I would like you to comment on it. Peter wrote about this in one of his first books, about demonetizing the world, right? Throughout human history, advanced technology has always been expensive. And today, for the first time in history, advanced technologies are cheap. We have solar energy, cheap sensors, open-source and free blockchain, etc. This
[00:18:00] should help a CFO or a large company: in the past, to do something disruptively new, you had to risk everything. Invest a lot of resources. The CFO hated you. He financed it and closed it as soon as possible . Today you can take a portfolio of very disruptive experiments on the periphery. So now it all comes down to thinking: are you ready to go for it? Do you agree with this and how often do you see companies doing this? I...I agree. I often hear, "Oh, well, Alphabet has a lot of money, that's why this is happening." I guarantee you, that's not the reason. We have tiny teams. The teams that come out of us, like Google Brain, started out with 18 people, and that's completely normal. We keep our teams small because I keep telling them, "I want you to find the cheat in the game called life." The moment of cutting the Gordian knot. If you can't find it, I don't want you to solve problems the usual way. I would rather
[00:19:00] advise you to stop and look for another “moonshot ”, because that is the essence of such projects. So I think it's mostly a mindset issue, and for those of you who are interested in learning more, we've created a certification workshop. This is approximately 20 hours if you complete all the tasks. That's about 3.5 hours if you're just watching the video. I encourage everyone to watch it . It's about breakthrough innovations and leading people to results step by step. Our secret ingredient is not a secret at all. We tell people in detail how to do this. And tonight, Astro, your team is hosting an evening program, a 50-minute condensed version of that 10- hour course. That's right. Tonight at 7:00 PM, Joe Sargent. This will be very interesting. It is dedicated to rapid prototyping. Surprisingly. I would also like to better understand and study the economics of "moonshots." On the one hand, it can be argued that the real cost of such projects increases over time, as expectations become higher, and we see complex problems being solved almost on the fly. On the other hand
[00:20:03] , it can be argued that the real cost of achieving success is rapidly decreasing because we have superintelligence and can solve a bunch of complex problems. I could also give some mediocre argument "in the middle". What is your position on long-term economic “moonshots”: hyperinflation, hyperdeflation, or stability? Well, we track this because we aim to improve the efficiency of implementing moonshots. We very carefully calculate how much it costs us to achieve results, and I can say that over the past 16 years this amount has decreased by about three times. There are many losses and difficulties there, and this is a lagging indicator. So, some of what you... Wait, wait, three times in that time—is that inflation? No, no, it's become three times cheaper for us to understand each other. This is a ratio... Okay, but this ratio of three over 16 years sounds like 10% per year to me, or at least like a market
[00:21:00] return. Maybe it's market profitability. We are talking about two different things. Yes, I'm talking about the cost of building Waymo. I guarantee that the value of Waymo is so huge compared to not only the cost of that team, but also all the other things that didn't work. This is a very good return on investment. So I'm not worried about that. I look at how much each individual thing that goes through our process costs, and it's getting cheaper by, I don't know, 10-20% every year, something like that. Partly because we do our jobs better. Partly because we finish projects a little earlier. Largely thanks to artificial intelligence. So it's a combination of all these factors. But now it is cheaper to implement ambitious projects than before. But perhaps, taking this idea a little further, some would say that solving the most ambitious scientific and engineering challenges will not be like a one-time, powerful leap. This will be the
[00:22:00] result of macroeconomic growth, when we holistically increase the potential of our economy, and this facilitates the implementation of specific ambitious goals. And when I hear about a threefold reduction in costs, probably in real terms over 16 years, to me, maybe you have a different position, it sounds like the result of markets doubling in size, say, every 10 years. The stock market doubles every 10 years. So the result of reducing the cost of ambitious projects sounds to me simply like our economy is capable of more, and the cost is actually the same. I disagree. I think you are confusing the cost of painting a house with the value of the house itself. These are different things. They are not the same thing . Of course. I'm talking about how cheap it is for us to paint our house, not how much the world will pay for it. Two ambitious projects a year, Astro? I mean, why don't you launch 100 little Alphabet startups every year? This is about a tenfold increase, right? Rightly. Again , if I were
