Latest / Cheeky Pint / Casey Handmer of Terraform Industries on solar maximalism, hard tech, and reclaiming the Salton Sea

Transcript

Casey Handmer of Terraform Industries on solar maximalism, hard tech, and reclaiming the Salton Sea · Cheeky Pint

• Casey (00:00:00): Non-alcoholic Guinness. I have an insufficient filter as it is. John (00:00:04): You're perfectly filtered just the way you are. Casey (00:00:05): You flatter me. John (00:00:06): You're a solar maximalist. People should fully update their worldviews that solar is going to beat out everything. Casey (00:00:11): The solar wave is exponential, right? People are really bad at understanding where they sit on an exponential curve and then correctly projecting into the future. My grand plan is you put another 3 million feet, acre-feet of water in the Salton Sea, bring it back up to its historical 1990-era level. Stabilize the levels, stabilize the salinity, and we could do that and it would pay for itself. John (00:00:29): Have you pitched the Trump administration on this? Casey (00:00:31): I did write a one-pager for the Secretary of the Interior, but I think they've got a lot on their plate. It's not rocket science and I'm sure they'll figure it out sooner or later. John (00:00:38): I mean, it's a little bit rocket science. A lot of meat in there. Casey (00:00:41): Actually, yeah, it's probably rocket science. John (00:00:43): We had an Irish man here recently and that was a two-pint episode, so that got nice and unfiltered by the end. It's like, not that many for a session, but actually kind of a lot for a podcast. Casey (00:00:53): Especially if you go back to work afterwards. John (00:00:55): Casey Handmer is the founder of Terraform Industries, which is developing a machine that makes synthetic natural gas from sunlight and air. He's also a prodigious blogger and all-around polymath. Cheers. So you started your career at Hyperloop? Casey (00:01:13): That's where I first escaped from academia, yeah. John (00:01:15): Yeah. And a working career. Is Hyperloop ever going to happen? Casey (00:01:18): No, I think I can safely say that now. John (00:01:20): Okay. Casey (00:01:21): It's really kind of interesting. You had a bunch of very talented engineers that came together to take a crack at a project that they weren't sure was possible. And it turns out that… John (00:01:28): It wasn't. Casey (00:01:29): Well, it turns out that of all the known unknowns, they basically were able to solve them. The levitation, vacuum, construction, all that sort of stuff. It was this weird subtlety that ended up, if I may say, derailing the whole project. Which is that the faster you go, the further you want to be away from the surface of the earth. And you would understand this as a pilot yourself. If you make a habit of flying through canyons and mountains, sooner or later, the cone of ground that you can get to with a 7-G turn completely intersects the ground itself and then you die. And the same thing applies whether you're in a tube or not. So the practical consequence of that is if you want to build a network of hyperloops across the United States, you have to drill almost the entire way, which is a lot of rock. You have to move a lot of rock. It's very expensive. And when you start to tally up the numbers, California High-Speed Rails are supposedly north of a hundred billion dollars now. And you say, “Well that seems like a lot of money. How many 737s could you build and operate for 30 years for that much money?” And it turns out you could move more people more quickly that way than with the Hyperloop. John (00:02:28): You first turned me on to Henry Kaiser. Can you summarize the Henry Kaiser oeuvre for people and maybe explain why he's so understudied? Casey (00:02:37): That's an interesting one. So on my way in here today, I passed by the Kaiser Permanente just down the road. Of course you have to pay your respects. Henry Kaiser grew up poor, left high school, started off in photography initially, then went into hardware, then into construction, and a bunch of other things. Every huge project you've heard of, he worked on. He worked on Shasta, he worked on the Grand Coulee Dam. He was the principal behind the Hoover Dam, which at the time they called the Boulder Dam. Just an insane work ethic. He'd work 20 hours a day, every day for his entire life, right up until he died at 85. And he really got going after the Hoover Dam at the beginning of World War II when just up the road here, at Richmond and a few other places, started the shipyards where they took the techniques they'd learned at Hoover Dam and then applied it to shipbuilding. And they were ultimately able to churn out hundreds and hundreds and hundreds of ships— John (00:03:28): Including the famous four-day Liberty ship that was a propaganda exercise, but still they built a ship in four days. Casey (00:03:34): The more typical duration would be more like 11 days. What a failure. But just the idea of building that many ships in today's America seems stunning. And this is 1940. They invented employer-sponsored healthcare, which became Kaiser Permanente. That was actually initially introduced a few years before that. John (00:03:50): He effectively built the town of Richmond, California. Casey (00:03:53): Yeah. Amongst others. His retirement project was the Hawaii Village Hotel in Honolulu. What else did he do? He did aircraft for a while. He was involved in the Philippine Mars. He was involved in the Spruce Goose. He did cars. He set up South American auto manufacturing, including Jeep—Willys Jeep was Willys Kaiser Jeep for a long, long time. Just a bunch of different things. Now, why is he not so well known? Well, he kind of took off at the peak of his game and went to Hawaii and his organizations, including steel mills and aluminum mills and cement mills and all kinds of stuff, continued to operate but he didn't really do succession planning very well. And so when he died, all his assets flowed to his second wife and not to his children or his surviving son who'd been running the organization and the other executives. And she basically had no interest or ability to run it. So she sold off pieces of it here and there. John (00:04:38): So the empire kind of fell apart a bit. Casey (00:04:39): But pieces of it still exist. Kaiser Aluminum is still a going concern. It's still a very large company and Kaiser Permanente obviously, but it's a really challenging problem. How do you keep it together? John (00:04:50): Yes. There's an Elon-like aspect to his character where… So he started as a roads general contractor. He built roads— Casey (00:05:01): Like 200-foot little strips of roads with mules and wheelbarrows and stuff. John (00:05:04): Exactly. So he was the guy you called to build a road. And then he seemed to keep jumping to new technical challenges, all of them with the attitude that it's not that hard, it's only physics. We can learn it. And so the dam, it's like, “Well, that's just a big construction project. I've done construction projects before. How hard can it be?” Casey (00:05:23): He'd done a lot of dams by that point. He actually formed a partnership with LeTourneau. So he was the first person, at least on the West Coast, to introduce mechanization to road building and canal building and so on. He did the Colorado River Aqueduct, which is this huge watercourse that actually runs very close to my house all the way through southern California, bringing water from the Colorado River ultimately all the way to Palos Verde. Just insane stuff. But yes, this idea that you build this organization of ultimately some dozens of very capable independent autonomous executives and then you compound and compound and compound and compound and just always double or nothing, find