The Google Engineer Who Set The Skydiving Record At 58 - EP 83 Alan Eustace

Sometimes there's a man. Sometimes there's a man who chooses to skydive from space and hurl himself 135,890 feet towards the ground. And we're speaking to that man today. He's Alan Eustace. He did indeed set the skydive freefall record in 2014. You probably think Felix Baumgartner, the Red Bull guy, holds that record. No. It's Alan, and he is a fascinating man. He was— he is a computer scientist. He used to work at Digital Equipment Corporation, which was, was the hottest computer hardware company in the world once upon a time. And then he went to Google, where I believe he was around employee number 50 and helped build much of the infrastructure that the internet runs on today. In all those senses, he is already remarkable man. But somewhere around, I think it was 2011, 2012, he decided— he, Alan, was a skydiver, he's a pilot. He decided he thought he could break the freefall record and do it from the edge of space. And he, along with the team, engineered his own system to go up in a balloon and then drop himself from space. He, unlike Felix, who did this with Red Bull and all the live streaming and all of that, Alan sought really no attention. He just wanted to do this himself. He is a brilliant, humble, fascinating man. This conversation is, I think, one of the best we've ever done. You're going to enjoy this. You'll hear the whole story of the skydive, and then we get into some computing history where Alan thinks AI is going. He's, he's seen the computing revolution. He's seen the internet revolution. He's a wonderful, thoughtful person to talk about all this. And then now Alan is exploring the ocean. Um, some people are just better than others and do wonderful, wonderful things. Alan is such a man. Please leave us a review or a like or anything, wherever you listen to this podcast. It, it really helps us out. It helps more people find the show and wonderful conversations like this with Alan Eustace. I had to chase Alan for many months to make this happen. I couldn't be more thrilled that it did happen. I am Ashlee Vance. This is Core Memory. Off to the edge of space we go! Alan, I've been stalking you for months. You're a guest that I desperately wanted to have on here. I just— I find your life so fascinating. So thank you very much for making time for me. I know you're very busy.

You're welcome. Congratulations on getting through all the armor.

I did it. I did it. And there's, there's many places we could go. Your career. Spans things I'm so fascinated by. I grew up reporting as a cub reporter on the Unix wars between Sun and IBM and Compaq and HP. And you worked at DEC. By the time I came along, I think Compaq had already sucked DEC up, but I was familiar with Alpha and VAX and all the history on that. And then, you know, and then you get to Google and it's kind of like my favorite time of Google when it was, it reminded me a bit of of almost like Sun Microsystems, which was my favorite company to cover, just really getting into computer science and advancing where data centers were heading. And then you go into a huge passion of mine, which is obviously this jump that you did in 2014 from space and all the history that led up to that. And so, and then now I think you're into eVTOLs and all this other stuff I'm into.

So. And underwater exploration. Yeah. Yeah. Using autonomous vehicles.

Yeah. I mean, I was going to say, start us with, with the jump because it's just such a unique thing. Are you, are you able, if we get there, are you able to talk about some of the work that you, you're doing today on the underwater vehicles? Okay.

Okay.

I didn't know if it was like stealthy or secret. Um, um, okay, well, let's do it. Let's do it. We were, we were chatting a little before we started. I ended up covering, there's a company called Worldview that made these stratospheric balloons. So they would take things way up basically to the edge of, of space. And, and I got into this company, I got to go see a launch. There's these two co-founders of the company, Tabor McCallum and Jane Poynter, who I got to know actually when I was doing reporting on the earliest days of SpaceX because they had kind of teamed up with Elon to see if they could put a greenhouse on Mars and, and grow, grow a plant on Mars. But they, they were always into space. They were always into pushing the edge of what's possible. And so they were doing these balloons. You're working at Google at the time, and then you, somewhere along the way, you have a, you have a desire to see if you could go up in, in one of these balloons, which we're going to get into. And then just as a result of doing all this reporting, I got so into the history of, of what we did to humans to, to see what, what the—

the early days were, uh, Uh, crazy. You know, it's just, just amazing. The only difference between what you said is that I called Jane and Tabor weren't into balloons until I called them.

I see. Okay. Okay.

The way it actually worked is, you know, I had spent like a lot of time, like I made my first skydive in 1975 and, and any skydiver will ask, first thing they'll ask is like, what's the highest skydive? And the first thing that leads you is Joe Kittinger. And Joe Kittinger was like an amazing human being. And what they tried to attempt in that group was at the edges of what people, what people can endure. Like famous, you know, Stapp, you know, he straps himself into a rocket sled, propels himself. He doesn't even want other people to do these experiments. He wants to do them on himself.

Yeah. Can I, can I set people up with a little context on this?

Yeah.

So like, like mid-1950s, we're starting to get planes that are going faster and faster. The US government knows we're probably going to send people into space. They're like, well, what can the human body endure? We don't know if you go that fast in a plane, do you just disintegrate?

Do you know what happens? Doesn't it remind you of the earliest balloon flight where they took a duck or something and put it up? Like, I don't know what's going to happen.

Let's see what happens to you. Yeah. And so, so there's this guy, John Paul Stapp. He's, he's born in Brazil. He grows up in Texas. He's a medical doctor, and they're sort of like, hey, why don't you see what humans can endure? And it's one of the wildest stories I've ever run across. There's a great book called, um, The Pre-Astronauts by a guy named Craig Ryan. And then Craig goes deep on all this stuff. He also writes a biography on John Paul Stapp. I forget the name of that. Um, Sonic Winds.

Oh yeah.

And, and so there's this amazing history. This, this guy, I always laugh because Stapp was, uh, he's incredible. And so like you mentioned, he, he goes out and it's either New Mexico or Arizona and builds this sled track and it's going like 700, 800 miles an hour on a sled. And then he'll just jar it to a stop to see if, if, if people's eye— eyes are coming out of their sockets, they're cracking their ribs.

And, and the point was by understanding human physiology, and what can you do to protect it? So it was, it was this yin and yang thing. Like, for instance, in the early days of flying, these— they didn't protect the pilot at all because the assumption was that every time you crash, you die. So let's not worry about that, you know, that. And so, but Stapp was proving that the human body could endure an enormous amount of G-forces and survive. And so they turned that into cockpits and setups and seatbelts and restraint systems that were state of the art. And, and after all of that, in the later part of his career, he was responsible for trying to get seatbelts into cars for the same reason. And this idea of what the human body is capable of doing in every different positions has had like a dramatic effect on that.

He basically invents the seatbelt. You could argue he might have saved more lives than just about any— I don't know if you count penicillin or something. It would be hard to track.

The airbag was also traced to him. I mean, you know, the early days. So he's, he's just amazing. And so I— Kittinger led me to that, you know?

Right.

Like the fact that I had read about Kittinger and the history, like, you know, the other people that had tried, you know, stratospheric skydives or people came after him and the failures and like, So for me, it was like, as an engineering problem, it's like, how do all these pieces fit together? And he did it, you know, it was 1960, you know, it was like, it was in the wildest, amazing time of engineering, you know, like Concorde and SR-71. And, you know, all these amazing things were happening in that period of time. And so I was just fascinated by it and loved it. So, so when I later on, when Red Bull is trying to do their attempt, which had a giant capsule and all this stuff. And I had what I thought was like a simple idea, which is why do you need a capsule? Like, you already got a spacesuit on. Like, why do you need that? You know, why transition from one environment to the other? And Nick Pinchinata, he was trying to break Joe's record, and he he couldn't disconnect from the oxygen system. And because of that, he had already undone his seatbelt. So he tries to disconnect from the oxygen system and can't do it. I mean, he's struggling to do it. And then all of a sudden he realizes he's going to have to parachute down from this high altitude, release himself, round parachute, swinging back and forth, holding onto a seat as hard as he possibly can. And it's like, why do you need to make those transitions? Because those transitions are where the problem comes.

Complexity.

Yeah. And the same thing when Felix left left in the Red Bull Stratos thing, he only had like 10 minutes of oxygen. Like, he had to figure out a way to cut away a reserve parachute if it actually came out ahead of time. So, so my idea was, well, let's just take a tandem parachute, tandem setup. Let's not put a passenger on the front and skydive. Let's just put a couple oxygen tanks and let's go up and do it tomorrow. So that shows how naive I was. But so then I did an enormous amount of research and, you know, you can't do that. I need to have a spacesuit. You know, the, the system itself, I mean, you know, the basics were there, and it will create— make a much lighter weight system, which means you need a much smaller balloon. So I had done the paper study on it, you know, a Google Doc, you know, that actually described everything.

Can I ask you about that?

Sure.

Okay. So, so Kittinger in 1960, he sets this record as something like 108,000 It's kind of a little bit debatable.

It's around 105,000, 105,000, 104,000 and change.

Okay. So, and he's part of this Project Manhigh. He's working with Stapp there. Stapp had done all the, a bunch of horizontal stuff we were talking about on the sled. And then it's like, okay, well now let's see what happens when you go real high. So the, the send up Kittinger to the, to the, you know, in the edge of space, just a bit below the Kármán line. And, and then all these years pass. I mean, I mean, that's 1960. And then Felix—

well, let me tell you one other thing about this interesting about the Joe Kittinger thing is like people don't realize there were 3 jumps. It wasn't just one jump, it was 3, 3 jumps. And on the second jump, Kittinger is, is going up and he's trying— he, he— the way it was set up is with a timer. So a drogue parachute will come out with timer. And he was testing at the time not just the suit and all this kind of stuff, but he was also testing parachute deployment systems. And the reason is, is because now jet— like, as you said, jet fighters were getting up over 100,000 feet and he wanted to test systems that would keep people alive in ejections from that. And so they were developing this parachute system called the Bopom system, I think it was called. And at the same time, so they were innovating in a huge number of areas. And so, so Joe is in this chair. He tries to get out of his chair. At, you know, I don't know, 90,000 feet, 100,000 feet, whatever it was, try to get out and he can't get out because of the expansion coefficients were different. So it was colder up there and therefore it was cooler and the system would not let him out. So he's struggling to get out of this thing and it was just a little square box that he was kind of in. So he struggles to get out and inadvertently during that struggle, he sets off this timer. So now his drogue parachute is going to come out super fast. So he actually launches himself off this platform. The drogue, instead of waiting 15 seconds for him to speed up, comes up immediately. It wraps around his neck— worst possible place for it to be. He goes into a spin, wakes up because they happen to have invented the automatic activation device that skydivers now use routinely, but But back then it was an incredibly rare thing. So the parachute opened itself. He wakes up under a parachute, lands, and then goes back and does it again.

Yeah, yeah, yeah.

Just set the record.

So in this hero, in the Pre-Astronauts book leading up to that, I mean, the things he endured are insane. Like, like you mentioned, I mean, we didn't know so much of this stuff about the temperatures, what was going to happen. So he's got a rectal thermometer in the whole time. There's times when they install the machine that's supposed to produce his oxygen, I think they put it in upside down, so now it's just like pumping CO2 into—

His glove failed on the front of the, uh, on the last jump, and so it depressurizes his hand, so his hand swells up to the size of whatever the inside of the glove was. He doesn't report it to his— to the, uh, to the doctors because they're afraid they'll abort this mission you think is so important.

I mean, there's this part—

what a guy—

there's this part where I laughed because it almost seems like the Army is— the Army is like, you guys go find out what can happen. And then it almost seems like Stapp takes it way more seriously than they were anticipating. And people's hands are swollen, their eyes are coming out of their socket. It's kind of like, maybe, maybe we should dial this guy back a little. This is—

this is— Stapp went blind on one of the tests. I mean, literally, he had blood in his eyes, and they weren't sure whether he'd regain his sight. And he knew that was a possibility. He's a doctor. Yeah, he went through with it anyway and he regained his sight. Yeah, it's just— I mean, it's just a different era of understanding. And you see, when you see that as a pre-astronaut thing, all of a sudden you get to see like, how did the Apollo program happen? I mean, how is it possible with that technology in those days to do things? And part of it is that they just had this kind of can-do, nothing will stop me. Pull it together, you know, like their risk profile on their, their vision and their risk were so much different than what we have today. Uh, and that's why it was possible. Yeah, otherwise it wouldn't have been possible.