[00:23:00] only interested in the projects themselves, and not the efficiency, I think we would have done that. Because I am obsessed with efficiency, there is a cultural rhythm and set of habits that are easy to describe but extremely difficult to get people to put into practice. Things like playing the long game, intellectual honesty, humility, teamwork, and a lack of egomania. Like, everyone says, "Yeah, of course, everyone wants that." But just take a moment and ask yourself as honestly as possible, is this how your organization actually operates today? Yes, I hear laughter because it's not true. It's not your fault, but someone would have to be maniacally determined to create a microcosm in which these things were actually adaptive behavior to do so. That's what I'm maniacally focused on— creating the culture necessary for people to effectively practice habits that drive innovation. Unfortunately, I
[00:24:00] know how to do this with a few hundred people. I don't know how to do that with a few thousand people. This is a cultural problem that I just don't know how to solve. So, Astro, are n't you saying that the main "moon project" is a meta- project to solve the organizational problem of launching new such projects? Um, yeah, I'll link it to your question because of how I think about it : I would say that the meta-project that we haven't talked about is how to systematize radical innovation . So, to answer your question, there is a meta-meta-project: how to systematize the systematization of radical innovations. How, how do you copy and paste the factories of " moon projects"? Which brings us back to what I was saying earlier: the world comes and asks us about it. We would be happy to help. I have n't figured out how to implement a meta- meta-project yet, but we're happy to help. Organizational singularity, Salim. Yes, I think we might have this methodology for you because we've been building our social tier for 10 years . We are now 50,000 people
[00:25:00] worldwide, and we know how to create an environment where you can launch these things very quickly. So, it seems to me that you've pretty much solved the inner layers of Stuart Brent's "tempo levels" model, and we've solved the outer ones. True love for the Earth. So, this could be a powerful collaboration. Dave, did you want to say something? Well, I would like to take your quote and say, "Listen, you know, here's a dirty little secret for everyone." Making a " monthly project" is very, very easy. You take fanatically focused people who care about the mission, who are very, very smart, and you put money into them. Perfectly. Now comes a world where you replace people with AI and money with computing power. This world has not yet arrived. There has to be some kind of transition period, which we're in now, where there are three AIs and three humans, and I'm putting computing and money into them. And then there will be a different ratio, and... Have you already started this path? Quite. I mean ...Okay. Deep, I will relate your question to a small digression.
[00:26:00] On the one hand, if someone came up to you and said, “I want to introduce my company to you. Are you ready for my presentation? We use electricity." You would think I was out of my mind . On the one hand , you would be scared if I didn't use electricity in business, but you would say, "So what? This is just a technical detail. What do you even do? Yes. I feel the same way about AI. I generally try to use the term "AI" as little as possible, because I'm already annoyed by how we... We should say... We should call it "AI"—extraordinary intelligence. Sorry, I got my PhD in artificial intelligence 30 years ago, so I'm a bit stuck on this. But me too. Er... On the one hand, as you just described, every one of our breakthrough projects is like that . It is a group of agents and people who work together to solve a problem. This is just a technical detail. We're obsessed with the problem itself: whether it's clean water, or how to create industrial- scale energy-saving systems, or education, or repairing power grids, or how to make our world truly circular
[00:27:01] so that we're not sending trillions of dollars to landfills every year. These are the very things we live by. And in the background—yes, the people and agents who make it happen. But our main focus is on the problem itself and how to implement it into the business so that it grows and does not require excessive costs. Let me translate this into budget language. If you say: clean water is a great breakthrough project. I need... what, about a cent a liter to change the world? Yes. This is a great project. If I said: let's do it... Do you already have a team for this? This ... There were teams in the past . Yes, we have a team working on this now. Okay, okay. So which is bigger: spending on AI or on salaries? In this particular case, there are still salary costs, but to be honest, I don't know exactly how much. So I don't know for sure what the share is, but I can say that they are full partners in the innovation process. And, actually, you know, we're constantly finding ways to automate parts of our " breakthrough