ways to do it. But also, he had personal relationships with the bankers since forever. And so Bank of Italy came in and underwrote the Hoover Dam project for example, and they all made a lot of money. But in those days, the way that the contracting was— John (00:06:09): Bank of Italy, which was an American bank that became Bank of America. Casey (00:06:12): Yeah. Sorry, for those of you not familiar with the intricacies of finance history. Well, Bank of Italy, which was a bank based here in San Francisco, was the bank that opened its doors the day after the San Francisco earthquake and started handing out capital to people to rebuild even though it was going to take them six months to straighten out their accounts because everything had burned down, which they ultimately did. They understood the role of the bank, which is issue credit. Just fascinating stuff. And then in the late 1980s—so he died in 1969—late 1980s, two biographies were published back to back like six months apart, and then nothing. It's kind of bizarre how quickly you vanished without a trace. John (00:06:48): Yeah. And again, there was this first principles thing where getting into shipbuilding, getting into cars, getting into planes. What is a plane, but just a car that flies? Just a willingness to jump into new industries that, again, I associate more with hardware entrepreneurs than software entrepreneurs. Maybe it's something to do with the hardware space. There's more opportunities for new startups. I don't know. Casey (00:07:08): He was a tinkerer, so he had multiple workshops including some up at Livermore and some down here in Oakland where he'd have people on payroll working on stuff, just trying crazy ideas. He was into speedboat racing. He had an estate up in Tahoe where he'd sponsor speedboat races and then ultimately pulled back a bit when a few of them died, I think. But he just loved that sort of stuff. He was a mechanic. He was inventing things his whole life. But yeah, he had his hands in the stuff, so he said, “Well, we can figure it out.” John (00:07:35): Are there other underappreciated industrialist tinkerers? Casey (00:07:39): Oh, probably all of them are underappreciated. If you look at the Gundo guys or whatever, like this kind of new generation of industrially ambitious hardware entrepreneurs, almost all of them have an origin story that goes back into their early teens or earlier of just trying to make stuff. And it's just this sweat equity. I mean, on the software side, I would say you started coding well before you went to school to learn how to code. John (00:08:03): Well, I never went to school to learn to code… Casey (00:08:05): The same idea. But I don’t know, I interview people every day and I can't tell you how sad it is to be screening someone who's got a 3.9 GPA out of UCLA mechanical engineering school, which is generally pretty good. And you say, “Okay, this is your portfolio, but can you give me an example of one project you ever built using your skills that was not required for a class?” And they're like, “Why would I do that?” And I'm like, “I don’t know, because you just spent four years of your life deciding to specialize in this particular field of all fields?” John (00:08:30): So if you interpret Henry Kaiser, Howard Hughes, Elon Musk, they decide that their circle of competence is not rockets or roads or what have you, but is engineering. And that engineering competence can be broadly applied to many potential domains. I associate this trait with you. Maybe for the listeners who are not familiar, summarize what Terraform is building. Casey (00:08:55): Terraform is building a machine that connects to a standard quantized solar array and produces natural gas. It’s a pretty neat trick. John (00:09:04): Yeah, explain it for people who remember their high school chemistry. Casey (00:09:09): Okay, sure. The philosophy of what this is doing is very strange because normally we would burn gas to make electricity, but the solar has gotten so cheap that you can actually kind of do the reverse process and it makes financial sense, increasingly so over time. In terms of the chemistry—hydrocarbons are made of hydrogen and carbon, it's not very complicated. So the source of carbon is like carbon dioxide in the air. Actually this beautiful table is made of a biopolymer, which was also extracted from water and CO2 that ultimately came from the air. We also make hydrogen from water, and unlike this poor tree, we don't have to use Rubisco to do that. We can just electrolyze stuff at a couple hundred degrees Celsius and then feed that into a high-pressure, high-temperature reactor of the sort that were pioneered in Germany in 1905. And then out comes natural gas. It's a very neat trick. John (00:10:01): Okay, so you take electricity from solar panels. You take carbon from the atmosphere. You take the hydrogen from water, and then you use those to produce, currently short chain, in future longer chain, hydrocarbons. And the reason that you get carbon emissions when you burn gasoline is we're taking carbon that was sequestered in the ground. We had it stored away through all the many, many years of breakdown there. Casey (00:10:27): Dinosaur stuff. John (00:10:28): Exactly. Casey (00:10:29): So actually that carbon was originally atmosphere too, but it's been stored underground and it would stay there forever essentially. And so when you burn gasoline or kerosene, you're participating in this broad industrial scale, civilizational scale effort to pump carbon out of the crust and put it into the atmosphere, with many positive side effects of that process. In our process, the carbon comes from the atmosphere and is reused. So it's actually a burning wood, which is technically carbon-neutral. So that helps on the climate front in that you can create an economic forcing function around going and sourcing our entire civilization’s carbon from the air rather than from the crust. It also helps from an energy abundance front, which is to say if you need sun, you go outside. But if you need oil, you have to be friends with Saudi Arabia or one of the handful of other places on earth that are generally uninhabitable but have oil. But everywhere pretty much that people live has enough sun that you can grow enough food to survive. And if you can grow enough food to survive, you can synthesize enough fuel to run your modern industrial economy. Plants are great, even this plastic one here, but their net throughput efficiency is about a hundred times lower than a synthetic process like the one we're using in terms of area of land into reduced carbon fuel. John (00:11:40): There are people who think that we should all be flying less because of the carbon impact, whereas you think, “Hey, this is just a tech problem, we should all be flying much more”. And you can make zero-carbon flying… The fact that flying is one of the places where it's really useful to have hydrocarbons, but you can just produce zero-carbon jet fuel and then you can fly as much as you want in a closed system that is not emitting net new carbon. Casey (00:11:59): Yeah, I mean would you disagree that flying is awesome? John (00:12:02): No, but I'm probably not unbiased. Casey (00:12:05): Yeah. One school of thought here would hold that flying aircraft accounts for about 2% of emissions and it's almost all relatively wealthy people who are doing it. And therefore it's like this moral valence to it in that we're screwing up the planet on behalf of the 10 million richest people on earth. And so what we should do in order to disincentivize this behavior is tax aviation fuel and then feed the money into something to try and solve that social problem. And that will make aviation more expensive and less accessible. That's actually