All right, what do we do at Core Memory? We cover innovative, fast-moving, forward-thinking companies, which is why Core Memory is sponsored by Brex, because Brex is the intelligent finance platform for many of these companies. 30,000 companies from startups to world's largest corporations rely on Brex's technology for their finances. They've got smart corporate cards, high-yield business banking, and expense automation tools that are fantastic. I hate doing my expenses, and Brex's AIs and software run right through those expenses, figure out where we're spending money, and take care of so much stuff for you so you don't have to waste your time on it yourself. Go to brex.com/corememory to learn more and just, you know, get with the program. Let's get going. Let's get out of this archaic finance software and move toward the future. CoreMemory and Brex. There's this history, like you mentioned, you've been skydiving since 1975. You, you've, I think you fly all kinds of planes.

Yeah. I didn't start that until I got out here, so I was a late learner flying, but yeah, so now I, I've flown a lot of things.

So you're into all this stuff, you're— and then if Felix went in 2012, what was the year that you first started thinking about doing this?

Uh, it was 2011.

Okay. And you're just sitting at home one day and you're like, you're like, why don't I try to break the record?

Well, I think I was reading about their effort because they were, they were pretty open about it because it was a— it was for publicity, right? So, and to Red Bull and Blend, it was by far the largest event, like livestream event that had ever happened. It was like— and it was wonderful engineering. My friend Art Thompson handled the engineering side. It was amazing. I have nothing but respect for that team and what they did and stuff like that. But I'm looking from the outside and saying, that's not how I would do it. Like, that's just not how I would do it. I would do it this other way. And so then, you know, because I mean, I always tell people I don't pick problems, problems pick me, and then I just can't let them go. And so I just kept working on it, working on it, working on it. It's too bad that, you know, we don't have the era of like, you know, DeepMind and stuff like that because it would take— it would save me a lot of trouble going through all this, you know, all the, you know, like what size balloon do I need? Do I have an oxygen system that's big enough? What happens to the body when it gets up? There's all these questions. But I worked for, you know, just a little bit, you know, at a time. But it took me, you know, probably 6 months to get to the point where I said, You know, I think this closes. I think this will work.

Did you tell your wife you were doing this?

Oh, yeah. But she would, you know, it's like, oh, I think of all these weird ideas all the time. You know, she didn't take it seriously enough, which was a problem. She, you know, like later on she told me, I shouldn't let you get that far down the road. Yeah, but I didn't think it was going to be successful because once you start things, once the ball's rolling, it's really hard to stop it. And she would definitely have liked to stop it if she could. So anyway, so, so, so anyway, so I worked out all the numbers and stuff like that. And then I wanted to find out whether I was on track or not on track. So I called Peter Diamandis from XPRIZE, and I— he was a friend, somebody who had come through Google a few times. I met him. Very broad man, a lot of background. And I asked him, who's the person that I can talk to to see if this you know, this thing works? And he said, I'd call Tabor McCallum. Yeah. And so he made the introduction. Otherwise, I don't think Tabor would have been taking the call. And I call them, I work through what I'm doing, what— why I think the things I need. And I said, what do you think, Tabor? Is this— is this possible? And his company, Paragon, builds environmental systems for toxic environments and space environmental systems. So like, they're the one company in the world that actually has all the pieces. They do. We're in this with all the space suit companies and stuff. And he says, not only do I think it's possible, but I would love to do this project with you.

So, okay. And then Tabor and Jane, they had done Biosphere and all these different things. And then, yeah, like, I mean, essentially they're very deep part of the origin story of SpaceX because Elon had made some PayPal money and he was trying to see what NASA was doing on Mars, and it turned out to be not a lot. And so Elon thought, well, to get people interested in Mars again, I want to send a greenhouse up, up to Mars, have it be— have— and he wanted to livestream a plant growing, you know, from Mars just to get people to kind of be like, oh, we should— this is a thing. We should think about terraforming this planet and exploring Mars. And, and so he recruits Paragon to help out with this project, which obviously—

Wow.

Yeah.

I didn't even know that part of the story.

Yeah.

That's really cool.

It never— what ends up happening is then Elon starts looking into, well, how would you get it there? How much would the rocket be? And so he starts digging into the rocket business and then he decides, oh, well, cuz he goes to, you know, cuz he is starting to go to Russia, um, to try to buy a Soyuz essentially for this mission. And he's doesn't like dealing with the Russians and he's looking at what's available in the US. He sees it's all extremely expensive. And so this— that's how SpaceX begins. Then he just decides to go do rockets and, and now here we are. So, okay, so and then when you, you dig into this and I don't want to be crass or anything, you know, you were an early employee at Google. I'm sure you, you did well financially, but I would think the sticker price on this would be like tens of millions of dollars. Do you start to think about that part at all of like, God, does this even make sense to—

Well, this kind of goes to like extreme optimism, you know, it's like, I think this is like not that hard. It's not going to take long. It's not going to take very many people, you know, it's like, and you know, this is like a common theme around here. It's like, people way underestimate, like, how hard something is. Yeah. And so, and that's what I did. I mean, I just underestimated. I thought it was going to take a year, uh, and I thought it was pretty straightforward to get all the pieces together. And it turned out, um, and, you know, maybe we could have thrown something together in a year, um, but not with the level of safety that was— that I required, my wife required, my family required. And, uh, and also that Paragon required too. And so, um, and so everything just took longer. And we, we, you know, we were actually prepared to do it at the 2-year mark. We were, we were getting closer and closer until we did the first airplane jump and in a spacesuit and all that kind of stuff. And like, everything went wrong. I mean, not fatally wrong kind of thing, but like all the little things of trying to put a complete system together, all the little pieces didn't hold together. And so I ended up not being able to steer. I couldn't depressurize the suit. I landed in the middle of a desert. You know, my team had to come get me. I'm going to say it was a low moment of the whole thing. And that required an entire year of redesign to fix.

Okay.

To fix, not just to fix what was broken, but to rethink some of the core decisions that we made and see if they made sense in light of the current— what we had learned in this job. And so then after that, it went super fast. We did 5 airplane jumps and fixed all the problems. I mean, those jumps were almost perfect, except it took a little while to figure out how to fly. Like, that was my problem. Like, what you do in skydiving and what you do in a spacesuit are exactly the opposite. So you have to totally unlearn something that you learn in skydiving. And relearn this new thing. But once we figured that out, and that was the safety divers that actually looked at the video and figured out what, what you needed to do to control this, uh, and so, so once we got that done, we actually did 3, 3 balloon flights in like under a month in which New Mexico— okay, okay, it was actually— and we had 4 dates. One of them, one of them we had a, a problem with a, uh, with one of the systems to bring the balloon down. And so we had to cancel that one. But, but 4 opportunities, 3 of them worked inside a month, which I think might be a record.

And so you're paying for like a team of how many people? Yeah.

So if you— I don't tell anybody how much it costs. So, but I was talking to a team that's not trying to do that, break that record, but another one. And I said, you know, if you want to figure out the cost of something, the way you do it Like, first you forget, people are always fascinated about what's the cost of a spacesuit or something like that. Forget cost of the spacesuit. Just figure out how many people are gonna be there the day of the event. You know, like there's the people, you know, obviously that are the medical team. You know, there's the spacesuit people. There's the people that transport you to the thing. There's the people that do the balloons. There's people that interact with the FAA. There's the people that are,, in mission control. So figure out, and then figure out how much training over how long and where they're going to do it, which is all in Roswell, New Mexico, far from home. And then you look at all the little pieces and all the tests you have to do, and you start adding up all the salaries, and then you get a number, right? You know, and so I think that's where people miss, is that, is that every engineering endeavor is about people, and people are expensive.

Did you have an upper limit of where you would cap things off?

Uh, I mean, yes and no. Um, but I looked at it differently, which is I looked at like, like, is this test necessary? Like, do we have to do this? And we ran like 250 tests. I mean, we ran a lot of tests. I mean, not all of them involved me, but like the strength of different materials. I mean, we had a, we had a test to try to find when the, uh, when the failures of nylon would happen and at what speeds. Like we put like giant rockets made of concrete that we dropped out of airplanes and they whistle down. And then we at 200 feet throw out a parachute and then see what's going to happen to these materials. And, you know, I mean, they're very strong materials and we managed several times just get them to totally, you know, like disintegrate. Yeah. Because of the forces that are involved. So to learn And it's impossible to learn these things other way. One of the rockets, missile, it's not a rocket, it's propelled. It's just a gravity thing. But it hit and the team found where it hit, you know, and they walked over and they knew it was where the thing, there was no sign of actually any part of the rocket, but they knew where it was because there was a small depression. And when they put their hand there, it was hot. And they dug down and found a fin and sent it to me. Holy cow. That's how much force was involved. So these tests, I mean, just think about what you have to do to ensure safety. Like, you know, another example is spins. Like, we were really worried about spins. We thought a drogue would solve a spin. We took the, you know, we took a dummy up to, you know, 120,000 feet, dropped it, spun, I mean, 170 RPM or something like that. It's like, no, that's not going to do it.

Bad news, yeah.

We got to go back. And actually, the Red Bull team was actually really helpful. Luke Aikens, was incredible. He actually— they had a backup emergency parachute system that Felix never used. I came close to needing to use it, but so I went and we went and talked to him and he told us the secret, which was like the smallest little change. You know, it was changing attachment point of a drogue from like right, you know, like on your butt to above your neck. That's it. You just move that attachment point and you go from 100— it'd probably be 220 revolutions per minute from that altitude down to like 7.

Okay.

Okay. So, yeah, incredible things you learn along the way.

Okay. So you're reading about Felix is going to make this— Kittenger's record has been there since 1960. You're reading that the Red Bull team is going to try to break this. You think you have a different approach. They are doing this, you know, it's Red Bull. They're doing this massive media campaign around all this stuff. He's presumably— well, he's going to try to to break the record and then they have a head start on you.

Yeah. And then 10 years. I mean, it was a long head start.

Okay. And then— and so you— part of your motivation is clearly you think there's a different way to do this. And then is part of your motivation, well, I'm going to beat the record that Felix sets too.

You know, I— the, the main thing for me was I just thought there was a better way to do it. And in some sense, it didn't— like, I actually thought we were going to beat Red Bull to it because we had a year.

Okay.

They did in 2012. We started the project in 2011. And so I actually thought we had a chance of beating them to it. But of course, they beat us. And not only did they get there first, but they went to 128,000 feet, which is like— like, Kidinger's record is higher, 4,000. You have to beat a record by 5% for the FAI. They didn't have to go that high, but they did, which was impressive. It was just impressive. But, you know, the, the ace in the hole that we had that they didn't have is we were a lot lighter. Like, they weighed 3,200 pounds and we weighed 600 pounds or something like that. And so even though we were much later, our system's max height was way higher than they could have. So we actually beat the record. They used 128 Um, uh, let's see, what was the, uh, what was the, uh, they used a 28 million cubic foot balloon to get to 128,000. And because of the way our system was designed to lightweight, we used an 11 million cubic foot balloon, okay, to get, you know, 7,000 feet higher. Okay, so we had an— we had a technical advantage.

All right, pod friends, I am here to tell you about SendCutSend. They are a manufacturing phenomenon. Whether you're working on your own car or you're a giant automotive company, a rocket maker, if you need metal cut, machined, or bent, you're going to SendCutSend. They'll do it for you right, and they will do it for you fast. Go to sendcutsend.com/corememory for a 15% discount on whatever you're trying to make. We love SendCutSend. Go visit them.

