Dr. Astro Teller, Captain of Moonshots, X
In conversation with: Terri Burns, Fortune Brainstorm Tech 

Brainstorm Tech 2023: Fly Me To The Moon 

Conferences

Dr. Astro Teller, Captain of Moonshots, X
In conversation with: Terri Burns, Fortune Brainstorm Tech 


Joining us today is someone who has, has had to think big but also make sure these innovations. See the light of day, Dr Astro Teller, a scientist, an author and an entrepreneur with a pretty unique job title. Captain of Moonshot Dr Teller leads X Alphabet's Moonshot Factory where they dream up risky and audacious ideas that are grounded in science and technology, but do have the potential to push the boundaries of innovation and are not just pie in the sky. So without further ado, let's welcome him to the stage, Dr Teller and he will be interviewed by brainstorm tech co chair, Terry Burns. Thank you. Hey Astro, how's it going? Great. I'm very excited to get into everything X. Thanks for being here today. So again, chatting with some folks as we've been network in and chatting, I've been hearing so many people have had enthusiasm and have heard a lot about X over the years but not so much in recent years. So to start, I would love to hear a little bit more about you, your role, your amazing job title and also just in general, what's going on at X these days? Well, that was a lot. Let me start at the beginning to create a little context, right? So I'm the captain of moonshots at X. It's Alphabet's Moonshots factory. And our job basically is to invent and launch breakthrough technologies that hopefully can help tackle some of the biggest problems in the world and are designed to create at least the foundations for large enduring businesses for Alphabet. Awesome. And go ahead. Well, so let me give you a few examples of some of the things that have come from X just to give you a sort of a background, a backdrop. So sometimes we make things that turn out to be right to move back to Google instead. So Google Brain which actually has come up during this conference is sort of one of the early epicenters for the machine learning explosion that's happening right now that came from X Verily, the Life Science business that came from X Waymo, the self driving car company that came from X Wing, the drones for package delivery that came from X Intrinsic more recently, which is our Moonshot for industrial robotics that came from X. And very recently just this January, we put out a new other bet, a new part of alphabet called mineral, which is our Moonshot in the agriculture space. I'm curious to hear about you. You've just listed a number of amazing projects that have come from X. I want to talk about failure a little bit. What are some or an example of an X project that maybe you would categorize as a failure that you and your team have learned from. Well, one of the ones that comes to mind for me is loon. So we had this aspiration that if you could somehow magically get cell towers into the stratosphere floating around, they could go where they wanted to go. But then somehow stay, even with the high winds just where you wanted them to be and beam internet down to the ground you could create for people, particularly in outlying areas, not in large cities, connectivity that several billion people in the world don't have today. We were really excited about this. And amazingly, after nine years, we had it working these balloons, hundreds of them were flying around world were creating these ad hoc mesh networks in the sky. Beaming four G LTE and five G to people, particularly in Kenya and Peru or two of the early places. We were testing this. I remember the buzz about all of that and was so uplifting because of this goal of solving for connectivity. And it turned out the techno economics per bit were good but not great. They weren't as good as we wanted them to be. And we just didn't figure out a right way to solve with the telcos who are our partners who are going to be our connectivity to these people in various countries. And one of the things we learned from that is if we have to get everything working perfectly before the ecosystem will talk to us about whether they really want it or not, that's not worth it. So we're not going to do that again. That's a big takeaway that we had. Yeah, thanks for that. You all sort of know the drill. Now, I am going to turn it over to audience questions in just a minute. But before we do that, I would love to get here the latest, what are some of the big moon shots that you and your team are working on right now that you're really excited about? I'll give you two. But let me start with, we're fond of, of talking about Moonshot compost of recognizing that the people, the code, the patents, the learnings that come from anything that we do, even when it doesn't work out can supercharge the next attempt for us to solve the same or similar problem. So in the case of connectivity, there are still three plus billion people in the world who have little to no connectivity. We're still really passionate about solving that these balloons. I was just talking about up in the stratosphere were talking to each other using lasers. And so someone said his loon was winding down. Well, hold on a second. What if those things were brought down to the ground? Six years later, our project Tara is using wireless optical communications, these thin beams of invisible light, which can go up to 20 kilometers and get the benefits of fiber optic cabling without needing to trench cables. And we're rolling these out in Africa, in India and in a lot of other places around the world, we're really passionate about how well that's working. And amazingly, we're now moving more bits to real customers every single day with Tara than we did in the history of living. Wow, that's amazing. So you are literally able to take this failure if you will and leverage that technology to one of the most exciting projects that you're working on now. Exactly. Awesome. Let me give you another example. We're really passionate about the electric grid. The electric grid of the world needs a Moonshot. The electric grid of the world is the largest machine humanity has ever made. It's the most complex machine that humanity has ever made. It's the most expensive. So no one really knows what's going on with it. You would think that the people who move the electrons around, who make the electrons, the generators who transmit them, who distribute them at the edge would know where every wire was, where every inverter was, where every piece of equipment that it takes to keep the grid together, where it is, how it's functioning. And they don't, they just don't because it's too complex. So tapestry, our Moonshot for the electric grid is working to virtualize the grid to first help the grid just maintain what they have. If they don't know what's broken or what might soon be about to break, they can't fix it proactively. Then number two help them to be able to plan for