super regressive because people like you and I will still be able to afford to fly. We're not super price sensitive when it comes to flying United, but actually it's the people on the margin who could… It makes a difference between them flying, say once in their life and five times in their life. It makes a big difference to have access to those opportunities and those privileges. So my radical point of view is, I want to find ways to make aviation fuel cheaper and to build hypersonic magnesium aircraft that use liquified natural gas as their fuel and go much faster. And I want to expand the access to this privilege to everyone, not just the 10 or 15 million richest people on earth. John (00:13:09): If you say you want to avoid carbon emitting use cases, and so you want to avoid having carbon emitting flights or carbon emitting cars, there are two ways we can achieve that. One is we substitute a lot of internal combustion cars with battery-powered cars. And that's one way where you just stop—you eliminate the hydrocarbons altogether, or you could start using zero-carbon, zero net new carbon— Casey (00:13:33): Synthetics. John (00:13:34): Yeah, exactly. Synthetic fuels. Where do you think the world goes? Which things go towards electrification and just removing the hydrocarbons altogether and which things do we go to using clean hydrocarbons? Casey (00:13:44): Yeah, I mean I have opinions on this. At the end of the day, the market will decide. John (00:13:48): Sure, but what are your predictions? Casey (00:13:49): My predictions? John (00:13:50): You watch this market. Casey (00:13:50): Electric cars are really great and that— John (00:13:54): We just had RJ from Rivian and it's a great product. Casey (00:13:56): Yeah, exactly. And that over time we will see more other successful, hopefully Western, electric car manufacturers that can compete with Tesla, and Tesla can continue to grow its market share and so on. And it would not surprise me if when we are like crusty old men sitting here, let's say in our late eighties, that pretty much all the cars and all the trucks were electric, all the ground-based transport was electric. But at the same time, wouldn't it be amazing if aviation by that point had grown by 20x or 50x and so instead of 50 million people flying planes more or less, occasionally it was 500 million or a billion people flying planes, getting to fly in planes. And just that by itself would more than use all the oil that humanity currently uses. So I think the really energy-intense, high-speed transportation stuff will stay fuel. Rockets, obviously, will stay fuel. John (00:14:40): Yeah, to give people an intuition for this. Isn't it the case that versus current battery chemistry, jet fuel I think is two orders of magnitude—it's 100x more energy-dense by unit weight than batteries. And so it could be a while before you could build the battery power of 777. Casey (00:14:55): At the end of the day, we have a planet that has a certain size and a certain gravity, and if you want to fly from one side to the other, then you have to go 20,000 miles. That's actually a little far. John (00:15:02): 10,000 miles. Casey (00:15:02): 10,000 miles. If you think like, well, what could a battery chemistry look like that would have the energy density comparable to LNG? It's 55 megajoules a kilogram. I'm not saying it's impossible, but I would be surprised. I would be very surprised if that occurred. That said, I would not be surprised to see some really high-performance, electric-powered aviation in the future, particularly power densities are great. So you can do I think potentially much cheaper, much more safe electric VTOL. John (00:15:26): Yeah, there's clearly a bunch of exciting stuff. Casey (00:15:27): Short range, supersonic— John (00:15:28): Beta, Joby, all these guys. Casey (00:15:30): Yeah, exactly. John (00:15:31): But sorry, you're going back to dividing up the world. So you think big aircraft stay on hydrocarbons aircraft— Casey (00:15:36): Fast aircraft, big aircraft, long-range aircraft—hydrocarbons. Probably long-range shipping would stay that way. LNG, it's just so cheap. It's going to stay that way for a long time. Paints are going to stay hydrocarbon-based. Chemicals, pesticides, herbicides, pharmaceuticals, all that stuff. Plastics. John (00:15:51): Industrial gas turbines are really backlogged right now. I think people don't realize this, but one of the easiest ways to bring new power online, if you're building an AI data center for example, is you get an industrial gas turbine, which is basically pretty similar to a jet engine. In your vision of the world, shouldn’t those data centers be powered by solar arrays and batteries? That isn't happening yet, to the best of my knowledge. Why not? Casey (00:16:17): It is a very small scale, but I think we're just at the beginning stages of the market working this puzzle out. It doesn't happen instantaneously. So it seemed pretty clear to me, I think two years ago—I've got a blog post on this—that we're going to have to go to solar and batteries for AI data centers at scale. And back then everyone was still talking about building SMRs. And I love SMRs as much as the next guy. I used to teach nuclear physics. John (00:16:38): Small nuclear reactors. Casey (00:16:39): Of course, they're great for submarines. But you and I talked about this years and years ago, like, okay, what's it going to take to get a hundred gigawatts of SMRs up and running? It's not going to happen overnight. John (00:16:51): And we're getting into the thrust of the—as I think about “the Casey worldview.” You're a solar maximalist and you think that people have not internalized the cost improvements and the learning rate of solar enough and realized just how radically different the world will be as a result. And you're building Terraform as one bet on this, but you think just more broadly speaking, people should fully update their worldviews that solar is going to beat out everything. And then you're also a techno-hardware optimist that you think we should just be doing more hard things. We should be willing to make improvements. People should start hardware companies. Is that a good summary of your worldview? Casey (00:17:28): Well, I mean people in general maybe shouldn't start hardware companies. It's kind of cursed in a way. But if you have the right skillset and the right mindset and you're determined to just do the most ambitious thing you could possibly achieve, you should do it. Henry Kaiser could have had a grand career doing road building for his whole career, but he saw an opportunity and he saw that the American West needed water and he's like, “I should be part of this.” It didn't take Einstein to figure that out. And he's like, “Well, the government thinks that the Hoover Dam should cost $38 million.” That’s about a billion dollars in today's money to build. ”And it should take seven and a half years and they're going to pay us monthly. And I think that I can figure out how to shave two years off that schedule. So we'll bid $38 million, we'll get the bid and then we'll finish it in five and a half years.” And they did. Eight million goes in their pocket and that's how the government used to run acquisitions. So it works quite well. But at the same time, the solar wave is exponential. And we know from COVID, amongst other things, that people are really bad at understanding where they sit on an exponential curve and then correctly projecting into the future. John (00:18:27): What is the rate of panel cost, of solar cost improvement? Casey (00:18:30): It's like 40-something percent per doubling of production and currently production's doubling roughly every two years. So you could say 20% a year. That's the panel by itself, right? That's