Today.

And then you choose not to do really— I mean, you're like doing this in secret and you're not making a TV spectacle out of this. I mean, you just wanted to do this just for you. I, you know, like there's no live stream. You—

nothing. Yeah, no, we had— they had like 700 reporters. We had one reporter.

Markov.

Yeah. Yeah.

Okay, so here's the story for people. John's a dear friend of mine and Here it is, New York Times. I mean, this is a printout, looks better in the paper. This is October 24th, 2014, parachutist record fall over 25 miles in 15 minutes. And John Markoff, you know, probably the best tech reporter of—

And I loved Markoff. I mean, I just felt like he's one of the few people that I think really understands tech at sign of kind of a, it's almost, he understands the tech, but he also understands people and he understands importance of things.. And I just thought he was— he'd be a beautiful person to, like, be able to tell the story.

Great choice. But the funny part is you should have picked me. But John's okay.

The funny part about the whole thing is, is John, like, they won't let me pay for anything. New York Times won't let me pay.

Yeah.

So he gets one trip to Roswell to watch a launch. And the first one that we had, he decides not to come. And I don't know whether that's why, but he had some kind of conflict. And that one, we had a failure. We didn't actually get to launch. So he comes to the second one and that one's successful. So, you know, and then he writes this one article. But part of it is, you know, it's an engineering story. I really just wanted to share it with my team that we worked 3 years together. We loved each other. We were a band of brothers and sisters that kind of came together to pull off this thing. And, uh, and the other people I want— and we had— I had a few family members and stuff there, but, uh, but that's the group that I wanted to share with. Yeah, I didn't want to share with a million people. Also, there's pressures that come from having a lot of people watching, and, you know, like maybe a launch on a day that you probably shouldn't launch or something like that. So, so for me, I, uh, that wasn't what it was about. And also You know, I just, I won't, if you, like the jump happened, I came back, we flew helicopters back, it was really nice. And you know, there's like 40 people that are standing there and we could all hug together, we could celebrate together. We went to dinner together at the Mexican restaurant, which by the way is an interesting thing. So we're at this, we're celebrating this Mexican restaurant. And we're all eating tacos, drinking margaritas, having a great time. And I get a text from my sister Barb, and she lives in Altamonte Springs, Florida. And she goes, you'll never guess who I'm standing next to. I said, I have no idea. And she said, she goes, Colonel Joe Kittinger. And I said, prove it. And so she basically sends me a picture of her and Joe together.. And she said she, she talked to the bartender, said her brother broke this record. And the bartender says, you want to meet the guy who broke that record in 1966?

Just a coincidence.

Total coincidence. Walked up and she, she said to Joe, hey, my brother just broke your record. Like, no, like, introduction or anything. And so Joe and her hit it off. And, and he said, please, could you take a picture of my my business card and send it to Alan. I'd love to talk to him. And so of course the next day I'm like, like shaking as I call, you know, like the Joe Kittinger. And he says, uh, and, uh, so I call him and he is like the perfect gentleman. And then, you know, congratulatory and all this. And then he starts hitting me with technical questions. Like, like they're the perfect questions. Asked in the exact order.

And he's in what, his 80s?

Yeah, he was in his 80s. And he's like, he's like, it was like, it was like the coolest technical discussion. So then he goes away. We say, oh, we'll get together sometime. But, you know, it's like, you know, the song, he said we'll get together, but yeah, you know, it'll never happen kind of thing, the Harry Chapin song. But anyway, so then my mom calls me and my mom says, hey, do you mind if I invite Joe? She's not Joe Kennedy, Joe over to Christmas dinner. And I said, sure, Mom, you can invite him. But I mean, what are the chances? This is Joe Gittner. And so the next time, the first time I met Joe Gittner in person was when I walk into my mom's house and there's Joe and Sherry sitting in the living room.

This is crazy.

What are the odds?

This is insane.

I know. It's like, oh my God. Anyway, amazing guy. And I, for the rest, for, you know, for the duration of his life, he and I were very close and we spent a lot of time together. You know, at many, many events. And, and God, I just love that guy. He's great.

It doesn't seem like you're the kind of person this would annoy, but if you go to some random person on the street, you're like, do you remember that guy that jumped from the edge of space? They'll be like, oh, the Red Bull guy, you know, because of all the spectacle attached to that. I mean, do you— does that bother you? I mean, I don't know that.

No, no, it doesn't. Actually, it's kind of like But you're exactly right. You know, if you, if you go to somebody and says, who holds the world record? Or you say, so like, I'll, I will never admit that. I, I mean, I'll, I'll admit it, but I'm never telling anybody, hi, I'm Alan Eustace. I'm the guy who has the older, holds the world record. That's just not me. But somebody else will say, you know, it will introduce me and say, you know, Alan holds this world record. And like every single time they'll say, I watched that on livestream. That was the most exciting moment. And I have to say, no, sorry, that It wasn't me. Yeah, but it was— they're right, it was the most exciting moment, you know, in a day. And I still— like, here's, here's the, here's a funny thing, like, it— that comes up in my feed all the time. Like, I don't know, I want my YouTube feed, like, I don't know, every month, his jumps come up in my feed. It's like, uh, anyway, kind of crazy.

But how— remind me how old you were when you did it?

'58.

'58. I say this next part with all due respect, I mean, you know, you're thinking, uh, it's a Daredevil type feat, you're thinking of, uh, Chuck Yeager kind of thing. I mean, here you are, you're a computer scientist, 3 degrees, you're working at Google in the data center stuff. And, and, you know, I wouldn't be the, the immediate picture of the person who's going to set the world skydiving record.

Yeah. And it was funny because there was a website or something that came out right after the jump, and it did a head-to-head comparison with me and Felix. Like he was like young and attractive, you know, like a daredevil. And he did all these base jumps in here. And like, I run an engineering organization, you know, I graduate computer science. Like we could not have more different backgrounds than we did in that. But, you know, for me, it's like people say, oh, that's just such a daring, you know, thing. And it's like, no, that's You know, it's exactly the opposite. You know, it's like, it's like great engineering takes things that appear to be really dangerous and makes it really safe, you know? And that's how I approach these things. There was a guy, Eric Roberts, who worked at Systems Research Center at DEC. And I one time asked him about the Diablos. He's a juggler, big juggler. He taught me a lot of juggling-related things. And he said, I said, do you ever do the Diablo? And he goes, no, that's not what I like to do. And I asked him why. He goes, well, that's a trick that looks easy, but it's actually quite difficult. And he goes, I like the opposite. I like things that are actually easy but appear to be quite difficult. And so when you look at the effort, the whole thing was how do you take out risk? And behind the scenes, the Red Bull team was doing the same thing. Like they were pulling out all these risks so that something looks dangerous was simple. But for me, You know, that's really the key is that, like, I could not come up with a, you know, one of the things my wife made me do, she had a long list of things she made me do before the jump, but one of them was to do a video for each of my kids in case something went wrong. And so I did, I made a video of each of those. But what was interesting is, like, there was, you know, I could not imagine what could go wrong. I mean, it's possible that that there's something we didn't think of, like an unknown unknown. But every single thing that we knew about, we had mitigated. We had multiple redundant systems that save us. So for me, the surprise, if we missed something, would have been just so large. And so, you know, it wasn't a daredevil exactly. I mean, I, you know, risk, we had done such a good job of like eliminating risk. And, and, you know, I'm like big on actuarial tables and it's like, you know, at this, at this, at my age right now, not at 58, I don't know what the numbers are at 58, but you know, there's a 2% chance of me dying this year. You know, it's like, I mean, that's just fact, you know, maybe I'm a little better shape, maybe a little worse shape than the average person, but, um, uh, But so there's always risk, right? And there's no way we, I mean, I looked at this thing and it's like one in a million. I feel like the chances of something happening. So for me, the cool engineering was how do you take something which appears to be dangerous and turn it into something which is very, very safe.

Magic or something. So how many people have had to make a video like that? How long?

It was relatively short, actually. She had like 10 or 12 things that she made me do. I was like, like, I love my wife, but she's like, I had to— she was one of the overwhelming themes of the list itself was, I'm not going to clean up the mess that you just created for me. So you're going to have to do it. Like, like an example is I had to write my will. That's kind of fine, you know, trust, all that kind of stuff. But then I had to write my— I wrote my own obituary, which was short, pretty sweet. Honestly, he was a really smart guy until the end where things didn't go so well. I wrote a check to Patrick Poucet, who was the CFO at Google at the time. And I said, I hope you haven't cashed it. But that was the amount to basically do the whole funeral arrangements. And he said he's never been so happy to return a check uncashed, you know. But yeah, she, you know, like the process of taking a body and moving it from New Mexico to to, you know, California if something went wrong or something, you know, just those kind of things, which for most people, I think, I mean, what you want, it was for me to think twice about the whole project. But for me, it was kind of just a checklist, went through each of them. Yeah. And, but, but yeah, it does make you think.

Yeah. No, absolutely. And I know you've had to do this a million times, I'm sure, to at least your friends and people around you. But going through the experience, I rewatched the video this morning. I mean, the, There are a bunch of parts that struck me, but you're in, you're in this suit and when it's on the ground, I mean, it just looks incredibly unwieldy. You're kind of like on your, like a turtle, you know, like that's been turned over on its shell.

That's a great description.

Now I think about it before you get to take off, it looked, it looked kind of uncomfortable, but you're throwing up, uh, you know, thumbs up sign and, and you seem to be talking like you seem pretty relaxed about the whole thing. And then Then you kind of shoot up. And so what's like as you're going up and again, you've gone through, you've done all these tests, you have some sense of what this is going to be like. And I'm sure you're just kind of running through checklists in your head and they're, they're running through checklists in the comms. But is it, is it relaxing going up? Is it exhilarating? What was that? Are you at all having a second thought?

Well, the one test that I think was the most— like the airplane test, like the first airplane test, there was a lot of stress associated with that. I mean, we had never done it before. It's the first time that a spacesuit had ever been kind of out in the wild with— first time I had been around a spacesuit that I didn't have, you know, two space tech, uh, spacesuit techs, like standing right next to me with hook knives ready to cut me out at any moment. So, so there was a little bit of stress associated with that. And then obviously when the jump went wrong in a bunch of different ways, um, you know, there was a little stress there, but I was mostly relaxed about that. I mean, in some sense, I remember the days in the early days of skydiving, you had round reserves, which we call them pointable, not steerable. You can point them in any direction, but you were going to go wherever the wind went. And that was kind of the situation here. I was I was under canopy. I was going down slowly. I didn't have control over the canopy. We were going to land in the desert, but we were in a place where, you know, you could pretty much land anywhere. So, so, so for me, you know, that piece wasn't— and then we got more and more comfortable and confident with the plane jumps. We fixed some of the problems of the first one, all the problems of the first one. But then the first balloon jump is where the whole system has to work. So we had done independent unmanned balloon jumps, but now we take me and we put me underneath a balloon and we actually do a real launch. It's kind of funny. It's only like 60,000 feet, which in some ways to me seemed really low. But I called my friend, you know, Scott, who had helped me on a lot of projects. I said, well, we just finished the first one. It was only 60,000 feet. He shakes me, goes, Alan, he goes, only 60,000 feet? I mean, you're like in the— You know, top 5 people ever, you know, and the, you know, the number of people that have been in a spacesuit above 60,000 feet on their own and done a skydive is like almost zero. And so, so anyway, he had to shake me out of it. But, but that was the jump where that was the, the launch and the jump that I thought had the most risk because there were the— and then every time we added something that added additional risk, but we thought we had it covered, but we were very careful, like the first one was there's a zone at which you cannot open a parachute. You're going too fast. So at the beginning, you can't open a parachute because you're going too slow, although our drogue system made it possible to get the drogue out, but very quickly. But then there's an area where you're just going too fast. The parachute, if it was to open, would cause either damage to you or the parachute. So, so there's a, there's a zone. And so by going at 60,000 feet, we reduce that zone. Like, we made it pretty small. It was only like 16 seconds or something like that before we got out of that zone. And so at that point, there's 16 seconds you couldn't open the parachute, but the whole rest of the time you could. And so then the next jump— and that jump went really, really well. Everything was like perfect on it. And the next one was 105,000 feet, and, uh, that was mostly to get us in the range of Joe's record, not, um— and we might have even beaten it slightly. There. But that was subsonic. So long delay, but subsonic. And then, uh, uh, sorry about that. So, um, so there was a long delay, um, but it was all subsonic. And then the last one was where supersonic came in. And so that was all about how to control yourself when you're going supersonic.