the future as most people in the audience know right now the wait times on getting renewables onto the grid in the United States is like seven plus years and getting longer by the day, not shorter. And the reason is because the system operators don't know what it would do to the grid to plug a new solar field onto the grid. And so they spend weeks playing a few 100. What if games about the worst case scenarios to try to figure out if it was going to be safe or not? If you had a virtualization of the grid, then you could answer that question. Will it be safe or not in a few minutes instead of in a few weeks? So that's the kind of thing we're really excited about with tapestry and that's tapestry. Got it. Ok. I want to turn it over to the audience. Do we have any questions for Astro one up here? Hi, Catherine Ducker with Hewlett Packard Enterprise. I think all of us in this room probably massively envy your job. And I would love to hear a little bit about what factors guide the type of initiatives that your team pursues. What sort of that checklist of. Yes, this initiative has the potential to do Xy and Z we're gonna take it on at the most basic level. What we say is we won't even start on a journey until it has a Moonshot story hypothesis. And in order for it to have that, it has to have three basic things. Number one, you have to be able to name a huge problem in the world you want to solve. If there's not a huge problem you want to solve, then it's an academic exercise. Number two. And so that should be pointed at either a market that currently exists or a market that could be made if there's a huge problem that's being solved. Number two, what's the radical solution you're proposing and help us understand why it's radical because if it's what everyone else is doing, at least when we start, that's not the thing we should be in the business of working on. We should be exploring blue oceans, not the places that people are already fishing. And so those radical proposed solutions should in principle be science fiction sounding, but they're products or services once we develop them, if we can. And then the third thing that we require is some kind of technology that would be breakthrough if we could do it. And that gives us a reason to believe in the early days, we can test this hypothesis. It's OK. If the hypothesis is wrong, it's not OK. If it's going to take us many years to find out if we're wrong back to this issue about loon and getting as good as we can, as fast as we can, as efficient as we can. And answering the question is this very unlikely sounding Moonshot story hypothesis of yours going to make it or not? Because 99% of the journeys we start out on don't make it. So we can't say making it is what counts as good. It's how efficiently we filter them. That's where we think of getting our alpha. Any other questions from the audience? I'm curious to hear you mentioned Google brain. Google brain spun out of X 11 years ago and now everyone's buzzing about deep learning ai generative ai there's a lot of investors in the room myself included. If you could predict, what is that X today? That's going to be the big buzz 11 years from now. What do you think that would be first caveat? I don't believe in predicting. I believe that our job at X is to be efficient explorers so we can discover what's next. So I'm going to make what sounds like a prediction but it is, it has come through the process of exploration. Not like you know what I think would be great. Yeah. Not, not a crystal ball. Yeah, exactly. You got it. One of the things that I believe is getting to the place where 11 years from now we might be sitting around and talking about this subject primarily is synthetic genomics. I believe the movement from biology as a science to biology, as an engineering discipline is almost there, all of the pieces are just coming together and we don't have everything ironed out yet. I mean, we, the world, not just we x but I am seeing enough interesting pieces and early green shoots and enough even different parts of things we're doing at X that's starting to feel like one of those things that's going to be a really big deal in the future. Absolutely. I think that that's really compelling investors take notes. I'm curious to hear more about, we talked a little bit about failure, the opposite side of things. What does success look like at XX? And when something is successful graduates, how do you define that? What does that look like in the long run for a Moonshot to be successful? There almost inevitably has to be some tech innovation. There has to be product innovation and there has to be market innovation. There's kind of no way to innovate in any one of those areas without having to innovate in the others, particularly if you innovate in the tech area. So by the time it leaves X for sure, we need to believe that the tech has been meaningfully de risked. We understand it better. That doesn't mean that there's no risk left, but we should have moved to a place where execution risk is more than 50% of the risk. But we also need to see some early demonstrations on the product and market side that make us believe that this isn't just a build it and they will come kind of situation but that this really has a viable chance to become an alphabet scale business that could be really great for the world. Our final question is around ethics. X is trying to think about massive problems that theoretically can impact the lives of everyone. How do you bring an ethical lens to those explorations and the development of that, of those technologies with an X that's worthy of an entire hour of conversation. So I have 60 seconds, but in 60 seconds, I would say, I do not believe that having ethicists on board and somehow just passing it over to the ethicist to pass judgment on things is the right way to solve this problem. I think you need to wire an appreciation for we're in this with the world, for the world not to see if we can get some money and run before they figure it out, helping everyone to really feel that is job one and then job two is you don't know what's going to happen, what the unintended consequences are. So X has become expert at getting out into the world early in safe sandbox ways. So we can start to be in dialogue with regulators and with the public, not just in Silicon Valley, but all over the world. Where we test the things that we're doing. So we can say, how are you experiencing this? Which parts of this in Kenya? In New Zealand, in South Africa? In Peru? How is this better than we may have hoped? How is this worse or different than what we were hoping it would solve your problems? And that learning turns into an iteration with those people. And I think that's how we avoid the kind of unintended consequences that happen when people just mix until they're done and then sort of like dump it on the world at the last minute. Absolutely. Astro. Thank you so much for. Thank you. Thanks everyone.