not necessarily the cost that you end up seeing at the wall. John (00:18:40): Right, and that's the objection people make to this line of argumentation. They say, panels are a pretty small component of the overall installed costs, and so costs are not improving that quickly anymore. Casey (00:18:49): So that’s true, but what tends to happen is the fundamental commodity at the base of your tech stack like de-costs and then other things kind of fall in behind it. So it is, for example, the case that the cost per transistor for a computer chip has fallen astonishingly quickly over time, and yet a Microsoft Windows license will still cost you a hundred bucks. Why is that the case? And it's true that actually Windows has gotten cheaper over time, especially when taking into account inflation, but it hasn't tracked the same cost curve. And that's the same sort of thing you're seeing here. But if you're able to build systems that sit on top of the fundamental technology with fewer intermediating layers, then you can take advantage of those cost savings. And then essentially you're trading the headaches of integrating with the raw technology and then pass those savings onto your customers. That's absolutely essential for synthetic fuels. There's no two ways about it. You cannot run synthetic fuels unless you have the cheapest possible electricity you can get. John (00:19:42): And you think in a way solar power will get much simpler and dumber where, for example, people used to optimize solar power by having movable panels that track the sun. And again, what happened as the panels became cheaper is, just put more of them down, don't bother having them follow the sun. And yeah, it costs a little bit more in the panel cost, but you save all that money in the super advanced hardware. And you think we have further to go in making the panels simpler. How far does that go? Casey (00:20:11): Economically speaking, we should have deleted trackers in 2016— the trackers of the systems that move the panels around. Obviously we're still in 2025 installing trackers. John (00:20:19): Why? Casey (00:20:20): Well, because— John (00:20:23): But the people who run these projects are smart, and they want to make money. I don't think “people are acting irrationally” tends to be a great explanation. Casey (00:20:30): Some amount of irrationality, I think, is involved. And I also think that there are certain participants in this market who are quite happy with cozy oligopolies and whatnot. So there's a lot of—I'd say the area’s ripe for disruption. So that was the first half of the question. The second half was, can panels get cheaper? They certainly can. So if you look at capital flows into solar panel manufacturing facilities in Southeast Asia, or the United States, or Europe or whatever, what assumptions they must be making about the economics in order to justify that investment. They could probably see line of sight to another 2x reduction in cost. So we go from 8 cents a watt down to 4 cents a watt. Just to pause for a second, 4 cents a watt is an absurdly low price. So the electricity that comes out of that is a thousand times cheaper than the cheapest food that humans can digest on a joule-for-joule, calorie-for-calorie basis. The limited physics on how cheap a solar panel can be is at least 10 times cheaper than it currently is. John (00:21:17): How would you achieve that? Casey (00:21:19): A very, very thin layer of structured silicon on some kind of inexpensive plastic backing or something like that. But in terms of, well fundamentally what do you need? You need some kind of transparent diode that creates this foamy-level nonsense that allows the electrons to do their thing as they go around. And in terms of limited physics, that could be the most generic thing imaginable. John (00:21:39): What was your line? “We're blessed that one of the most abundant elements on the earth squeaks electrons when you shine a light on it.” Casey (00:21:44): I don't know if that's my line, but yes, it's certainly the case. John (00:21:46): Well, it's a line you told me. Maybe you didn't invent it. Okay, so solar panels can get a lot cheaper. Why aren't you working on, why does Terraform not produce really cheap solar panels if you think everyone's doing it wrong? Casey (00:21:57): I think that the companies producing cheap solar panels are doing it right. They're doing a fabulous job. It's very competitive. John (00:22:02): But you just said people are buying expensive solar installs when they should be buying cheap ones. Casey (00:22:05): Well, so you've got the EPCs, which are the companies whose job it is to sell a solar array development to a buyer like a utility. They'll be buying the modules from probably a retailer, probably not directly from the factory, and they'll be buying other various components and doing the engineering and the procurement and the construction. The construction will be subbed out as well. So that model obviously needs to be refined. Why am I not competing in that? Maybe I am. But making modules themselves, I don’t know why you would go into that fight unless you absolutely had to. There's a lesson here. This is what David Senra talks about a lot which is really good businesses are built by commodifying a complement. And the complement being commodified here is, as much as possible, of solar development, which is pretty much reduced to practice and commodified already at this point. And the part that you sit on top of that is the thing that takes that and turns into something useful for the end customer. And then you end up providing that value transformation. John (00:22:59): You sell a kit. I think your nominal design size is one megawatt, right? Casey (00:23:01): Yes. John (00:23:02): That's what you guys are building for. And one megawatt, for context, is a few hundred homes? Casey (00:23:06): Maybe in Europe. John (00:23:07): Okay, yeah, yeah. In America— Casey (00:23:10): 20 or 30 homes here, I think. John (00:23:12): Oh, okay. 20 or 30 air-conditioned homes with the pool and everything. Casey (00:23:12): Ten thousand square feet. John (00:23:13): Exactly. Casey (00:23:17): Like a kitchen on the kitchen on the kitchen. John (00:23:19): Exactly, yeah. So that's a megawatt. And so you sell a one megawatt kit that, again, captures the solar power and produces the hydrocarbons. Casey (00:23:26): Instant oil well. John (00:23:28): Exactly. When will you have your first customer paying you for hydrocarbons that doesn't want you to turn it off? So not like a demo pilot or something like that. Casey (00:23:38): We're in the process right now of attempting to acquire land, which would then result in construction and development of these. All goes well, it could be early next year. I strongly suspect at this point—I’ve got a few bruises—not everything will go well. So we'll see. John (00:23:49): But potentially next year, you're selling oil? Casey (00:23:53): Well, natural gas. Yeah, so actually there's an interesting point there, which is obviously there's been some political shenanigans here. And we've got to the point now where actually we're looking at doing methanol accelerating the methanol pathway as well. The reason for that is that natural gas prices are basically set by the marginal producer, which is not us, and it's pretty cheap. Now, I think that as our export capacity to Europe increases and our fields continue to deplete and so on, we're going to see that price rise quite a bit, but probably not next year. So in the meantime, we can produce methanol and sell that at roughly six times higher revenue per carbon atom handled. Assembling molecules atom by atom is tedious. John (00:24:28): I'm sorry, where do we use methanol in the economy? Casey (00:24:30): Methanol, it's used