So, and you'd— so you'd broken the sound barrier before the final jump?

No, no, the final jump was— it was the only one that went supersonic.

Okay.

And so the previous one was long delay, subsonic, and the last one was very long, but not even that much longer because you're going so fast, but then supersonic.

Okay, okay, okay.

And you—

so when you finally end up doing it, how long are you— how many minutes are you doing the rise up to your final?

It's not that long. It's 2 hours and 7 minutes.

No, but I mean, when you're doing the float up to the, to space, to the edge of space on that final jump. I mean, isn't that, isn't that like 12 minutes or something?

Or no, no, the, the, from the ground.

Yeah.

When we launch, the balloons are really interesting. They, they go up at about 1,000 feet per minute, maybe a little more. And so when you launch, and they don't, like airplanes kind of taper off, you know, they, they struggle with less. Not a balloon. The balloon starts at 1,200 feet per minute, ends at 1,200 feet per minute. The only reason it stops is the balloon fills up and it starts venting gas out of the bottom, basically, which keeps it from exploding, which is a good thing. But 1,200 feet per minute the whole way. So when you launch, you know exactly when you're going to get to float altitude. Matter of fact, I kept asking, like, when are we going to get there? I felt like a child in the back of a station wagon. When are we going to get there? So I actually, they bought me a watch, you know, that long story about the watch. But anyway, so 2 hours and 7 minutes on the way and on the way up, totally relaxed.

So you're just for 2 hours and you're going up eventually to 25 miles. And so you're just, you're just there thinking about just relaxing.

And I have done so much emergency prep on the previous ones, like, you know, and I did on this one too. Like, you know, if I turn, if I start turning right and if I can't control it, what do I do? If this thing's, if that, you know, do this. If that doesn't work, then do this. If that doesn't work, do this. You know, it's like, like I had all that. I had grilled that so many times in all the previous ones that I spent a little bit of time on it. And also then if I open, you know, when I open, if there's a malfunction, what am I gonna do about it? You know, what handles do I choose in what order? And so I went over all of that like so many times. So in the final chop, it was actually like super relaxing.

Okay.

Like I was as close to sleeping as I could possibly get because the more relaxed you are, the less oxygen you use. That actually.

Yeah.

So yeah, it was funny because on the previous one, the heart rate monitor worked, but on this one, the heart rate monitor failed. And so we'll never know what my heart rate was, but I'd be super surprised if it was more like 70 or something like that.

Okay. Okay.

It was, it was a very, very— and I tell people it's like Google Earth in reverse, you know, it's like like you just, you start, you start by seeing pebbles on the ground and then you start seeing small rocks and you start seeing, you know, you've got people and later see cars and then you see cities and then you see, you know, towns and then you see states. And I mean, it's like, it's like this amazing gradual thing. And I remember one time I was saying, you know, people ask me on the last one, could you like pay attention? So could you see the Gulf of Mexico? Could you see the Rocky Mountains? Could you see this other thing. And I remember one of them, they— I said, no, I can't see the Rocky Mountains. And then I went and looked, and I could see the Rocky Mountains. They just didn't look like mountains. They were looking like tiny little bumps down there, kind of thing.

So anyway, I think it's the Gulf of America now.

Oh yeah, sorry about that. I grew up in Florida, right? I swam there a lot of times.

So you go it up, and then when you Is everything okay? First of all, so the suit is giving you oxygen and then temperature control. Any other, like, giant—

Well, the temperature control is an interesting one because, yeah, the suit is basically providing everything that I need. But there's so many nuances of it. Like, one thing that's interesting about spacesuits is I blew the spacesuit up from inside. Like, I breathe in and then I breathe out. And that's what expands the spacesuit. So the idea of the pressure inside the spacesuit is not like— I always thought, you know, it was like putting air in your tires or something like that. Shit, no, you actually— it starts out, you know, like, like around you and then slowly you blow it up. And then when it reaches the pressure at which it's destined to operate, it has a dual suit controller which basically vents at that point. So So then it just starts venting. So it stays at exactly the right pressure. But, but it's, you know, it's like, but the suit, it was interesting because we thought we were going to be really cold because Stempeljup, it was like super cold. It's like -120 degrees or something like that. So we built an entire system to keep me warm. It's like a heater and it's got these, these tubes that go in and around me and they come back to the heater and pumps and all this kind of stuff. And we tested that by buying a giant walk-in freezer and then attaching it to a liquid nitrogen tank and putting fans in the air. We got these cool videos of all that. That test we ran like 5 times. It failed like the first 4 times for like lots of reasons. And then I go up and I get through, get through, the heaters are on and I get through like 60,000 feet and I'm starting to get a little warm and So I turn off the heaters and they stayed off the entire rest of the trip. Never, you know, we never used it again. It's like, well, why? What happened? Why did it? And it turned out we just massively overdesigned the system because when you get up high, the temperature is measured by the movement of molecules, right? So the molecules are moving really, really slowly up there. But there are so few of them that they don't conduct any electricity, I mean, any heat. Heat. So those two balance each other off. So when I'm up high, I'm generating some amount of heat because I'm alive, and that is balanced by the fact that none of that can actually conduct away from me. And so we just totally overdesigned that. So, so even though we built all these great heating systems in it, I almost didn't use them.

Okay, okay, okay. And then, and when you got up to where you wanted to be, was everything automated? For the release point, or you're kind of like, oh, I want to—

It's funny you mention it because on the very first time, the 50,000 or 60,000-foot jump, we went up, we had not even talked about the actual release. Like, there's 3 ways to release. There's, well, there's 2 different ways to release. Well, 3 different ways to release. 2 of them are from the ground and 1 is a big handle that I can pull. And so, But we didn't talk about it. And we knew that the ground was going to do the release because that allows me to get in the perfect position before the release. So we had already set that up. But I'm up there and, you know, I'm talking to Sebastian, who basically is the mission control flight director for this. And he's saying, you know, are you ready? And I said, could I at least get a countdown? Because we hadn't even talked about it. I'm up at 60,000 feet. Asking for, you know, could I get a countdown? He goes, "Oh yeah, sure." And so anyway, I think we counted down from 5 or something like that and then we did the release. But, but yeah, it was one of those items that we just kind of forgot about the actual countdown. And then in the final jump, it was interesting because we had a downward facing camera from the balloon which would show me the entire time and, and the process for actually working the mic. Is— was— is hard. Spacesuits are like horrible, you know, they, they've got all these bearings. And so in order to just like grab this glass or something like that, you have to go this way and this way and get used to it. But the process for using the push-to-talk switch was, was painful. And so we decided never to use the push-to-talk switch. And so every time they would say— they would give me some information, I would move my leg. And the downward-facing camera would see it, and that was the acknowledgement. But then the last thing is pilot ready for release. That has to be oral.

Okay.

And so I moved it, moved it up there, and I pushed the push-to-talk switch, and it goes beep, which is a sign that didn't make a connection. It's like, shoot. And so anyway, so I go and do it again, beep, and then So anyway, on the fourth time it actually did make the beep and I said, "Pilot ready for release." And then they did the countdown. So anyway, well, yeah, yeah. Well, then nothing that you think is absolutely totally simple turns out to be perfectly simple.

Was there a time limit on how long you could hang out up there once you got there?

Yeah, yeah. We had enough extra oxygen to last for like 2 hours up there.

So you didn't want to just kind of take it all in?

You know, no. Well, first, it's not super comfortable, like up there. Like, the suit is— I mean, it was in a position that's actually— you could fall asleep in, you know. I almost fell asleep many times. But, uh, but the, um, uh, so it's as comfortable as you can get. Uh, but for instance, there's no water. A long story about why we have no water. We, we tried a system with one of the backpacks kind of thing with the water that comes in and you could suck But when I fell over backward on the, on a, we try to fall where we fall over backward on a mattress just to see what would happen. That valve that holds that basically just exploded. And I got this giant, like, like squirt of water in the back of my throat and I thought I was going to drown. And so it's like, no, forget that. So you don't have any, you don't have any, any water for the whole thing. The view is beautiful. But you've been watching it, you know, change for 2 hours and 7 minutes. Also, we were starting to drift over some places that were more congested. Like, the area we were over was just perfect. I mean, it literally— I landed, I don't know, 300 yards from a road. We have all these helicopters and chase helicopters and all this kind of stuff, and the truck was like the first thing to get to me. Yeah, it was like, it's like like it was perfect place. And so yeah, we just decided to end it there. And I don't regret it. I mean, an extra, you know, given what I cared about, I, I think getting close to our auction margins, you know, like, nah, not worth it.

And then, and then the fall takes about 15 minutes to fall.

Well, the fall part is super fast.

Okay.

Uh, 4 minutes 27 seconds. Okay. So it's super fast. Um, because you're going Mach 1.22. On the ground, they heard a sonic boom, which is crazy. Super cool.

Yeah.

I wish they had taped it, but, and I talked to Art Thompson at Red Bull, and they also heard a sonic boom. Most people didn't think it was possible for a human being to create a sonic boom that was where you could hear it on the ground, but both their project and my project heard it, so.

So you're like a human missile.

Yeah, it's like, it's cool. You know, it's really cool that you could do that. And so, But yeah, it doesn't take that much time. And most of the time on the freefall, I'm just concentrating, you know, because Felix had heading control problems. I mean, he flipped over and did it. It was an inverted flat spin, which is like a scary thing. But for— because he's a superhuman skydiver, way better than me, he managed to invert the thing, come down, you know, land safely. He— my friend's landing— I mean, we call our— it wasn't really a landing. It was a crash, you know, like I'm in a 305-pound thing that doesn't bend except, you know, in the knees. Everything else is straight. And so I'm not going to land. I'm going to crash.

I saw you kind of flexed your knees right before you hit the ground.

It was like, you're going to just go.

I mean, we—

before we had this elegant way I was supposed to turn around, but it didn't work. And what we learned is I'm in an airbag, a giant airbag. I've got metal on both sides of me. Me. I am the most protected that you've ever seen. I've got giant boots on. I got this helmet that you could hit with a baseball bat all day and wouldn't think— like, we didn't need to do any of that. You could just— I mean, we did do as good a job as we can of flaring and stuff like that, but it was going to be a crash. And so I tell people when they see the video, it's like, here comes the landing, or more appropriately called the crash. And in that one scene that you saw where I'm laying on the ground like the turtle flipped over. And then the coolest part for me is I've got this giant piece of vegetation over here on my left side, which I thought, oh yeah, that's perfect. Well, the funny part about the whole thing is though that, you know, we hit a rock, we broke off this little piece of things. And when the suit is in the Smithsonian, the Air and Space Museum out at Dulles, and it's great, you should come visit. People visit it and then they send me like selfies of them with the spacesuit. But we asked him, do you want us to fix this little hole? And he goes, no, that's exactly what we want. We want this exactly as it lived.

Yeah. How many— I mean, how many times have you had to tell this? Do you get tired of telling it? It must be—

I don't tell it too often right now. I mean, it was a pretty long time ago.

