♟ The Art of Failing Forward: Insights from the CEO of X, Alphabet’s Moonshot Factory, Astro Teller

🌎 Welcome to the Geopolitics of Business Newsletter, where I'll be sharing frank discussions with the world’s leading investors, CEOs and politicians about navigating our ever-changing global landscape. 

💫 Hit subscribe and join me as I seek to understand the interplay between global business, geopolitics and how we can work together for a brighter future.

A Conversation with the Captain of Moonshots - Astro Teller

In this edition of The Geopolitics of Business Newsletter, we take a fascinating dive into the world of Moonshot thinking and the relentless pursuit of innovation for the betterment of society with Dr. Astro Teller. 

Astro is the co-founder and CEO of X, Alphabet's moonshot factory which is responsible for inventing and launching breakthrough technologies. I wanted to speak to Astro on the podcast because science, technology and innovation now cuts across every aspect of our lives from economic prosperity, to health and even geopolitics. 

I’ll never forget our memorable first meeting at the Alphabet office in Mountain View, California when Astro rollerbladed into our meeting room!

Astro’s background as a serial tech entrepreneur and his expertise in artificial intelligence made him particularly suited to lead an outward focused team at Alphabet that is dedicated to solving real world global problems. He has successfully systemised innovation at a cultural and organisational scale, enabling his team to generate unconventional and valuable ideas more efficiently.

In this conversation, we covered his insights from pioneering an organisational culture of learning from failure, and navigating cultural and regulatory barriers to implement radical innovation on a global scale:

🚀 How does X navigate regulation when developing breakthrough innovations that sit outside current frameworks?

🚀 How can innovation effectively address the problem of climate change? And ensure that progress is globally equitable? 

🚀 What are the risks and opportunities of AI for the future of work? And what regulatory approaches should be taken given how transformational artificial intelligence is going to be?

🚀 And what can businesses and leaders learn from the radical culture of failure, progress and positivity Astro has cultivated at X as Captain of Moonshots?

💡 Here are four insights from our conversation:

1️⃣ Strategies for navigating regulation to catalyse breakthrough innovation:

Astro shared X’s approach to ‘sandboxing’ new projects in different countries around the world.

He advised that when trialing prototypes that are outside the traditional regulatory frame of reference, a deep understanding of various regulatory regimes and their openness to innovation is critical. Ultimately, new innovations will involve failure but it is more expensive to fail to learn from mistakes, than to fail, learn and move forward. 

Additionally, Astro recommends working closely with regulatory bodies from an early stage. Rather than presenting a finished product and demanding approval, his approach is to seek feedback and address concerns from regulators from the very beginning: 

“because our interest is not getting a quick win, our interest is in lasting innovation…we’re actually trying to make the world a better place. So we want to work with the regulators to make sure that what we're doing is for the good of citizens.”

2️⃣ Ensuring innovation is globally equitable:

Over half of all X’s projects are devoted to addressing climate change and half of their testing is south of the equator around the world. This ensures that the impact of innovation reaches places where the need is highest. 

Alphabet's Moonshot Factory recognises the significance of learning from diverse regions and engaging innovative projects in locations where new technology can make the biggest impact. Astro gives the example of one of their moonshot projects, working on the electric grid, which aims to virtualise the grid for better management and integration of renewable energy sources. 

Their team is currently working around the world, including in South Africa as it has one of the largest, dirtiest, and most brittle grids in the world and there is an urgent need to fix infrastructure issues. 

3️⃣Why AI is just advanced math & we should not fear progress: 

Astro demystified AI, describing it as "algebra on steroids" and emphasised its potential to improve society. He also discussed the challenges of regulating AI and suggested focusing on outcomes rather than the technology itself.

His argument is that AI is not something that can be discovered or regulated like a physical object. It is simply numbers and algorithms that enable computers to process information and make decisions. Therefore, instead of focusing on regulating AI as a separate entity, the emphasis should be on regulating the use and impact of AI technologies. Similar to crash tests for cars, Astro proposes evaluating the safety and effectiveness of AI systems based on their ability to protect or better the situation of people. 

“If you get too down in the weeds, you're arguing about things like math that don't have good answers to them. But up at the top, we know what we want. And so I think that that's what we should be requiring, is that the cars, for example, drive themselves better than humans drive them. If we hold that bar, then exactly how they accomplish that…is a problem for the designers to solve.”