for all kinds of things, but it's a precursor. So again, this is a somewhat subtle point, but Terraform’s job is to take sunlight, which is intermittent, but it's free. And in the form of solar panels, make electricity that's intermittent and then do a bunch of very scary chemical engineering to turn it into a tangible, storable, off-loadable commodity. So we make hydrogen, but I would never dream of trying to sell that hydrogen as our major product because there's no mechanisms for transporting that to market. No one wants it in bulk. It's an intermediate product, not the final product. We make carbon dioxide, same story. Put them together, now we have something that you can store in a tank. Methanol itself is chemically, in terms of boiling point and so on, it's very similar to water. So you can literally store it in a tank. Then if you want to, you can pipe it across the road to a standard old-school refinery and they can turn it into gasoline or something. I could do that myself. I could build a refinery, but what is Terraform good at? We don’t know much about making large-scale chemical plants and we'd have to go and license the MTG process anyway. Methol a gasoline process. So why not just hand that off to someone who can do that already? And we just focus on the thing that we do really well. John (00:25:39): The methanol to gasoline process is like a proprietary process that you have to pay— Casey (00:25:44): A local development in the seventies. John (00:25:45): But is it patented? Is it a trade secret? Casey (00:25:47): Almost certainly had a patent at this point, but— John (00:25:49): It's just protected by trade secrets? Casey (00:25:51): So the thing is, even if I had all the patents in front of me right now and blueprints of the plant, right? John (00:25:56): It's just process knowledge. Casey (00:25:57): And access to half a dozen expert chemical engineers had done this before and then I went and built it. I promise you it would still not work. John (00:26:03): It's Walter White's cook. It's just the process. Casey (00:26:06): I've not seen that show, but I mean it's the same as brewing beer. I can buy all the pipes, I can make something undrinkable, but— John (00:26:14): I believe that. Casey (00:26:14): Oh yeah, but– John (00:26:17): I mean you're Australian. It's in your culture. Casey (00:26:19): Now, now. I'm American now as well. Fifteen years [on] Sunday. It's a very finicky process. I think it involves blood sacrifice to get it right actually, to be perfectly honest. John (00:26:30): Okay, so that's methanol. And sorry, you mentioned you could do cement. How does that work? Casey (00:26:33): Well, so we use what's called lime calcite cycle to capture CO2. And so that involves calcinating calcium carbonate to make calcium oxide, which is the active ingredient in cement. So that's 60% of the difficulty of making cement there. So we do that already. And then cement is about 7% of global emissions. The challenge with cement is that it's 40 bucks a ton at the plant. And so if you want to compete—one of the big mistakes that I see in my sector generally, which is the climate-aligned energy sector and so on. This applies not just to the climate companies, but they say, “We figured out how to do this.” That's very impressive, but are you able to actually sell the setup price that people want to buy? John (00:27:12): Then having a shiny demo of something that works but is not commercially viable. Casey (00:27:17): Even that is impressive in itself. I mean, Rivian for example, they make a beautiful product but they're yet to make money with it. But just getting a product to market that people like… That in itself is worthy of admiration. Beautiful car. They've just got to figure out how to delete about 30% of the components and they'll be in business. And far be it for me to criticize. But if you are able to make cement with a carbon neutral process and you're like 60 bucks, you probably have an argument that you could find enough off-takers who want to pay for the carbon credits or whatever. Then you can stay in business. But if it's $6,000 a ton, it's very slim. John (00:27:51): How causal were the solar subsidies along the way, like Germany's, to solar success on the current cost curve? Casey (00:27:58): That's a good question. I think it's very hard to run the counterfactual. I think that the Germany solar subsidies did not accrue the appropriate reward to Germany that they probably should have. John (00:28:10): But I don't think they were intending to— Casey (00:28:12): But it would've been nice if they had. John (00:28:13): Sure. Yeah. But I think they were more motivated by do-goodery than… Well, these things are always bootleggers and Baptists. Casey (00:28:21): The total amount of money they spent on, it was not all that much in the grand scheme of things. John (00:28:24): But I guess what I mean is, learning rates are about the idea that with increased volume comes lower cost and you get to be competitive. And again now— Casey (00:28:34): So in the counterfactual where Germany didn't do it, right? Well, Spain still did it a little bit and Australia did a little bit. Maybe it would've delayed things by a year or two. But as soon as, say China, started to do it at scale, they would've slurped up that time difference and we'd be where we were today. John (00:28:46): Well, and then how impactful were the Chinese interventions in the market— Casey (00:28:50): Enormously. Enormously, right? John (00:28:53): So Chinese subsidies accelerated solar. Casey (00:28:55): China has deployed more solar in the last three months than the United States has ever deployed, ever. They're becoming an electro-state, and this is an important point that I think is widely underappreciated. The United States has formed an alliance with Australia and Japan, and I believe India—I can't remember exactly—to basically try and control oil transiting through the South China Sea and adjacent areas, because they believe that will give them the political leverage they need to force China to play nice. And China currently imports something like 12 million barrels a day from the Middle East across sealands they don't control and can't control. Past countries who are not all that friendly to them in general. And they understand this. And the Virginia-class nuclear submarine is being built in part to service the needs—allied needs—to police the Strait of Malacca. That submarine is due to be delivered to Australia in 2035. Do you think by 2035 China is still going to be as dependent on oil through the Strait of Malacca? I don't. At the rate that they're deploying right now, they could cut off that supply if they needed to. In the same way that Germany was ultimately able to cut itself off from supplies in 1939. And it didn't really kill them until 1944. John (00:29:58): But it did kill them. Casey (00:29:59): Ultimately, yes. But that was mostly the combined bomber offensive delivered by a fleet of hyperloops flying from—oh no wait, they were planes. John (00:30:07): And so the way that China reduced its oil dependence is via electrification or via Terraform-style synthetic fuels? Casey (00:30:16): As far as I know, they're not working on synthetic fuels, at least not Terraform-style, but it is not rocket science. And I'm sure they'll figure it out sooner or later. John (00:30:23): I mean, it's a little bit rocket science, a lot of meat in there. Casey (00:30:26): Actually yeah, it's probably rocket science. John (00:30:27): Going back to the subsidies, how did the Chinese subsidize solar panels? Casey (00:30:32): I actually don’t know the specifics. John (00:30:33): Okay. Casey (00:30:34): What I do know is that they basically offered government-backed, zero-interest loans to manufacturers, for