I mean, I know, but I just imagine you go to a dinner party and somebody's like, well, what do you do? Oh, you know, that's the guy that jumped.

That would never come out from me.

Me.

My wife will occasionally tell people about it, which, you know, in some ways it wasn't her favorite event of life in some ways, but she's got it on her phone. She's got the video on the phone, you know, so she'll show it to people. But, but no, I mean, it's—

this is just how you're wired. So, and you won't— you don't— I, it's clear you're not like a flashy, braggadocious kind of person, but you wouldn't even say now how much it cost because you just—

it's none of anybody's business or You know, I, yeah, I just, you know, it costs what it costs. You know, it's like I was fortunate, like super fortunate that I got a chance to work in a company that was successful. And so, and so the money, I had the money part. I mean, you asked the question a while back, like, like at what moment did you decide to do it? And I remember I was on a flight to Europe or something and visiting one of the engineering offices there. And there was an in-flight magazine and it had a story of Red Bull and what it is. And this is after I had gone and done, you know, like done all the homework. And I was just staring at it and say, you know, like, do I want to live the rest of my life saying I had this idea but never doing it. Like, do I just want to put it in a drawer, which would have been the fiscally responsible thing to do, is slide it in a drawer, declare victory. But I don't know, some miswiring of my brain or something said, no, I want to do it. I mean, I don't want to just let it go. I don't want to let a good idea go. I want to do something about it. And that took me on this path. That turned out to be more expensive than I thought it was going to be, turned out to be harder than anything. But the value to me in terms of the team and what we accomplished and stuff like that was— it was worth every bit. Yeah.

No, I think I only harp on this because I'm such an aerospace junkie and I always love to compare the time cost of the initial plan versus—

I know—

the end, because it's usually like 10 to 100x.

Actually, the very weird— this is the weirdest part The one of the weird parts, my wife guessed the total, like, way ahead of time. Look, from the very beginning, she goes, that's going to cost X. And I said, no, no, there's no way it's going to cost X. We got this whole team. We're doing this all this kind of thing. And then in the end, she was right, which occasionally she reminds me of. Yeah.

And so, and you're saying, okay, because I went to visit Taber and Jane to see one of these Worldview launches where they did a balloon and the idea was to do I mean, they kind of had two parts of their business. One was to fly up cameras and sensors and things like that. And then the other goal was eventually to do tourism where you could hang out at the edge of space and not have to go up in a rocket. You could go up nice and smooth and have a cocktail up there and maybe even a bathroom and, and, and, you know, look down on the Earth. But so you're saying this, your mission is really what gave birth to that because they had to build these balloons for you.

Well, no. So, so what happened was is that neither of us knew anything about balloons. We didn't know anything about balloons. And so we ended up looking through all the people that knew things about balloons. And we went to Danny Ball, who ran the NASA program. And, and we also met a guy named Julian Knott. And Julian was from Santa Barbara. He had been approached on these kind of projects for a long time, but always said no. But I spent some time with him and, and, and so we started to learn about balloons and getting a balloon was actually one of the hardest things. It wasn't that balloons weren't available, like, like Raven Aerostar was a maker of balloons and been making scientific balloons for a long time. But they got out of manned balloons actually because of the death of Nick Piantanata, because he had used one of their balloons. And then later they got out of hot air ballooning and all this kind of stuff for the same reason. And so we didn't have a balloon supplier. Red Bull got around this problem by actually buying balloons that Raven Aerostar had sold to somebody else and then buying it for them. And, but we didn't have that option. And so Julian introduced us to this group, TIFR, Tata Institute of Flight Research. Of something, something research. And they, they built balloons in India.

Okay.

And so, and Julian was really good friends with this guy Shree there. And Shree ended up joining us on the project and helping in lots of different ways. But he, but that's how we got a balloon. But before that, we actually tried, you know, can we use weather balloons and tie, you know, like 50 of them together and all, which was a horrible idea, by the way, even though it had been done historically by the SkyDelft. But, but anyway, we ended up getting the balloon from them. But the balloons they needed for WorldView, for the space tourism, because they were trying to lift out like a 10,000-pound capsule, there were no balloons that were available that would be able to do that. So they had to take on the challenge of making balloons. But that happened much later.

But they got into all this as a result of this project.

As a result of that project. That's where we both learned about balloons.

Okay. And then even— and then they started a new company after, which is doing Space tourism.

Yeah, Space Perspectives.

But we still haven't— we still haven't got to this point where other people get to do this. Do you think we, we will?

I think it should work. I mean, there's nothing technically that makes it so that it's not possible. And I would have loved it because I, I really do think the idea of— I mean, balloons are, uh, the most amazing form of travel, gas balloons in particular. They are Zeppelin, and they are beautiful. And it's, uh, there's a few places you can fly gas balloons occasionally in the United States. In Europe, it's more, much more pervasive. They have hydrogen balloons they fly in. I still think of all the types of flying hydrogen, flying a gas balloon is one of the most special.

I went in a Zeppelin in Germany, and it really is such a different experience to a plane. I mean, like you said, it's very quiet and you can like open the window and stick your head out and like the ascent is this is very gentle and I hadn't really— you've flown all these different crafts. I haven't. I hadn't really appreciated how it's a completely different experience. And there's something just awesome.

It's really magical. Like when I need to get a gas balloon license, it's not really called a gas balloon license. It's a, it's a lighter than air license without an airborne heater restriction, which is what hot air balloons are. But I went to Germany to do that. This guy, Willy Eimers, who's a famous balloonist, has a place where you can go and you launch the way that you used to launch. You launch by, you know, sandbags basically that you drop out sand and you go up, you know, to 1,000 feet and then you— it's kind of a long story, but you get the balloon into this, this this position where just the tiny— like, you take a little trowel and you take a little bit of sand and you throw it out, and that balloon climbs. You are such— in such amazing equilibrium. And the balloons fly pressure altitudes. And so when you have, like, small hills and things like that, the balloon will automatically fly them all for you, and you can hear every single sound on the ground. Count. Like, we were flying overnight and there was a party going on, and we were flying up— they didn't probably even know we were flying over at 1,000 feet, and they were having this conversation in the backyard. We could hear every word. It was like, it's just so special. And, uh, yeah, so, so I think there will be a time, uh, you know, Space Perspective kind of ran out of money at an inconvenient time, um, which was too bad, um, because their their vehicle was actually coming along really, really well at that time. But, uh, but I, I still think they're, you know, that dream's still alive.

Yeah.

Uh, and there's nothing technically that prevents it from happening. Um, but, you know, like anything else, it's more expensive and you need to find the right investors for the right period of time and tolerate it. Yeah. But I would have loved it. Yeah. Yeah, it would have saved me a lot of money.

It did.

Yes.

Did Do you, again, I'm sorry, I'm sure you've had to answer this question, but I mean, did this experience change you at all?

You know, they talk about the overview effect and looking at the Earth and humanity in a different way. For me, that wasn't what, you know, I think it changed me in the sense that then And it didn't as much change me as kind of reaffirm the core belief that great engineering matters and that teams of people can accomplish amazing things. And that being a part of something like that is like really worthy. Like, you know, people will tell me, well, 'Well, you know, why'd you have to do that? Why didn't you just buy a ticket on Virgin Galactic or something?' And, and it's like, you don't get it. You just don't get it. It's like, it's like that, the, the, the process and the work at which— and the learning. Like, like, why do you do things? It's like, because I learned so much. I mean, when do you get a chance to be in the same room interacting on on a very important topic with flight surgeons and balloon people and environmental engineers and, you know, like, uh, spacesuit designers. And it's like immersing yourself in that environment with those people that you would never come in contact in a normal world, and then sharing something that's so personal that if I saw any of them today, it would be like time is like not going by.

Yeah.

It's like, how do you describe that feeling?

Yeah, I think it's, it's, uh, I mean, my only comparable-ish experience is when I work on these books for many years or a movie. By the time it comes out, it's, it really is the process that's like the fun part. And for me, getting to inhabit these other people's lives to get this knowledge from them. And then all the hard work that goes into it. And then on the movies, like a team that you're working with and, you know, by the time they end product, I mean, it's like you want to see it and it's great to have it out.

But I do think there's a moment like when we did the movie. So we put together a movie. This is Jane and Tabor's idea, which is, you know, let's document it and let's have people videoing things so we can remember what it was like because And so we ended up hiring Jerry Colbert and Tex, and they came and they actually, on a very small budget, like there are two of them and one film guy, you know, kind of thing. They put together the whole thing and they made a movie. And, you know, we debuted at Tribeca, which is like a great venue. And I will say the moment, the really interesting thing was first time somebody who has no knowledge of the project sees what happened, and you get to, like, see the reaction and feel the reaction and feel the audience and things. And I mean, Jerry made the thing way more melodramatic than it really was. But, but, but still, it was special because, you know, that story wouldn't have been told. Yeah. And that story would have gone away. And now, even at a 10-year anniversary, one of the things we do I did was go back and watch the movie. Yeah, it was like, wow.

Yeah, no, no, I can, I can imagine. I mean, it's extraordinary. I mean, you're, you're nuts.

It was fun.

Did Larry and Sergey go to the, the actual launch?

No, no, they hated the idea, by the way, you know, because they didn't— they cared about me. Like, they didn't want anything to happen to me.

Yeah.

And, uh, uh, but they were happy for me to see Sergey's airship. I did. I actually got a ride in the airship.

You did?

Yeah, yeah, like one specialist of the special. Yeah, I am so grateful. But yeah, I follow the project. I've helped them in little ways for a long period of time. I think it's a brilliant, brilliant— I mean, it's like, it's like I feel like that technology, you know, has languished for so long and, and that way of travel is so special. And I'm just so grateful that Sergei and that team brought back, you know, and extended what that technology is capable of in that way. And I must say, I had goose pimples like the entire time.

Oh yeah.

So cool.

So I followed it for years. Al Weston's a dear friend of mine. And so he, you know, he was the original—

Oh no, I know him well.

Yeah. And what a character.

Oh, I like him a lot.

It was extraordinary to watch them build that thing. And yeah, amazing. Will you indulge me a little bit more just to ask you a couple questions away from this?

Yeah, it's okay.

Okay. Okay. Just because you're, again, you did so much stuff I'm fascinated by. I, when I, okay, when I first started out as a reporter and I mean, it was amazing, you know, I was 22 and knew nothing about tech and I had to learn all this stuff on the fly and I got thrown into covering Solaris and Tru64 and Alpha and Spark and all these things. And it was, it was such a cool time because I used to go to everybody's research labs and, and all these companies were working on the, the every, you know, they were making chips, operating system, file systems, storage system, networking, everything. And everyone was a computer science lab. And it was exhilarating. And I loved the, this, the competition of it all that still goes on today, but it is definitely different. I mean, I, I, I don't really get invited to people's labs hardly ever. I mean, Google might be one of the rare exceptions that still has something close to this. Yeah, I was just kind of, I was curious. I mean, have you, does this ring bells? Do you feel like the Valley's, I mean, maybe, you know, probably didn't make sense for everybody to be spending billions of dollars making their our own chips at some point, obviously at some point. But yeah, I mean, does the Valley have like a little bit of a different feel to you in that regard or all that stuff matured and we're just working on different problems now?

Well, I mean, I mean, I mean, it was a really interesting time. You know, I feel like I grew up in like, you know, the best possible time, you know, like when I was going to college, you know, the VLSI revolution had just started and the idea that anybody could build their own VLSI chip and there were rules to be able to do it and then you could have it fabbed and you get it back and you can run it. And that was when the MIPS chip and the RISC chip were coming on and before it was something that was only relegated to very large companies and then that spurred a whole round of innovation in architectural ideas and software ideas and hardware ideas. And then, you know, Xerox PARC comes along and, you know, they build the first bitmap display and, you know, the first, you know, printers that can, you know, and, you know, use of a mouse. And I mean, they, you know, I mean, there was—

Ethernet, I think Ethernet.