[00:28:00] idea factory" to work faster and more efficiently. What is your guess? You know, if you focus as much as possible on the efficiency of this innovative machine, there is a case to study. Their AI spending is now less than their personnel spending. When will the day come when this ratio reaches 50/50, 75% AI, and then 100%? Well, first of all, I'm not sure that we'll get to 100% in the next decade, because I think that identifying what problems are worth solving and ensuring that they are solved in a way that is acceptable to society is an important part of the task. If you need distributors to help sell it, if you have to implement it in communities in a way that they find acceptable, then you need people in that process for at least the next ten to twenty years. Of course. So, look, it will only develop as fast as it is effective. We are not going to use computing power just for the sake of power. There were people who launched agents simply because it felt good to feel powerful by launching them. But there is a
[00:29:00] difference between running an agent and actually benefiting from it. Mhm. Ashton, go on, Salim. Yes, I divide breakthrough ideas into two classes, okay? The first is “pain points”: overcoming hunger, improving education, solving anything. And the second are ambitious, far-sighted projects . Elon, you got us to Mars, right? Solve the problem of aging, etc. What does your work with breakthrough ideas focus more on: the former or the latter? Is it balance, or are you only focusing on the problematic side? Well, if you worked at X, I would say, “Oh my God, can’t you get both? I want you to find the most daring things to work on and turn them from a crazy idea to something smart in less than a decade, and then into a sustainable business in less than another decade .” So if you want to go to Mars, I would start by saying, “Okay, I’m listening. How will this turn into a sustainable business in less than two decades? I'm not saying it's impossible, but that would be my
[00:30:00] task. So, you're limiting yourself in time to understand whether we can reliably work on this problem in the next decade and identify what will become the main criteria. Because it sounds logical. Since I get money, our investor— Alphabet—I have to report to Alphabet: this is what I do with your money, and they have to be happy with the flywheel I'm running for them. This is true. Here's what really fascinates me—is that you reject the vast majority of the ideas you initiate. Give us some numbers. The total number of projects started versus completed, is it less than 1%? Uh, I think it's closer to 2% , but let's say it's about 2,000 projects over the last 16 years that have been given code names, from which we have , depending on how you count, between 35 and 50 graduates. So tell us about this idea of the importance of giving up your own "children" in a timely manner. Well, um... No no not his real children.
[00:31:00] Yes. I, I, I'll use , I'll use the analogy of the monkey and the pedestal, which I know you're very familiar with, but imagine that we're working together on a breakthrough project, all of us, and we're trying to teach a monkey to stand on top of a 10-foot pedestal and read Shakespeare. What should we do first? Train a monkey or build a pedestal? It's so strangely tempting, if you're being honest with each other in a situation that's not that extreme, to build a pedestal, because then you're halfway there and you can get another round of funding to live on. You can go to your boss and say, “Can I get a raise? I'm half done, man. Can I get a bonus? This is what makes the world go round, and it drives me crazy . So we at X say: forget about the pedestal, just focus on the monkey. If you can train a monkey, we can always build a pedestal later, and if you can't train a monkey, thank God we didn't waste Alphabet's money and our time on that. And so, it's not because I
[00:32:01] want to lose. I'm like the crown prince of failure, but not because I like to fail. Because you don't learn anything when you're right. You only learn when you make mistakes. I like it. And the Crown Prince is a failure. So if you agree—and it seems like you do—that this is true, and breakthrough projects are a process almost entirely about learning , then we are embarking on an adventure that is only limited by how fast we learn. This means that if we feel bad at a learning moment because our model of the world turned out to be wrong , we will unconsciously, and perhaps consciously, avoid that moment. So, we're going to take the stigma out of failure. If you work at X, I need you to share with me the mission of optimizing X's portfolio. It's not that I don't want the teleporter to work, Peter, who's working on it for me, but I want to
[00:33:02] find out as efficiently as possible whether it will work at all. I don't want you to deny the fact that you just really want the teleporter to work, and we spend five years, tens of millions of dollars, and the project still stalls and becomes a "zombie ." How many of you have spent at least 5 years on something that was a "zombie," and you knew it was a " zombie" for at least two or three years? What is your worst failure? For example, that failure that made you think, "This is really crazy." And then it turned out: “You know what? "That was really very stupid." Oh, we've had many such cases. We have one right now that is almost certainly false. Um, that would be a problem for physics if he turned out to be right. So, um, but it works. We don't know why he works in the lab. We work frantically to find where we are deceiving ourselves so we can shut it down. We've