example. And they also— John (00:30:39): Cheap capital. Casey (00:30:40): That's right. Yeah. Cheap capital. So a lot of these technologies are developed in Australia. John (00:30:46): I would like a government-backed, zero-interest loan. That sounds lovely. Casey (00:30:48): Yeah, it'd be great. Actually, to be fair, the financial and legal situation in the United States, as much as we complain about it, would you rather have built Stripe in Hong Kong or something? John (00:30:56): Oh no, no I agree. The US is a fabulous business environment. Casey (00:31:01): It’s a great business environment. The capital is cheap enough. That's the important thing. John (00:31:03): Where I'm going with this is, if the German subsidies were somewhat impactful and if the Chinese subsidies were very impactful to the development of solar, what are the implications here for your grand theory of industrial policy and intervention? Where, if we want to be able to have nice things like solar panels and things like that, Silicon Valley tends to have this libertarian bent, you know, keep the government out of our innovation. But it seems like top-down government intervention was impactful in bringing solar to us sooner. It would've happened anyway, but it would've happened slower. And so, what do you think sensible industrial policy is for the US or for whatever government with lots of stake— Casey (00:31:47): A bit above pay grade there. I mean, it's an interesting point. You say, well, the counterfactual surely would be if China was a Western liberal democracy that was also grappling with the fact that it was enormously dependent on oil imports and it was unable to deal with the Middle East and maybe wasn't totally friendly with the United States. But they were Western liberal and they had the same business freedom we take for granted here. Would they have been able to mobilize the same amount of capital to go and build solar panels? And I think the answer is probably yes. John (00:32:10): But no, the counterfactual is just, the Germans never get around to doing the subsidies. The Chinese never get around to doing the subsidies. All the other subsidies are too small and solar happens much more slowly. Casey (00:32:19): Maybe. But at the same time, actually, I kind of sit on both sides of this issue. You can look at the process of Soviet industrialization. There's a fabulous book called Red Plenty by Francis Spufford that is kind of a fictionalized account of this. And you can say, “Isn't it amazing that the Soviets managed to go from this kind of desperately underdeveloped backwater then level up prior to and during World War II?” And then World War II, having lost 20 million people or something, put the pieces back together and then execute a five-year plan after five-year plan after five-year plan. And then by 1980 they were more or less at parity with the poorer parts of the West. “Isn't that amazing?” You say, “Yes it is. But if you look at GDP growth over that time, it's actually slower at its fastest point than the United States was at its slowest point,” which indicates to me that miracles were achieved in the Soviet system. But even so, it would've been more efficient had they had a freer market, for example. John (00:33:02): Sure, yeah. But I think we can have something between “no intervention at all is bad” and “communism is bad.” Somewhere along that spectrum. Casey (00:33:09): Somewhere between? John (00:33:10): Yeah, yeah, yeah. Casey (00:33:11): I don’t know. I feel like you’re— John (00:33:12): I'm baiting you, yeah. Casey (00:33:14): You've come from the British Isles and it's a very British Isles thing to say. No, I don't think anyone in Silicon Valley really complains when the government shows up with a big check and says, “We would very much like you to now build the atomic bomb or to build nuclear submarines.” Palantir is doing great business right now. AWS does great business with the government, et cetera, et cetera, et cetera. And I think that's actually a pretty good example of that working quite well. And actually a lot of the secret history of Silicon Valley is all government work, Raytheon and Lockheed and so on. And indeed Terraform is right across the road from Skunk Works, which was itself obviously largely government-funded. I think that's a pretty good way. I think that Congress, the way that US soft power always works, is with a fat wad of bills and that tends to be metabolized well by the US economy. Do I think that we should raise taxes to spend more money on DOE picking winners in the US domestic solar panel manufacturing space? No. I think that probably distorts the market overall. John (00:34:05): But are there places where we should be picking winners? Casey (00:34:07): Probably, yeah. Or there are at least places where there should be more like decisive government intervention. Absolutely. I think on the energy front. Countries that have cheap energy can afford to screw up a lot of things and they're still basically, everyone is prosperous. And actually I would characterize the United States as one of those countries. Not an overgoverned country, and in many ways a barely governed country. John (00:34:28): I think people in the US don't realize how expensive electricity is in other Western countries. Casey (00:34:33): They probably don't. Yeah. Although the people on the Twittersphere who are like the “Euro poor” meme or something. By the same token, there are very, very few developed Western liberal industrialized economies that have very expensive energy where they do have good GDP growth, for example, and good wealth. And you say Japan has expensive energy and their GDP has been pretty much flat, and UK’s GDP, flat. Singapore’s GDP per capita is not spectacular by comparison to the United States. And so the moral of the story is when a new technology like solar comes along, you need to make sure that your markets are able to deploy that at the scale necessary to make sure that you stay ahead of the pack. Because we're already in a situation right now where China is out-producing the United States on all the metals like steel, aluminum, magnesium, and so on and so forth, by a factor of between 10 and a hundred. Because their energy is cheap, because they basically have interest-free loans to completely unregulated coal mines, amongst other things. And basically, to the victor go the spoils. So that's something we should be careful about. John (00:35:31): If you could wave a magic wand to speed up US solar deployment—not manufacturing, just deployment—what would you do? Casey (00:35:37): The single biggest intervention you could allow is by right development of solar panels on private land, and ideally even like BLM land would be great. BLM land has been set up for ranching and you can run your cows on it with very little friction. And I think sections of that should be set aside for— John (00:35:54): It's just electron ranching. Casey (00:35:55): You can put your solar panels there and you have something like California's fire insurance scheme, so that if one batch of solar panels goes bad, or after 50 years we have fusion and we want to take them out, there's money there that we can take them out and put them away. It goes back to being desert. But right now to permit a solar array, even on private land for the sort of applications we want to use, it's often just as tiresome and expensive and difficult as permitting a new chemical plant or something that's— John (00:36:23): Much dirtier. Casey (00:36:23): Much dirtier and much worse for the—actually, I think solar overall is a net benefit for the environment. We should be encouraging it. It should be like, you get credits for deploying more solar on government-owned land. Ninety percent of Nevada is like