Yeah, Ethernet. Yeah, I know David Boggs I work with at Whirl. He was one of the co-inventors with Robert Metcalfe.

Stuff.

But yeah, it was just like a— and, you know, ARPANET was starting, and, uh, and NSF had a network, and, you know, I mean, it was just, it was just like before you were isolated and now you're connected to this larger world. I mean, it was a, it was an amazing time. And, uh, and every— it was a, it was a new time where everybody felt like they could now do anything. You know, there's a book by, uh, I think it's Tracy Kidder, Soul of the New Machine. I, I might get that wrong, But it's a book about a small team building a processor or building a computer. And it was awesome.

Like I have that book sitting right over there.

I, you know, it just still gives me goosebumps right here. You know, it's like, it's like crazy. Like this idea that you could do things far and above what you were able to do before and expanded the horizon of ideas that could be realized. And so I loved it. But, you know, I look now and, you know, there's another set of places where our toolset and our ability has now allowed us to expand the horizon of what we are capable of doing. And that has like profound implications on like the way we work and other things. And I feel like, you know, the Valley is different. In some sense. Um, but in some ways, this idea that we are, um, that we have a new set of tools and a new set of ideas, and we are now capable of expanding the horizons, and if we do, it will have dramatic impacts on us and society and things like that. And the key for me, I mean And, you know, this is a longer, more philosophical discussion, but fundamentally we are about to build or encounter a smarter society. Like the IQ of society is about to— I mean, it's been going up, no doubt about it, but we're about to make a step function change in our intellectual capacity. And how we choose to use that is really the fundamental thing. And if we work on the same set of narrow little problems, then, you know, all hell will break loose, all the jobs will go away, all this kind of stuff. But if we as a society expand the frontier of what we're going to attempt, like we're going to cure cancer, we're going to solve transportation problems, we're We're going to do farming a lot better. We're going to be faster and more efficient. We're going to generate like way more interesting companies. We're going to, you know, we have the potential to take an innovation just like VLSI did and turn into something which is, which everybody will in the end assume it was kind of always that way. And I kind of like think in the back of my mind, it's like, Like, do we want to go back? Like, you know, after the tractor was invented, you know, it's like how many people 20 years after the tractor said, you know, I really thought we were better off without that tractor thing. Yeah. You know, I really thought pickaxe was the right way to go. And now history has proven me right. You know, why? Well, why? It's because society moved on in a way that we're, where the inventions caused us to be able to like feed more people, to be able to lower prices, to be able to farm larger tracts of land, to be able to farm land that wasn't practical to farm. Like when the steam engine was invented, trains were invented, and then we expanded the horizon. Like now we're gonna build trains that go all the way across the country. You know, like, can you, like, across the Rockies and across the, you know, the Sierra Nevada, and we're going to actually go on both ends. I mean, it's like we expanded the vision of what we could do as a country, and that expansion more than made up for the fact that the technology had both allowed us, enabled us Uh, but it made us much better. And so, so I don't know, for me it's like almost impossible to believe that 25 years we'll go back and say, man, I wish we were a dumber society. I mean, I think we really made a mistake by getting so smart.

Well, okay. And let me ask you one more question along these fronts. So you, you live through this explosion of computing power. These chips come along, you help even push this forward. You're all the stuff you're talking about, VLSI, you're witnessing the birth of the— not the birth, but the fulfillment of the microprocessor revolution. You watch then the internet go from these early days to this thing that takes over the world and also unlocks human potential in new ways. We still have debates whether LLMs are sufficient to, to get to this, I, whatever anyone wants to define as a superintelligence, but clearly they're doing incredible things already. So based on everything you're saying, I'm just getting the sense you're very much by, you've seen enough of what we've done with AI in this moment in time to think we are for sure on the cusp of this another revolution that is either comparable or bigger to what you've seen before.

Yeah, I would definitely say that this is revolutionary technology that can dramatically improve. I mean, here's another way to say it, is that we have pushed our brains, our kind of boring little brains, pass what I think anyone would expect us to be able to do. Like if you go back 300 years and you say your brains are capable of doing this, people would say no way. And we've done it because, because we've abstracted problems away. We've worked really hard. We've put a lot of things in place. We've built on this, you know, stood on the shoulders of giants. We've moved the things forward in fits and starts. We, we've designed machines that can now give us more information, which can change our ideas. And we— this flywheel has been able to allow us to proceed. We are so limited by our own brain and the amount of complexity that we can understand, and then we're super limited by our inability to communicate. Like, this conversation is the lowest bandwidth communication in the world, and no disrespect, but, you know, it's like, you know, it's, you know, it's a bit rate thing. The bit rate thing is super low, you know, and And so we, and therefore teams like can't get, they can get large, but the inefficiencies associated with are so high. And I'm proud of what we've managed to do with our brains, but the idea that we can remove some of those limitations that we have and be able to take on harder and harder problems because we are no longer limited. And also just the abstractions that we need to do to be able to like, you know, get as far as we have, you know, like in my world and maybe in your world, object-oriented programming is a good example. Like, you know, maybe we can describe, you know, objects and the interaction between objects and therefore we create a level of abstraction between them. We do that all the time, even on the Fusion case where I'm on the board of Commonwealth Fusion Systems. Systems, trying to build, you know, fusion power because power is one of the most important things in the world. And we have to basically figure out a better way to provide firm power all the time. And fusion is like a lovely technology and it's an amazing team and they're doing great things and they're going to prove it. They're absolutely going to prove it. It's not going to be decades away. It's going to be in the next couple of years. We're going to have power plant scale fusion. And I'm so lucky to work with that team. But, you know, solving, you know, kind of that level of problem, even in that level, there's abstraction that they use to basically, you know, this building holds cryo, this building holds RF, this building holds, you know, like energy density, this one has tokamak. So in every way, we have to build abstractions around things to make them work. To make them worthwhile. And it's just, and that shows our limitations. But once we get past those limitations, once we can think about much more complex things and complex solutions, 'cause you know what comes down to the human body is pretty darn complex. Like why haven't we solved the problem of the human body? Why haven't we solved the problem of cancer, all these diseases? Like, you know, it's because the human body is complex. I mean, even the atom, like, you know, like Bohr and the atom, you think, oh well, we're done. We're not done, you know, uh, the cosmos, like, like we have all these bigger and bigger telescopes that finally give us way more information, and we're, we're, we're asking more questions. Every time we get a new tool, we ask— we get to ask and sometimes answer new sets of questions, um, which expands the frontier of what we're trying to do. And I just feel like, boy, am I glad to have a power tool. When I was working with sticks and rocks and stuff. Thank you.

Has the— has the rate of recent progress felt arresting to you? Does this feel different? When we had Moore's Law, we were on this roughly 18-month cadence. We could see computers doing more and more things. When the consumer internet came along, you had this explosion of interest and and hope about it. Fiber optic cables were being laid everywhere. All these data centers were getting built. Then you had this interim period where the infrastructure outstripped the demand, and then eventually the demand exceeded people's wildest expectations. But this took, this took 10, 12 years, which is a relatively short amount of time.

Yeah, people were right about the impact, but they were wrong about the time.

And so over these last, say, since like the transformer came out in say like 2018 or I don't know, pick a, or ChatGPT came out. Do you find the rate of progress over the last 2 or 3 years familiar or is this, is it arresting and exhilarating and frightening?

I think it is all of those things. I mean, I feel like, I mean, I am surprised. Well, first I'm surprised that language was the thing that came first. You know, it came right down. I thought language would be the last thing. You know, I thought, you know, I don't know, you know, you know, I thought there were much easier problems out there. And I thought language is very complex. And, but what was interesting, and so the whole LLM thing came as a complete surprise, not because the technology didn't exist, but because I just thought language was so complex. And, you know, if I'm smart, I should have like figured it out earlier. 'Cause in the early days of Google, I remember we had a research project around machine translation, like language-to-language translation. There's a guy named Franz Och at a company called ISI. They did this BLEU score test where they could, where they'd get a set of translations to make on like Monday, and then by Friday they had to basically turn in their translations that was all done by machine. And then, but Franz Och and his little computer basically had the best translation in the world using a set of techniques that kind of were related to machine learning, more probability-based. But so anyway, he was the expert. And so my team said, what can I, I mean, I asked my team as I always do, is there anything I can do to make you go faster? And he goes, yeah, you need to hire Franz Och. And I was like, so I try to hire him and he says no. And he says no. And he says no. And eventually I I flew down to LA and met him at the old, you know, the restaurant right by the airport. And that's where I met him. And in the first time in Google's history, to my knowledge, maybe it happened before, but I actually made him an offer without an interview, like never done before. And so I made him an offer and I said, look, we really want you. And he came, he eventually said yes, thankfully. And he came up and was building a system. And then I I connect him with Jeff Dean, who's, you know, like world famous, you know, one of the, one of, if not the best computer scientist at both Google and the world. And they together took a process that took 5 days to do a print translation and turned it into seconds. And then we shipped it and all of a sudden now you were not limited by a single language in what you were able to access. And so this idea that language is more predictable, and that the— I mean, in the end, the whole key is that you're, you're able to predict what's coming next.

Yeah.

And that's the key. But, but I still think LLMs are wonderful, but I think the other side of, you know, the whole AGI debate and trying to figure out like how systems work and how we learn and learning and reinforcement learning and being able to iterate things and stuff like that more like humans learn. It's like the difference between AlphaGo and AlphaGo, you know. It's like I feel like there's steps that still need to be taken to be able to get to the point where we close these loops and that there's some limitation right now on the amount of training data that's there. And training on generated data might not be as good as training on original data. And so, so, so, so, I don't know. I, I, I, I'm a little bit disappointed that so much effort is going into the LLM, LLM side and not on some of the other things. That I think could— like, I don't think cancer is going to be solved by an LLM, and I feel like that's a pretty big and important problem. So, you know, so for me, I'm, I'm thrilled. Um, I, I wish— I mean, if I had my druthers, I would push out some of the other frontiers, you know, probably as fast or faster than, than the LLM side. But anyway, that's just me.

That's— you would know far more about this than me. This is my I've got instinct just from having covered different waves of technology and then just where people are putting— it's like, I don't think we need $5 trillion companies trying to copy each other. It would be, it would be good if maybe, you know, a couple of these went off and threw all that into other spots. But there are many stocks.

Yeah, like materials. Like materials is like a super exciting area. We have the ability to build materials we've never seen before. I feel like like a great set of materials would allow us to do lots of interesting things.

I wish I could cite it. Uh, didn't Claude Shannon in the '60s write this paper about the limitations of language? Yeah. I mean, didn't he show that, um, we think we have all this variety and choice when we're picking our next word, but that in actual fact, it's much, it's much more limited. I mean, I guess whenever I have looked back at that paper, it just seems like, oh, I guess he, I mean, Claude Shannon, and sort of knew everything. So it just felt like he had put some architecture that, oh, maybe language could be solved if you had enough horsepower. Yeah. There's a lot of like early people that had really, really, I mean, you look back like Vannevar Bush had lots of really interesting, you know, uh, I mean, those two Bush and Shannon and, oh, somebody's escaping me at this moment, but, um, sort of like this, like Von Neumann, this, this window into the way the universe operates is see the future. Incredible.

It's incredible. And, you know, like I was talking to somebody the other day, like, what is it about, what is the biological mechanism of insight? Like, what is it that all of a sudden causes somebody to figure out something that, you know, for entire human history we've never figured out? And then to be able to predict it, you know, decades before, you know, it actually can. And some of the things are just, just hard, you know, and they take a lot of work. Like the basics for fusion. I mean, there was a— the Lawson criteria was basically invented in like 1958 or something like that that basically told you exactly what conditions you need to make fusion, you know, power. And but, you know, it's taken a pretty long time to basically be able to recreate those conditions, even though we know exactly what they are.