We also discussed the impact of AI on job displacement. While Astro acknowledged that certain jobs may be lost, he argued it is as important that we account for the number of jobs that will be transformed for the better. He gives the example of the invention of spreadsheets, which pushed many bookkeepers to become accountants to illustrate how new technologies can create new and more interesting job opportunities.

“So we rely on artificial intelligence every day in hundreds and hundreds of different ways. It has made the world a better place, whether it's drug discovery or being able to use an ATM and sort of get out cash or put in cash. Anything which allows for humans to step up their thinking and allow for computers to fill in behind them is, I think, ultimately going to be good for humanity.”

4️⃣ Reward failure and social positivity to promote innovation:

Astro’s biggest insight is that rewarding failure promotes innovation. 

Taking us behind the scenes at X, he revealed the importance of creating a culture that supports and encourages experimentation, even if the outcomes are not successful. By doing so, individuals are more likely to take risks, try new things, and ultimately drive progress.

While he acknowledged that nobody (including himself) likes being wrong, the mindset of embracing failure as a learning opportunity and being open about what is being learned enables teams to move faster and make progress.

He recommends businesses and leaders build reward systems into the culture of an organisation that encourage social positivity. Recognising and celebrating individuals who take risks and learn from their failures creates a supportive environment where innovation can thrive.

By creating a culture that supports and encourages experimentation, individuals are more likely to take risks, learn from their failures, and drive progress. This mindset, coupled with teamwork and collaboration, allows organisations to discover the future and make meaningful contributions to society.

That's it for the final edition of The Geopolitics of Business for 2023!

You can listen to our discussion on The Geopolitics of Business Podcast, or scroll down to read the interview in full. Stay subscribed for more insightful discussions.

I’d love to hear your thoughts on Astro’s approach to innovation. To get in touch with comments or requests for future newsletters, email info@thegeopoliticsofbusiness.com.

We'll be back with Season 2 of The Geopolitics of Business later in the new year.

Thank you for your support and wishing all readers a wonderful and safe festive season and happy new year. 

Sam

Q&A with Astro Teller

Sam: We first met when I visited your office in Mountain View, California, and I had just come out of a spell in politics about a decade. I'd been a minister for a long time and I wanted to spend time with people who thought and saw the world differently. And you achieved your mission by rollerblading into the meeting room, I remember. And I thought, okay, I'm in the right place. So I want to start by asking you a question I didn't get to ask you that day, which is, how does one become captain of moonshots at Alphabet?

Astro: Well, a long time ago, 13 and a half years ago, Larry and Sergey were already thinking about a future where something like Alphabet would exist, and they were interested in having a part of Alphabet that would, instead of be focused inward on the things that Google was currently trying to solve, would be focused outward on new, interesting problems that Alphabet could have. And that's only interesting, really, if we can find solutions for some of those problems. So they asked me and one other person to co-found what ultimately wasn't named at the beginning, but what became Google X. And then when Alphabet was announced and X became one of the other bets, we became X, the Moonshot factory.

Sam: And what was it about your previous background that made you attracted to this opportunity? There are many other things you could have done. You're a published author. You’ve got a Ph.D. in computer science. What attracted you to this particular endeavor?

Astro: I think it was two things. I'd been a serial tech entrepreneur, so I'd spent quite a while on a number of companies trying to do really audacious things. And so that had built up a sense for me of why it's particularly hard to do that and what it might take to make an ecosystem that was more conducive to taking moonshots, but also because as someone who got a Ph.D. in artificial intelligence and was really interested in the idea of an invention machine, and I'd been thinking about that for decades, the idea that you could systematise the process in a computer of helping people to make ideas, inventions, designs, faster, get to a particularly unusual and valuable good for the world, ideas, quicker, get to more valuable states. And so that interest in mine, which started as a software thing because of my experience as an entrepreneur, sort of expanded into what it would be like to systematise innovation at the cultural scale, at an organisational scale.

Sam: So can you tell me about one of the sort of technologies that you've worked on and developed at X that has been successful as a moonshot and the process of getting there, the challenges that you had to overcome?

Astro: There are many. I'll give you an example. Wing Sure. Wing is a fun example. 