[00:34:00] been trying to figure out this cold fusion-like moment for about 5 months now, even though I hate cold fusion. It is not cold and not synthetic. But that's okay, okay. It still didn't work. It's okay . We don't want to be that group, right? We try to be the opposite of that group. tell us what it is . No. Oh, well, that's it. OK. If everything works out, we will tell you about it . Everything, for example, causes me deep discomfort. Deep excitement. Okay, a few questions from the audience. Rich asks: Are there any " moon projects" that you've turned down in the past, but that you would pursue today because the world's challenges have changed? I mean, we made a lot of mistakes. So I'm going to answer "no" to that question, but I don't want that to be confused with the fact that we've never been wrong. We make mistakes all the time . The question, however, is, Do I regret this? But was it something you dismissed and didn't even try? Okay, fine. I want us to understand everything clearly. The cost of a false positive, when we believe a project is breakthrough, fund it for years, and then it turns out that it is
[00:35:01] n't, is very high— potentially tens of millions of dollars. Yes. The cost of a false negative result, when it's actually a breakthrough but I gave up on it, is zero. As long as you believe that there are an infinite number of problems in the world and an number of creative solutions, we can always just try again to find new breakthrough ideas. It doesn't cost us anything to say "no." Therefore, I refuse and do not regret even those things about which we were wrong. We're probably not doing it fast enough. Paul asks, “What about breakthroughs in materials science? Do you see any particularly promising areas for research in this field? Well, a small digression: I am excited about the field of new materials. Yes. And I think people haven't fully thought it through, because there's such a hype around technology right now, and everyone says, "And then we'll build a business on this ." Yes, it's more complicated than it sounds. Look how many centuries it
[00:36:00] took us to truly perfect steel. Steel radically changed the world, but not before it began to be produced on an industrial scale. There is a lot of new material; If you could create even flakes of room-temperature superconductor, that would be really cool. Nobel Prize, hurray. But this is not a business. This is only the first 5% of the business. How do you produce thousands of tons of such material every day? How do you make it plastic enough to be able to, for example, make wires and such? All of this—the other 95% of working with new materials or something else—is what is needed for a real breakthrough. So I'm excited about a lot of things, like new materials, but I'm focused on the next step, not the one that others are fixated on: how to scale these discoveries, not just how to find their מל start. One last quick question from the audience and hopefully a quick answer. So, other
[00:37:01] companies are constantly trying to emulate X and failing. What is the most difficult part to reproduce that you think most outside observers still underestimate? I truly believe what I said: you need a leadership team that is maniacally focused on creating a culture where people can act in ways that are most effective at driving radical innovation . And, if you allow a little advertising, just like you have a podcast about " moonshots," we have one too . Two seasons have now been released , where you can hear the stories of people who have gone through this path. So this isn't just me talking about "moonshots." These are real people who are involved in these projects, talking about their successes. Can they find it on YouTube? If you just search for Moonshot Factory or Moonshot Podcast, you'll find them wherever you listen to podcasts. I would advise people to listen, because you will hear exactly what
[00:38:00] I was talking about: a culture, a microcosm that encourages people to do the things that really drive innovation. That's obvious. You all know what it is. It's about creating that space, that protected place for researchers. This is the most difficult part. I would like to say something to answer this question. You created a mechanism, structured it so that the parent company left you alone for a long time. I've never seen another company do this. In every other company I've seen, the immune system eventually breaks down, reaches out its sticky little hands, takes everything back, and inevitably ruins everything. So congratulations on sticking around for this long. And I applaud the Google executives who allowed this to happen and had the discipline to keep it on the periphery, because that's where all the magic happens. Friends, let's give Astro Teller a round of applause .