So in its heyday, DEC was almost like the OpenAI or the Google of its time. I mean, the super sexy—

I think they peaked the moment I joined. I don't take complete responsibility for the demise of DEC, but yeah, it was, it was, you know, you know, and it just, I think it's an interesting interesting story in technology and companies. Like they built the VAX and the PDP-11, and the VAX was by far the best minicomputer of its time. I mean, they sold them like hotcakes. I mean, and they actually had their own operating system, VMS, that went into it. But, you know, there was this interesting moment where people would buy their machines, but they would strip out every single piece of the software on that machine. Machine, and they would load on it Unix onto that machine. And that probably should have been a wake-up call for them. And in my, I went to University of Central Florida, it was called Florida Technological University at the time, but the Unix people could only use the machine at night. Like the machine was used during the day by the VMS people. And then every night at like 9 PM, we would go in, there was these giant hard drives and you'd spin down the hard drive and you'd pull out the platter and you'd put it over to the side and you'd put the new platter in. It was Unix and it would run Unix until 9:00 AM the next morning. And so all of us would work basically all night 'cause that's the only time we had access to the operating system on which all of the interesting innovations in VLSI were being built. And so, but yeah, they were great. I mean, some wonderful company, but they also, You know, their money was in VMS and not in TCP/IP, and they had DECnet at the time, and they kind of missed where that technology was going. And then they were very early in building your own chips, like, you know, like MicroVAX, Alpha, all this kind of stuff. They even built some of the very first low-power ARM chips were done there on the StrongARM processor.

That's right, I forgot about that.

Yeah. Which is sold to Intel. Intel actually followed up on low-power processors, it would've changed everything. But they killed that project basically because, you know, they wanted one architecture and one CPU and stuff like that. And if you look at like many of their problems, it dates back to whatever that decision, I don't know who made that decision, but at DEC, the same thing. It's like inside of DEC, I remember a set of meetings that we were in and, you know, should you choose the, the big mainframes or the mini computers or the micro computers. And there was a lot of internal debate. And I actually, I was actually in a high-level kind of executive review of the three. They all used roughly the same technology. They all cost about the same amount of money. But on this one, on the microprocessors, we made almost nothing. Mini computers, we made something. And on the big computers, we made a lot. And so we decided to chase IBM. Instead of chasing like the mini computer people. And that was a mistake. That was the moment. So it's like, and maybe that wouldn't have been a great business, you know, maybe we needed a different company to be able to pursue low margin businesses like that. But it was sad because, and then Compaq ends up acquiring DEC, which is like, you know, the upstart, you know, they didn't even build their own chips and stuff. And they acquired us not for the technology, they acquired us for like our professional services organization, which was even more like a dagger to the heart.

Get into the data center.

Oh man.

Well, so the—

The people were great though.

Well, that's what I was, I mean, the talent that comes out of there, I mean, Jeff Dean, who I mentioned.

Oh yeah, the— Chuck Seitz. I mean, Microsoft, you know, their whole operating system.

Gordon Bell.

Yeah, I mean, it's just like the number of like amazing people, you know, Jim Keller, Pete Bannon, you know, I mean, it's like a who's who list of all these people that were incredible talent and they went on to do amazing things.

Google, Microsoft, Intel, building all these new architectures computers. And then Sun is in many ways the inheritor of this. They're the next DEC in, in a lot of— you could argue, you know, the, the next hot new thing. And they've, they've taken the— we've gone mainframe, we've got minicomputer, now we're at a workstation. And then I was, I was just so curious. I mean, right as you're joining Google, Sun is about to be hit by many of the forces that you just described, which is Sun's making its own version of Unix, its own chips, Now this Finnish kid creates a Linux kernel and, and you can put it on Intel's chips and it turns out it's quite powerful and does a lot. And then Google, more than any company on earth, recognizes the power of this to go build data centers that are much cheaper, these enormous clusters instead of having these, these high-priced Unix machines. Was there any part of you in your heart as you're like doing some of the work to make this happen that you felt some sadness for the era of computing you were about to put out of business?

No, no. I mean, the thing about Google, which was special, is, you know, we had a lot of super smart people. Like, they understood, like, a lot about all the pieces, how the pieces went together. But the thing that the previous era had created a set of tools that, that provided the framework for what we were trying to do. And, um, and, and search itself was, was interestingly partitionable, like these, um, the indexes with basically, you know, arrays of words basically, or arrays of references to words in all these different documents That was actually pretty easy to parallelize, not simple to parallelize, but it was easier. And so all the innovations of the past, like Linux and fast networking, other kinds of stuff, those were all the pieces that needed to be put together, like disk drives and stuff like that and large memory systems and other things. But there was a, but what happened was And so what Google did was start to do a vertical integration because, and they didn't have to invent everything, all the pieces, like the puzzle pieces were on the table and all they had to do was fit it together correctly. And then they would expose the holes in the system, like networking was a huge hole. You could have Ethernet, which is great, and it got you up to some point, but we needed a lot more networking than that. That. And all the things with network, lots of computers together at high bandwidth, those were prohibitively expensive. So we made the investments in people to be able to take and build commodity networks at much higher performance or to spur that effort with different people. And so, so I feel like we took the best of all the worlds and put it together in a way that we could control the whole ecosystem, and that allowed us to move problems around. Like, maybe you couldn't solve this problem here, but maybe there was a software solution, or maybe there was an operating solution, or maybe there was a, you know, some hardware solution that would get you there. And so by controlling a whole stack, then you got to do things that you couldn't do before. And also by parallelism both at the rack level, at the data center level and other things, you didn't have to worry about some of the problems that you, like saw, like, oh, Tandem Computers basically, they were building triply redundant computers, 3 computers that would basically execute in parallel and then you compare results. You know, that was a very expensive solution that in our case we just had, you know, things and racks, and you could run the same thing on a different rack, and you could compare the two at the end and not, not have to build any special hardware but still get the same advantages of that. And then, you know, if a, if a, if a system went down, you didn't care. You didn't have to build a big UPS thing out in the back that would keep all these computers running. I mean, one rack would die, the other one would take over. One board would die, the rest of the rack would take you know, and, uh, and, and in the end, if one data center died, you know, another data center would take over. So, so building in simple commodity, uh, systems that can, you know, kind of punch above their weight and solve bigger problems, that was super interesting to me. And I feel like, like, uh like it was the next generation. And, you know, I don't know why the other— I, I think it's an overwhelming, uh, an overreaching question. It was like, why don't big companies— like, they have all the resources, they see everything that's going on, they have very talented people, but somehow the inertia keeps them from reinventing themselves. And I don't, you know, I don't know why.

There's books written about this, but still nobody ever— Yeah, exactly.

We still seem to be repeating the same problems. You know, it's like it doesn't go away.

The stuff you're talking about with Intel, I remember going year after year after year to their labs and they would show me a smartphone and they, like, way before the iPhone came out, Intel. It was essentially— it was like a shitty iPhone, but it was an iPhone. And they, they knew what was coming. They did the same thing with Microsoft, and it was just— could not swing the organization to, to make that kind of like priority one for, for what would be next. And then you end up—

I miss mobile, and then you miss, you know, AI. Yeah, I mean, it's like those two revolutions, it's like, how did Intel miss both of those two?

Would ever have imagined that Nvidia would be worth this much more than Intel? Intel was, was like, for— there was a long time where even when Intel made a mistake, they just turned the fab up and obliterate what had come. It's the thought that like Arm and Nvidia would be worth more than Intel is like unfathomable.

Yeah, yeah. No, it's just, it's, it's— it's hard to see, you know, it's hard to kind of understand. I mean, even at the time there was Intel and there was AMD, right? And they were both making chips and systems and stuff. And I, you know, we obviously knew people and, you know, we, for our data centers, we would build an AMD system and we would build an Intel system and we would compare them. And even in the cases where AMD wasn't good, we would give them a lot of feedback. Like, here's where you're failing, this is where it's not working, you know, well, 'cause we always wanted, you know, like two systems. We didn't want to basically have one system that would always basically go with. And then sometime, and we made a promise to AMD that says, look, if you do this and you are the best solution, we will buy you. And we did, we worked back and forth over a long period of time and supporting an ecosystem that was, friendly to innovation was like a super important, you know, part of what made us successful is preserving optionality over time. And so you look at, you know, NVIDIA, I mean, you know, they didn't know that their GPUs were gonna be good for AI. They had no clue, you know, these kids up in Toronto or Montreal, AlexNet, you know, Alex and, you know, as part of the acquisition that we did of which Jeff Hinton and Ilya and Alex brought in. But that was the first time I heard, it's like, hey, we figured out how to use GPUs and then we can make things, we can make it go a lot faster. And if I was smart at that moment, I would've gone and bought a chunk of any— I mean, I did okay with Google, but it's like, But even, you know, in that situation, you didn't realize that a bit had just flipped. I mean, I talk about this in Fusion a lot, which is that technology doesn't go from like impossible to maybe slightly impossible to okay, mostly impossible, but maybe there's a few things to neutral to, you know, so, you know, slight possibility. It's like that. It doesn't do that. It goes from impossible to inevitable. And it happens in a heartbeat.

Yeah.

It's like a very, very fast flip. And you don't, most people don't recognize it when it happens. And in that moment, you know, NVIDIA went from impossible to inevitable. Fusion, same thing. Right now, I don't think the world's seen it, but Fusion has gone from impossible to inevitable. And we're about to show that. So, but it's interesting, you know, how quickly the world changes and how quickly the world reacts to the changes that have been put in place.

In 2008, I wrote the first time anything outside of a trade publication. I was at the New York Times mentioning a GPGPU and having Jensen quoted. And it was mostly for scientific workloads at the time, but he does, you know, he starts talking about it. These chips will be useful for new types of computing and intelligence. And I'm not allowed to invest in tech stocks, but that would've been a very good time.

That would've been a good time for you.

I think NVIDIA was still basically from year to year going between near-death experience and, you know, based on whatever, however, what new gaming console had come out, how their fortunes were doing.

So yeah. Yeah, Bill Daly was one of the chief scientists, good friend of mine. And yeah, he's been along with the whole ride. And it's like, you know, that you talk about companies and it's like, you know, they were an overnight sensation, but they're not an overnight sensation. Yeah, it's a long period of time with really, really hard work by really smart people. And, you know, I always tell people, it's like you can't make a wave, but you can see a wave coming. And if you do everything right, you can catch a wave. Wave. And, uh, and I, I think that's, that's, uh, that's how technology works. And a wave's coming. Yeah, might already be here.

Yes, I think it is.

Some people are on it.

Well, I'm going to regret doing this because I'm going to point out to you that I've already gone past the amount of time that I, I told you about. Um, I will talk to you forever. Do you want to talk Fusion and what you're working on now, or do you want Or not.

Give you a quick— I can give you a quick—

because I've been to Helion.

Yeah.

And then I've talked to my— I've got friends at Commonwealth and they don't care for Helion. And then, and then I keep getting invites from Commonwealth to come see.

Yeah, you should definitely go see. 100% chance you should go see. It is, it is an incredibly impressive facility they put together. I mean, people don't realize that the scale at which fusion is happening right now And, and it's all based on a very simple thing. You know, Dennis White told me a story. He was walking down the hall. He runs into a, you know, another professor, shows him a material, very thin stainless. It's like a cassette tape kind of thing. But on that material, they've deposited the atoms to make a high temperature superconductor, and it's called a REBCO process. And so what was interesting about it is that material is a high-temperature superconductor that doesn't lose its high-temperature superconductivity in the presence of a strong magnetic field.

So, so this was kind of the breakthrough.