federal land and that's the best place. The reason it's federal land is they couldn't give it away back in the day because it's so inhospitable. John (00:36:41): I'd never thought about that. It's the land that's left over. Casey (00:36:43): Yeah. John (00:36:44): Canada has this too, where a huge amount of Canada is Crown land and it's just the untitled document of Canada. It's just the land that they never did anything with because all the settlement happened right next to it. Casey (00:36:52): No one ever ran a railway through it because why would you? Yeah, that would be nice. And every time I fly to Texas or wherever I look out the window— John (00:36:58): Nevada's very empty. Casey (00:36:59): There's about a hundred valleys in Nevada, and one of them is already basically used for nuclear testing and one of them is for Vegas and one of them is for Reno, and there's like 97 left over. One’s for Burning Man. John (00:37:08): Have you ever looked out the plane flying near Vegas and you see the big solar collector? It's like something out of a sci-fi movie. Casey (00:37:14): Ivanpah. John (00:37:14): It really catches you by surprise. Casey (00:37:16): They're going to close it down, so enjoy it. John (00:37:17): Are they? Casey (00:37:18): Yeah.. Never really worked very well. John (00:37:20): Okay. Casey (00:37:21): But as you're flying past the Henry Kaiser Monument, Hoover Dam near Boulder City, there's a huge solar array right there. If you've ever landed that airport, basically all the way from there down to the border is solar now. John (00:37:31): That's cool. Casey (00:37:31): Which is good. John (00:37:31): We're talking about Chinese manufacturing here, and obviously the American dynamism movement has a lot of energy right now. It feels like there are some things where we are obviously not going to—obviously to me—not going to be competitive in manufacturing. Like, just general light electronics. That stuff doesn't seem to me like—consumer electronics—is not necessarily coming back. However, we just had Keller from Zipline on, talking about the drone manufacturing they're doing right here in South San Francisco. You guys are manufacturing in Burbank, for your sins. Do you think it kind of makes sense to think about areas of manufacturing we should be trying to win or trying to succeed in? Casey (00:38:13): Yes. John (00:38:14): Okay, then what are those areas? What would you say? “Whatever, it's not coming back, we're just going to let it go” versus the ones that we really want to win or those national security considerations or what have you. Casey (00:38:25): Well, conventional wisdom would say, well, we've got to maintain our supply chains for our national security stuff, so we'll keep them as a life support, but all the rest doesn't matter. And we've seen how well that goes. I think you actually can't afford to say, “Well, we'll just let the Southeast Asians keep consumer electronics manufacturing.” If you want to have the ability to rapidly prototype electronic systems, you've got to have that here servicing the US market, which is the strongest consumer market on earth. And there's no reason you can't do it. No, you don't want to have 14-year-olds in sweatshops soldering stuff together. But if you look at the Bastrop plant where Starlink terminals are made, it's all the United States. It's 99.99% automated, and that's how we can do it. We should be doing it here. I grew up in Australia where the limited manufacturing sector we had just evaporated overnight when China basically came in. And now Australia's in deep trouble really on the basis of this. But at the time it was like, “Well, we can't compete with China. It's impossible. You could never compete with China.” On what basis? There's no reason why you can't compete. Obviously your factors of production are different, so you have to balance them. But this whole idea, comparative advantage that the United States should just be intrinsically bad or uncompetitive at certain kinds of things is crap. It has every advantage you could possibly want. John (00:39:28): What success stories should we be looking at to model US manufacturing after? So you mentioned how about— Casey (00:39:34): How about its own? John (00:39:35): How do you mean? Casey (00:39:36): The US was, and in some ways still remains, the manufacturing powerhouse. Absolute titan. John (00:39:41): Yes. But again, I feel like that was quite a different environment of more trade barriers and different labor dynamics. And I'm saying, if you think about manufacturing today, SpaceX's Starlink terminal manufacturing plant, which is very impressive and a much larger scale than people realize. How many Starlink terminals are they producing now? Casey (00:40:01): It must be like a million a year or something. John (00:40:02): Yeah, it's a lot. But yeah, who else should we be looking at and saying we should be doing more of that kind of thing here? Casey (00:40:07): Well, in terms of us, I think Chris Powell is doing really well at Hadrian, but I think— John (00:40:09): What does Hadrian do? Casey (00:40:10): They make parts for military aircraft, mostly. But the idea is you increase the machine-to-machinist ratio with software, which is not exactly astounding. How many cores are in your phone? I mean, it's actually quite hard to do in practice. Casey (00:40:26): And I think the edges or whatever, the corner cases that the current entrepreneurs are grabbing onto for mass production is like, “Well, I don't have $10 billion to go and set up Foxconn 2.0 here in the United States if I even wanted to. But I could build this really high-value product for a discerning customer, US branded type stuff.” And so Sam D’Amico, for example, down at Impulse has done that with his stoves, right? It's a premium product when people need to replace their stoves, they do it in a hurry. His is available, has all these other ancillary benefits because it's so nice, he can sell it. John (00:40:57): But I think— Casey (00:40:57): They're actually being produced in China, though. John (00:40:59): But isn't it hard to rely on producing premium products? Casey (00:41:02): That's where you start, right? That's the beachhead and then you diversify outwards. But if you're able to build a mostly automated factory that can produce a few thousand units a year, then it's relatively straightforward to go from there to 10,000, 100,000, a million. And then as you get those economies at scale, continue to push the price down. I saw some stat the other day that Apple had spent a hundred billion dollars on manufacturing investment in China. Imagine if they'd spent it here. Do you think they could not have succeeded in building a really nice iPhone even here in Silicon Valley if they'd spent a hundred billion dollars in factories? John (00:41:32): People in tech follow Elon a lot and have read the Walter Isaacson book, and they're familiar with the oeuvre and know much about how he works. Casey (00:41:40): Ashlee's book is good too. John (00:41:41): Ashlee's book is very good as well. What do you think, nevermind people at a distance or people are very bothered by Elon or something like that, people in tech who kind of follow this stuff. What do you think they still misunderstand about the Elon production function? Casey (00:41:55): Oh, that is a good question. Well, I think actually I'd be more surprised if there was someone who had succeeded as well as Elon had in his various fields who was better liked. If that makes sense. You should expect to make enemies if you're going to be that effective in that sort of area for whatever reason. I mean, he has his own— John (00:42:09): So you think the hard edgeness is not an incidental trait. You think it's part of the success. Casey (00:42:14): Yeah. Yeah. I