So we have believed for a long time that moving things over the last mile, last ten miles. It makes no sense that a three or £4,000 vehicle should be bringing you something which weighs a few grams, maybe a few pounds. Something that actually has sound pollution. I mean, I'm talking about a truck or even just a car. Risk to pedestrians. It actually has a very high carbon footprint. Moving a pair of shoes or a bottle of aspirin or whatever to your house. It makes no sense when something could fly through the air and not have to stop at stop signs. A vehicle that weighs 100th as much. Almost a thousand as much as your car. That's just a much better, much lower carbon footprint way of getting you what you want when you want it. But there was a lot when we started this 12 years ago that was unknown about how to do that. And the team went through a lot of explorations about things like, for example, should you land and then drop off the package? Should you hover near the house and just let the package drop to the ground? Should you lower the package on a string and sort of release it and then pull the string back up? And we came through a lot of experimentation to an understanding about why lowering the package to the ground, releasing it and then flying away without landing was safer and faster. It would be cheaper. And it wasn't that we just had the idea. It's that we had 50 ideas and that one is the one that after a lot of experimentation and learning and iteration proved itself out as the right way to do things. It now turns out that that same string with this little hook at the end, this little plastic, we call it a pill, not only can drop off a package, but it can also pick up packages. So a lot of this sort of learning accumulates over time.

Sam: And Wing is operational now. It is being used to deliver parcels.

Astro: Yes, very much in five countries now, I think at least a little bit in the U.K. as we speak.

Sam: Which is which. I mean, one of the things that I'm interested about, obviously, someone who's sat on the regulatory side of the fence is when you're coming up with products that sort of outside the frame of reference and outside the regulatory frame of reference, which countries are open to it, how do you identify where you can go and test the products and which regulatory regimes would be more open than not? How do you think about that aspect of things?

Astro: Well, we go talk to them. If we show up to any regulator with any project, this isn't just about Wing, with the attitude ‘we're right and we need a yes from you’ it's not a great way to start a conversation with a regulator, but that's where a lot of companies do start because they've already finished what they're doing. 

So we try to show up very early with regulators and say,’this is what we're trying to accomplish. What do you think about our goals?’ And then if you can be excited about those goals with us, agree that they're good for the world, that they're reasonable ways to try to solve those problems. Great. Here's how we're trying to solve it. What do you think about that? Okay. You have some concerns. Great. That's exactly what we're interested in. How could we address those concerns? 

And so we've done a lot in the United States with the FAA, in Australia, with CASA, in Europe, with the ACA, and in the UK as well. So not every country has been concerned about the same things, but we've found regulators very open with lots of the projects that we're working on. And because our interest is not getting a quick win, our interest is in lasting innovation, finding some way to kind of like skirt around the regulators and sort of run for it. That's not what we're in it for. We're actually trying to make the world a better place. So we want to work with the regulators to make sure that what we're doing is good for the citizens that, you know, we're around or above in the case of Wing.

Sam: And in your travels, have you seen different cultural approaches to innovation in different countries and different regions and how that plays out in terms of what can be achieved in innovation?

Astro: Yeah, I suppose so. So I'm known for being sort of on the failure bandwagon. And I'm going to get back to your question, but let me explain it in this way. I hate failure. I have no interest in failure. But once you commit to the long term, then being obsessed with learning is the right way to produce as much value as possible. The longer you have, it's a time frame, the more you can harvest those learnings inside that time frame. And once you're focused on learning, because all learning happens in the moment where you had some model of the world, you presented your model to the world in some way and your model of the world was wrong. You get the feedback loop and that could be, you know, you're just wrong about the physics. You were wrong about the design, or people don't want to pay for it. They don't want to pay that much for it, whatever that is. In some way, you had it wrong by whatever name you want to pick that feels like failure. But that is the moment where we learn. So X is really focused on de-stigmatising failure internally, not because we have anything positive about failure, but because feeling negative about those moments tends to cause people to shy away from learning itself. And that's what we can't afford if we're serious about being efficient. 