Dennis looks at that and said if you could use that material, then you could build stronger magnets, and the power out of a fusion reactor goes up by the third or fourth power of the magnetic field strength. So if you could build a magnet that's twice are as powerful, you could absolutely totally build fusion machines that generate more energy, or, and/or you can also shrink machines. And so the combination of being able to shrink the machine and still maintain high power output, that was the key. And so the first thing they had to do is prove they could make magnets of that scale, which is like, you know, almost as big as this room. So they did that in 2021, raised another round to be able to build build, you know, the first fusion machine. And that's— it's going like really, really well. I mean, really well. So you should definitely go visit. I strongly recommend it. It is the coolest tour that you'll ever take. Yeah. And you should definitely—

okay, well, I will. I will do it. I— when I was reporting on Helion and then just having some back and forth with Commonwealth and then reaching out to other friends, still seems like— I mean, then I talk to other very smart people and they're just like, ah, no, we've been promised this many times before. Show me one system that makes any, you know, an iota of more power than it consumes. And, but you're, you're— so you've seen enough?

Oh, it's— I mean, you know, it's kind of like a weird moment. So you're— so I'm there at the beginning, you know, like the beginning of the company, before there was a company, and, and I'm standing in a room and they're, they're cooling this magnet down and, uh, And like the magnets themselves, like the toroidal field magnet is the coolest magnet you've ever seen. It's like, it's like, it's called a non-insulating magnet, which is the weirdest thing. Like one of the things about these really strong magnetic fields is that the, is that they're really, I mean, the structural constraints on the magnet are really high. This thing can't bend. It's got to stay really strong. So they came up with this idea, which is the coolest idea. They take a sheet of steel, they basically like carve in these circles around it, and they put the superconductor in it, flow solder, do all that kind of stuff, stack a bunch of these things down, and then they turn this thing on and they put the smallest amount of current they can across it, which of course goes right across the steel, doesn't even bother to even touch the superconductors. But then they start cooling it down, and the moment it gets to super, to the temperature to the critical temperature, then the electrons find this new path. And so the entire magnet is made out of non-insulating material and generates, you know, like over 20 Tesla, which is like unbelievable amounts of power. And now they manufacture all the magnets, they're slotting it in, the vacuum vessel's there. It's like the coolest time. So go visit.

And that part's in Virginia, right?

I mean, the part that goes in— No, no, it's in Massachusetts. There's two machines. There's a Spark machine, which is basically, you know, it's like 13 of the, or 10 of the 13 systems you need to make a power plant. But all the ones that I think you wanna retire a lot of risk in, the other ones aren't that risky. And so that's gonna go into, that's the one where we're gonna show Q is greater than 1. Q, we're gonna show Q is greater than like 10 or something like that. But it's gonna start with Q is greater than 1. And so that's what you'll visit. You'll get to walk through the room that eventually will be filled by neutrons and would X-ray you at the moment you walk through there or something. But right now you can actually see and feel what the very first fusion plant is gonna be like.

And then you spend most of your days working on Aqua Satellite, right? That's your—

That's true, yeah.

So that's your—

I have a little company that's like 13 people. And it's based on this premise, energy, I think, is one of the most important problems you can work on. But the fact we know nothing about the oceans is just really irritating to me. And I went in 2017, I did a search with a friend of mine who also worked on the Stratospheric Project to try to find Amelia Earhart. And so I funded and participated in that project.

Oh, wow.

And we didn't find Amelia Earhart, which is the first thing that that group relatively small group had ever tried to find that they did find.

So just where she crashed.

Where the crash was and where the aircraft is. And so we did that. But I came back from that saying, wow, it's, you know, I've always heard that we, you know, we know more about the surface of the moon or Mars than we do about our own oceans. And in that 2017 trip, I could tell you why. It's because the cost of placing a sensor anywhere in the ocean is prohibitive. Expensive. The ships are like $50,000 to $200,000 a day. It takes 18 months to be able to even get a ship into the right place at the right time. Uh, uh, it's like these ocean-going vessels are just so expensive. And so after that thing, I came back and I said, you know, I'm not going to go do another one of these things unless I can cover 10 times more area at 1/10 the cost. So I need to be able to get a factor factor of 100 out of the cost of ocean exploration. So, so anyway, so I built a team and we've been at it. We have vehicles that we're basically putting in the water all the time.

And is it in Florida or here?

No, it's here.

And I mean, Liquid Robotics and Saildrone tried to address the very problem you're talking about by building—

Both of those were at the surface.

Okay.

You know, so I don't— I'm not as interested in the surface because I think all the cool stuff is probably underneath the water. For me anyway. And so the key is building vehicles first that can go deep. Right now we're limited to 500 meters, but I think we'll get to 1,000 and 6,000, maybe 11,000 after that. But you want a vehicle that can efficiently explore the ocean. We're using cameras as our main sensor., and we have NVIDIA chips to do machine learning. So we can actually, as we're traveling under the water, we can learn a lot about what we're actually seeing and then make navigation decisions. And based on that, how do you—

I mean, you want—

don't you need light? It's electricity. And the trick to get— like, the ship provides power. That's the main thing it does, to charge batteries, to put things in there. And most of the vehicles Vehicles like the vehicle we used in Amelia Earhart search was a REMUS 6000. It's like a $5 to $8 million vehicle and it has a team of 5 that comes with it.

So there, but it's a drone, it's a drone.

This is on, on my, the one, and that's the REMUS 6000 is a drone too. Okay, uh, autonomous undersea vehicle is what they call it, but ours is like this big, it's like this wide, weighs about 70 pounds, and And it's autonomous, but it's also super smart. And the trick is how, if you, to make the economics really change, you have to keep it out there for a long time, which means you've got to figure out a way to recharge it. And so our mechanism to recharge it is actually to unfurl a solar array. So you come up to the, ideally what you do is you come up to the surface, you charge for 8 hours, then you do science for 16. I see. And then you'd go back and forth. And so that's the goal that we have, is to build those systems. And we've had a vehicle in the water every week for the last 2 years.

Where do you go?

Uh, sometimes, like today, the team is down in Santa Cruz area. Okay. But sometimes you're off Half Moon Bay, sometimes you're in Redwood City. So doing tests in the bay because it's a little closer for us.

Should we come see?

Yeah, sure.

Yeah, I would love to. I mean, I don't know, you probably— I'm sure you don't know, we— I make a video show where I travel around the world filming interesting science.

Yeah, no, it's fun. I mean, it's hard though. The ocean is hard, but eventually, I mean, the reason the company is called Aqua Satellite is because I feel like there should be constellations of satellites continually exploring the ocean.

Yeah. So even though— I'm sorry. I'm sorry. Well, just like, because I've done a lot of reporting on Saildrones. So even though they're doing all these, aren't they doing like LiDAR and all kinds of sensors going down? Maybe they They don't do LiDAR, but I mean, it's not good enough to point sensors down from a craft that's on the top of the water.

So, so I love Saildrone. I mean, I think what they're doing is really cool. And I love Liquid Robotics. I'm, I'm a huge fan of anybody who's basically doing interesting things in the ocean and trying to build a business plan around it, which is like the hard part. But the, but the surface is not as interesting to me. And I think it's hard. Like the surface of the water is where all the wind and the waves and the sun and the breaking waves and all this kind of stuff. And it's like, it's like not a, it's the hardest part of the puzzle in some sense. And the least interesting to me. It's like if I had a, let's say I had a terrain map, like a contour map of this area. Like would I rather have a contour map or satellite map? You know, would I rather want to see what's actually there? Or would I rather just the contours? And getting contours are— and it's a big thing, like multibeam sonar is a lot of power to do it. They're building huge sail drones that can do that kind of stuff. And they're— to a lot of people, like people that are like caring about channels and stuff like that, you know, there's a great business case for them to be able to do that. And, you know, there's people that are doing like really good work to finally figure out what the actual terrain— because the 'Cause the bottom of the ocean right now, I think for most of what you see on Google Maps, for instance, on the ocean, it's one data point every square kilometer, like one depth average over a kilometer. So that's not very good. And I applaud anybody who's trying to solve that. But the explorer side of me says, that's not what I'm interested in. Like, I wanna find Amelia Earhart. I wanna find MH370. I wanna learn about all these creatures that, you know, the 40% of them that we have eDNA from, but we don't know what they are. I wanna understand the migration patterns of whales. I wanna understand what, you know, these things eat. I wanna see, you know, like that's the cool part to me. And so for me, it's like, I'm happy they're working on those things, but to me, and maybe this, I mean, this is a pretty darn hard problem, honestly, to deal with pressure and all that kind of stuff. But it's a hard problem too to deal with the surface. So it's just, like, and also not that many people are doing it. Like, it's like, uh, like there's not like 100 companies out there that are trying to do what we're trying to do. Um, and so I don't know, to me, I like, like I have a little bit of resources and, uh, I'd rather apply them to something which is really hard.

Yeah.

But is really impactful. Like even like we had this case the other day, but you know, you can recreate it. Like you're standing on the Golden Bridge, you're looking out and you ask, what's there? Like what's underneath the Golden Gate Bridge? Like, I don't know that anybody knows.

Yeah.

Like I'm sure there's a bunch of sunken ships out there. I'm sure there's probably really cool marine life. But the idea that we're standing there, like, I mean, how many people cross that bridge like every day? That. Look out on that thing and like, like, I don't know whether they care.

Yeah, but I care.

It's like, oh my God, we have no idea what's there. And maybe we— I mean, the best thing we have is a contour map that's done by the, you know, that shows the geography, you know, for sailors to make sure they don't run into rocks. That's all we know. But boy, it sure seems like, like we as a civilization are capable of learning about the last unexplored place on Earth. And let's, let's spend some effort like trying and maybe failing, but at least trying to solve that problem.

No, I love it. So you're going to do, you're going to do the world's most impressive skydive and then you're going to find Atlantis. That would be a good, it would be a good life well lived.

It would be, it would be a good comedy. They said, find Amelia, find the MH370. I mean, really, I mean, my goal, I mean, this is like a dream. It's not really a goal. You can't really call it a goal, but I would love to have the very first satellite map of the bottom of the ocean. Like every pixel. And then everything we wanna learn about the ocean becomes a data problem. And I think that would be like amazing. And we are so close. So far away from it. But it's, it's an interesting thing to be able to know the ocean the same way that we know land.

No, I love it. I love it. I love it. Was all this as painful as you thought it might be, or was this okay?

Oh, this whole series? Oh, it's fine. I don't mind. I mean, it's not, it's not that it was painful. It's not that I didn't want to talk to you.

You're busy.

It's just I'm so busy. And like every moment, like I am like, I have like every single moment of every single day is just so filled.

Yeah.

And like right now I've got a, you know, like in my backyard I've got a, uh, I'm trying to figure out what, how these solars are, where I react to waves and long duration stuff. And I'm building a test in the backyard. And you know, my, uh, my middle child, and it was actually helping me program it last night. And it's like, and then I get this thing and then I need to run these tests. And stuff like that. It's like, it's like, I, I just don't have any time. Yeah. Yeah. And the, the doer part of me definitely like overwhelms the talker part of me. Yeah. Every single time.

Yeah. So, well, I deeply—

thanks for your persistence.

Thank you. No, well, thank you for your time. You— and you gave me extra. Um, thank you so much. And I've wanted to do this for so long, so it was an honor. It was a true honor for me.

Yeah, honor for me as Well, anyway.

All right. Thanks, Alan.

Okay.

Talk to you later.

Thanks, man.

The Core Memory Podcast is hosted by me, Ashlee Vance, and/or Kylee Robison, or both of us together. It is produced by me and David Mickelson. Our theme song is by James Mercer and Jon Sortland, and the show is edited always by the Jon Sortland. Thank you so much to Brex and Send Cut Send for all your support, and thank you most of all to everybody for listening or watching. We love you. Please leave us a like, a review, a subscribe, all those tremendous things. Thank you, and we'll see you again.