So interestingly, back to your question. There are a bunch of cultures, even some cultures, that have very strong engineering tendencies where being wrong is stigmatised much more. And so I would say those places can have a hard time with certain kinds of innovation because they go slower or even avoid the process of particularly unlikely outcomes. But if something is four times as valuable, but with twice the risk, there are a bunch of cultures where that's a particularly scary thing to do. If the culture that you're in is not amenable to those kinds of risk taking and getting a wrong a decent amount of the time.

Sam: But it can be expensive. I mean, how do you budget for failure? Is there, you know? What price do you put on that learning process?

Astro: You know, one of the sayings I don't know if it's worldwide, certainly here in Silicon Valley is there's never time to do it right. But somehow there's always time to do it over again. So you can say that failure is expensive. But I would say, especially when you're committed to the long term and to trying audacious things, to being innovative, you will be wrong most of the time. So we'll be wrong most of the time. That's not an option. The question is how efficiently you can learn your way through all the things you're wrong about to the right thing to do. So it's a fantasy that people have that you can somehow just be right the first time. And I don't think any of us can do that. And once you are humble enough to accept that that's not possible, then designing a system in which you can get to the right answer as cheaply as possible. But that's what X is focused on. So I would suggest that the ‘I'm going to do it right the first time’ is actually more a humility failure than a piece of wisdom, especially when it comes to the innovation space.

Sam: And listening to you, it sounds like a very matter of efficiency and innovation can co-exist in the same organisation.

Astro: They can. This has been a struggle for us, you know, because what we really want to do, we have these two separate dials, right? We want to have huge amounts of of magic, the good kind of chaos and unreasonableness that sort of creates the gold dust of over the horizon exploration and discovery. But at the same time, we want to do this as rigorously, as efficiently as possible. It's not rational for Alphabet to keep funding us if we're not really serious about that. If you start from an efficiency perspective, being efficient by itself is pretty easy. You just sort of remove all of the unknowns. You just don't change anything really don't ever be surprised. But there tends to be no innovation in that scenario. If you want radical innovation, you don't care about efficiency. That's actually also pretty easy. You just give a lot of money to very high energy, unusual people, and I'm sure you'll get some radical innovation. It just won't be very efficient. So you have to start and this is what we've done over the last 13 and a half years, more on the innovation side, on the exploration side, and then slowly turn up the rigour dials in ways that are done carefully so as not to kill off that creative and explorer spirit. And that's really the arc of X over the last 13 years, is constantly trying to turn up the rigour dial while still keeping the explorer, dial the innovation dial high.

Sam: Yes. And it's always been fascinating to me that for you, it's not just innovation, it's innovation that is solving some of the biggest challenges we face. And climate change is one of those. And I think more than about half of your projects, if I've got it right, are devoted to addressing some elements of the sustainability of the planet. I'd love to get your sense of, as you look into this problem, and it is a big problem not just for generation, but for generations to come. Is it technology? Is it the entrepreneurs? Or is it the capital? What do we need in place to address climate change?

Astro: All of the above, I'm sorry to say, is the right answer. Humanity is going to need to pull together on all fronts in order for us to maintain the quality of life that people are hoping for and simultaneously have a good planet to leave to our grandkids. I don't think there's a choice except an all of the above approach. So there's lots to be done on the public policy side. There's lots to be done on the finance side, even just the sort of more meat and potatoes finance like project finance for things that are already pretty well understood. But there is a lot to be done on the innovation side as well, even on the basic science side. Let me give you an example of the kind of thing that we're working on right now. 

One of the projects that's here at X right now is our moonshot for the electric grid. Right now, the system operators in any country in the world, including the National Grid, don't have a complete map of where every wire is, where every transformer is, where every inverter is on their grid. This is the world's largest machine, the world's most complex machine, the world's most expensive machine, the grid. It's been built over more than 100 years. And because of how it's been built up, pre-dating computers, most of it no one has a really detailed circuit diagram of their grid. So that doesn't allow them to, one, take care of it properly to plan for its immediate and medium term future. Which is why, by the way, you know, in most countries around the world, there are huge long wait times. Many years in the United States on the average, like 5 to 7 years waiting for a solar field or a wind farm to get onto the grid. They're already built. They just can't be plugged into the grid because the grid operators have such a hard time planning for whether it's safe to plug them on to the grid because they don't have a model of their own grid. And then ultimately, in the long run, you would want for the grid to become a marketplace for electrons where everything can either take or give electrons as it wants, depending on what the current needs and price is. In order to do all that, someone would need to build a virtualisation of the grid down to where every wire is, every inverter, every transformer. And so our moonshot for the electric grid tapestry has been building that. We're actually working with the National Grid. We're working with CTM, the National grid operator in Chile, we're working in Auckland, New Zealand, we're working in Australia, we're working in several states in the United States. And in each of these we are helping them to sort of take steps along this wave, virtualising their grid and then helping them to use this. For example, in Chile we're up to 30 times speed up on Chile's ability to think through, is it safe to plug this solar field onto the grid right now? Which is allowing them to start to get renewables onto the grid faster.

Sam: And that's innovation at scale. And particularly in this context, I'm always interested in how you make the technologies available to countries that don't have the kind of budget that certainly the G7 countries have and can deploy that much capital towards innovation. And how you think about your role in innovation and making it more equitable in terms of the impact?

Astro: Yeah, I'm glad you mention that. In fact, one of the countries I didn't mention that we've just started working in is South Africa. South Africa is one of the largest, dirtiest and most brittle grids in the world. It goes down all the time. There's a desperate need to fix it, and you can't fix something if you can't first understand it. So we're relatively early in our journey there, but we have started with that country as well. We are in some other projects doing work in India in maybe 8 to 10 countries in Africa right now. This is more on the telecommunications side, but we go to places where we think we can learn the most, where the need is highest, and that often is in the global south. So we certainly aren't only in the global South, but I would say half of our testing is south of the equator around the world.

Sam: I’d like to talk about artificial intelligence, which is top of mind for so many people in business, but also society causes huge public concern. So I’d like to start with what you see as the benefits that we're going to experience. If you can sort of enumerate that, because certainly from a political side, there is a lot of concern about technology taking over people's jobs or this technology that wouldn't be controllable or technology that could be that has the potential to be a threat to democracy. So bearing in mind those concerns, what are the benefits we can expect to see in a time frame?

Astro: Let me start with a little bit of a definition of artificial intelligence for your listeners. Artificial intelligence is just algebra on steroids. It's just math. So you can't go in with a microscope and discover the artificial intelligence inside a computer. It doesn't really work like that. It's just numbers. And artificial intelligence has been around and participating in society increasingly for almost 60 years. So this is not new. It's just people are starting to pay attention to it now. So the last time you got on an airplane, even ten years ago, when you got on an airplane, it flew most of the miles by itself, not because a human was flying it. And you get there just fine. And in fact, you're safer because the pilot doesn't have to be paying attention all the time. So we rely on artificial intelligence every day in hundreds and hundreds of different ways. It has made the world a better place, whether it's drug discovery or being able to use an ATM and sort of get out cash or put in cash, anything which allows for humans to sort of step up their thinking and allow for computers to fill in behind them is, I think, ultimately going to be good for humanity. 

You know, computers are levers for our minds. And artificial intelligence is just increasing computers' abilities to be levers for our minds. And our jobs will change for sure. But if you think back to the advent of spreadsheets, for example. There was an entire class of people who were bookkeepers. They were writing down what was going on financially in spreadsheets, but on pieces of paper, basically books. And then they moved over to being accountants, being analysts. And so bookkeepers as a profession went down, but the number of jobs that were lost there was tiny compared to the number of people who started working with spreadsheets. Now they had different names for those jobs, but I think what they were doing was actually a lot more interesting than what the bookkeepers were doing. 

I think what we're going to see with artificial intelligence over and over again is instead of focusing on the jobs that may be changing, focusing instead, or at least in addition on the jobs that are coming, I think they're going to be really exciting. It's going to allow more people to have more interesting work. That's my prediction for how this plays out. I understand the concerns, but let me flag two things. One, there tend to be concentrated harm and diffuse benefits. So anything in the world where 98% of the people are going to be better off and 2% of the people, at least temporarily, are going to be worse off. Those 2% of the people can tend to be very loud and understandably so, because they're the ones who are in pain and the people who are better off barely notice it. It's still a net positive for society. We society have a responsibility to help that 2% through that process. And I think that we let them down over and over again. But I don't think that the solution is for society to go backwards, but for us to get much better at seeing these concentrated pieces of harm coming and helping people with things like job training so that they can move to the next skill set to the next set of jobs. The other one is that there's the short term versus the long term. So when big changes happen in society, it is temporarily very painful. That's not necessarily an argument that we shouldn't move to a new version of society. For example, the Industrial Revolution, there was a huge movement of people from farms to cities to work in factories. And ultimately, I think, you know, net to net was good for society, but the change was incredibly painful for society. When we can see those kinds of disruptions coming and we can help society through these transitionary periods, that's what will allow us to get the most from these kinds of changes, like artificial intelligence, and minimise as much as possible the kinds of change or unintended consequences.

Sam: I mean, your definition of artificial intelligence is algebra is terrific. I've never heard anyone describe it like that before. But then it prompts the question for me. If it is not a thing, how do you regulate it?

Astro: Well, I'm sorry to be a bummer. I agree with people who would like to see regulation, but I don't agree with them that regulating artificial intelligence is the right way to think about it. You know, when cars, eventually started to become a thing and society eventually we said, hey, we should regulate these things. And so we designed crash tests as one of the main ways for us to ask, is this car safe? You smash your car into a brick wall at like 40 miles an hour and you see what happens to a fake person on the inside. You measure how that car crumbles and whether the dummy's okay afterwards or not. And you allow for a lot of creativity about how the car could be made, but you are inflexible as regulators about the fact that the dummy has to be okay when it hits the wall. The same thing should be true with respect to artificial intelligence. I would love to think that we are moving to a world of intelligent technology where we will stop, for example, asking the question. What happens when the car smashes into the wall? But how successfully can the car not smash into walls because it's more intelligent? But the question should be at the outcomes level, how great is this car at not hurting anyone outside of the car? How good is the car protecting the people inside the car? But how it does that, because there's probably lots of different ways to do that. If you get too down in the weeds, you're arguing about things like math that don't have good answers to them. But up at the top, we know what we want. And so I think that that's what we should be requiring, is that the cars, for example, drive themselves better than humans drive them. If we hold that bar, then exactly how they accomplish that. That is a problem for the designers to solve.

Sam:  And one of the things that occurred to me when we first met and you spoke to me about the technologies you involved, and I know it for a while, but it sort of hits you is how. Firstly, how little tech knows about government, but how little government or politicians know about tech. And so there is a question here about education and how people who are responding to societal concerns around new technologies to regulate, to put safeguards in place, actually develop the skills and expertise to do the job. And how do you see that playing out, given how transformational artificial intelligence is going to be as a technology, is already?

Astro: Yeah, I think it's critical that that communication be bidirectional and it be much better. Many politicians, I believe, are secretly interested and serious about understanding how the world actually works, even though they don't always speak about it that way in public. I think they genuinely do. I know you personally do. So the willingness is there and certainly the willingness of me, for example. But lots of people in the tech community are excited to share kind of where we are and where we might end up, how we can end up in as positive a version of the future as possible. But you're also right, we don't have, as a society, great ways of connecting, of trusting each other, of being able to reach out in the right moments and sort of sanity check on both sides. So I don't know exactly the way to solve it, but I would love to see better interactions.

Sam: I mean, you are the grandson of Edward Teller, who is considered to be the father of the hydrogen bomb. So I think I have to ask you, how does this legacy inform your work when it comes to innovating things that can both hurt and harm the world?

Astro: Well, I would say two things that I picked up. When I was a child. One, I was interested in working on things that were really clearly in the basket of good for the world. I've always been very purpose oriented in how I've spent my time so that something wasn't a rejection of my grandfather per se, but that certainly is something I was determined early on to not have to spend my time apologising for the work that I was doing just makes me feel better to know that the work that we're doing here at X, for example, we don't do it perfectly all the time, but we are as a group, very sincere about trying to get out in the world and focus on lasting innovation, as I was describing earlier in this conversation. The other thing that I picked up, though, as a child was that the Manhattan Project being able to get a group of people who were particularly bright and creative together, separated a little culturally and operationally from the rest of the world, and creating a sort of culture microcosm to get them to focus on making really strong progress in areas that matter. That is a legacy of the Manhattan Project. That's always inspired me and I hope in a very different way because we're working on very different things. You know, X has been inspired by and maybe, you know, has taken a few pages from that book.

Sam: We've talked a lot about technology, technological development, but you also lead a team of people as captain of Moonshot. So how do you keep the team spirits up if they are in an environment where failure is part of learning and reaching your goal?

Astro: Someone gave a talk at our all-hands. Every two weeks we all get together and sort of go over things that we're learning - successes and failures. And he talked about something he had tried, why it was a smart thing to try what he learned and the fact that it didn't work out and what he was going to do next and the fact that it didn't work would have been a good reason for him not to have given that talk. But he did. And he got a big round of applause. And I was talking to him afterwards. He is a very smart guy. And he said, I hate being wrong. But the fact that people will treat me like a hero, whether or not my experiment goes well, if I'm running a great experiment and I run it in a great way, has made it possible for me to actually try smarter things and to be more open and collegial about what I'm learning, which means I guess we're going faster. 

So it's that moment and there's lots of them here at X and we've orchestrated this. We've engineered it into the culture, but telling people it's okay, we support you. The real thing that's a mistake is if you have the data to know you shouldn't keep doing it and then you keep doing it. That's really bad. But doing a smart experiment and coming to the conclusion that that was the wrong direction, that's how most experiments go. We just pretend that it's not. So building into the culture, lots of ways of rewarding people, promotions financially, but also and maybe even primarily this sort of social positivity around this is how we do things. You know, a lot of modeling, I try to do it personally. If you have an idea and your ideas are better than mine, it goes a long way. If I just say, You know what, Sam, your ideas are better than mine. Forget my idea, let's do yours. Because it's not about the ideas. You're going to get a little bit of credit with me that it was your idea, but I actually don't really care about the ideas because everybody's got ideas. And ideas are pretty cheap. It's actually like how well we work as a team on the idea. That's what really matters. And I just created as much value by not being political or like hoarding ideas or trying to knock your idea down as you did by coming up with the idea. I think we all intuitively know that. But until people are modeling that kind of behavior, you just don't get good traction and everybody just squares off from their corners, which is what happens in most organisations.

Sam: So you're a published author, captain of Moonshots at X, and you work with some very smart people who all see the world differently. How do you deal with each other?

Astro: I don't think any of us can be much better than random at seeing around corners. We have one or two people at X who I think maybe can, but it's very unusual. But as a group, what you can do is you can think of a lot of futures. That's relatively easy. Then you can develop the muscles as groups to say, How do we feel about Future A or Future B or Future C? Is that a problem we'd feel proud to be working on? Is this radical proposed solution for getting at that problem something that we would be proud to be working on? Would it really? Even if we don't know how to make it, if that science fiction sounding product or service could help with, let's say, climate change, do we think there'd be unintended consequences with that? Would it make inequality in the world go up or down? Like, let's talk about that and then we can filter those ideas first, just conceptually to ones where we could feel really good about it and there's no least obvious problems with it. Then we say, Is it possible? Is there a hypothesis we can test? And each of these questions starts to filter down, down, down on this pile of ideas. We discover the future. We don't invent the future. We are trying more than a thousand futures a decade. And so if we look like we have ten really great ideas like Waymo, the self-driving cars or Wing the drones for package delivery, Verily, the life science business for Alphabet, Intrinsic, which is the moonshot for industrial robotics that came from X. In any of these things, they may look really great now, but it's through this process of heavy filtration on the basis of evidence that we've done over time. So I don't believe you can invent the future very efficiently. I believe you have to discover it.

Sam: And I've got to wrap up. So one last question for you, which is what's the best compliment you can give an employee?

Astro: One of the most fundamental things that we look for here at X is people who have high audacity and equally high humility. And it usually takes learning. Most people tend to either have higher audacity but struggle with their humility or have very high humility, but struggle with audacity. So one of the strongest compliments I can imagine giving to someone here is that those two things are both high and matched, coming up with a particularly gorgeous idea and being able to kill their own idea in the same meeting because we've sorted out that it's actually not as great as it looked in the first 10 minutes. That's like double high five. I'm jumping up to give them a hug.

Sam: Dr. Astro Teller, thank you.

Astro: Thanks for having me, Sam. 

#Moonshots #Innovation #ArtificialIntelligence #Sustainability #Regulation #Technology #FutureOfWork #Leadership #Teamwork