didn't always think that. I used to think that it probably was possible to build an organization as successful as SpaceX without being quite so uncompromisingly demanding of your employees. John (00:42:25): And then you started a company. Casey (00:42:26): Then I ran my own hardware company, and I'm like… John (00:42:28): And so how did you update? Casey (00:42:29): Many hard lessons learned over and over again until I finally worked out that actually, people are in general capable of achieving far more than they would think. But almost always they require extrinsic motivation to do that. And usually it's like, “I expect you to do better,” kind of thing. John (00:42:44): But there are lots of people who are very successful and generally are regarded as pretty nice people. Casey (00:42:49): Successful, yes. But successful in building competitive hardware manufacturing companies in the United States— John (00:42:54): Okay, so you think it's intrinsic to hardware? Casey (00:42:56): Hardware is, I think… Well, statistically speaking, it would seem that a software company is a surer ticket to ungodly riches than a hardware company. That would be my guess. John (00:43:08): Okay. So that's one thing. You have to be somewhat hard-edged. What else? Casey (00:43:11): I think he works really hard. I think that he compounds his own knowledge. I think he's super smart. I think he understands that physics is the law, everything else is a recommendation. But yeah, I don’t know him at all really. So I can only speculate from the outside. I think that one of the things that frustrates me somewhat about Isaacson's book is Isaacson is quite clearly a very sharp guy, and he spent a year and a half with Elon, and yet if you read his book, particularly the last third or a quarter of it, which is talking about all the recent stuff, the Twitter stuff and so on, it's just very clear that he doesn't get it. He doesn't understand why Elon is doing these things that seem completely crazy, but yet he keeps rolling sixes. Why is that the case? And my point of view on this— and I wrote a blog post about this at one point—was, if I see someone who is pretty successful do a series of crazy things over and over again, and yet they work out every single time. Either they have root access to the simulation or their model of reality is better than mine in some key way that I don't understand. But I might be able to understand if I assume that they made a sensible decision, given knowledge that they must have, but I don't have. What can you infer about this? John (00:44:17): You've talked about how many SpaceX alumni have gone on to start rocket companies that are essentially just trying to build Falcon 1. And despite the fact that SpaceX was kind of doing that for the first time and had to learn lots of hard lessons, they got to work at SpaceX and they knew all the lessons and they'd seen everything and still a lot of them have struggled. I haven't heard you espouse a theory as to why they've struggled. Casey (00:44:46): It's got to be really hard. Yeah. Well, I know that a lot of people leave SpaceX and a few of them are like, “I'm going to start my own company and I'm going to try what it's like to be Elon.” And then they get a horrible experience. And some of them are like, “I'm glad I got a horrible experience.” But I think a lot of them are like, well, “I always disagreed with this particular technical choice.” And then they make it differently and then the rocket explodes. And maybe it's because that choice was actually correct and they didn't understand it. And maybe it's because it's unlucky and maybe the organization didn't work well enough. In the space sector at least, it's quite astounding the extent to which SpaceX is dominant. This is not the case that Apple has, 20, 30, 40% market share on smartphones and Pixel has 10% and Huawei and so on has… This is like SpaceX has 95% market share for the entire world. And then the second place getter is China or something. And then the next closest was 15 years ago the United States' main launch company. And then it's very much what I'd call an “anomaly in the matrix”. (00:45:41): Solar getting cheaper every year is also another anomaly. It's like a big wave that can be surfed and people don't really recognize it or how to use it or how to lever it. John (00:45:48): What advice do you have for highly technical youngsters? Say someone who's in college listening to this. Casey (00:45:55): Damn. Well, don't validate my life choices by repeating them. But realize that your life is on a spectrum—a continuum. And certain things are easy now and they'll be hard later, and vice versa. And if you see yourself as someone who's the technological elite, you should challenge yourself as much as you can. You should go out and find a worthy challenge and just level up. Don't go to someplace that puts you in a holding pattern for 10 years. You've only got 40 something years of really good work ahead to level up as much as you can. John (00:46:25): Go work somewhere that's very demanding. Casey (00:46:27): As much as you can take, for as long as you can take it. Yeah, probably more than a hundred people have asked me over the years, “Oh, I got into grad school and I got into SpaceX, what should I do?” And I said, well, the psychic damage is the same, but the pay is better at SpaceX, which is hard to believe but true. And I really believe that. You want to go and work with the best people on the things that you care about. Whoever they are, wherever they are, whatever it takes. Don't settle for second best. On the other hand, if you've recently graduated or you're about to graduate, it's more important that you get a job and you kill it. Do really, really well straight out of school then that you sit on the job market for a year and then compete against the next round. I see a lot of resumes. No doubt you've seen a lot of resumes over the years, and really you're only as good as your most recent job. And if your first job out of school is super important to set the foundation upon which you can build the rest of your career. Even if it sucks, stick it out for 18 months to two years and make sure everyone there has something nice to say about you. and then you can build on that. But yeah, it's really sad when you see someone with an ostensibly decent degree out of decent school and they kind of failed to launch in some way. John (00:47:39): As Casey is describing, and as we've heard a lot about on Cheeky Pint, prototyping is easy but production is hard. A surprisingly large slice of human progress comes down to reducing the timing cost of industrial production processes. In Stripe Press’ newest book, The Origins of Efficiency, author Brian Potter breaks down precisely how that progress happens. He draws on a hundred years’ worth of research to explain the fundamental steps of a production process and how each can be made faster, cheaper, and more reliable. He includes detailed examples from steel and semiconductors, to wind turbines and container shipping, to Tesla and the Model T. And he discusses how the principles of production efficiency can be applied to domains like construction, medicine, and education, where much work remains to be done. And Brian should know, he's the author of the popular Construction Physics newsletter and worked for 15 years in the construction industry. The Origins of Efficiency just came out, You can buy it on press.stripe.com where we offer all Stripe Press books. How are you planning to educate your kids? Casey (00:48:38): Mostly benign neglect, I think. John (00:48:39): No, but actually. Casey (00:48:40): Well, they're in the local school. John (00:48:43): But presumably that's not the limit of your plans. Casey (00:48:46): Yeah, I mean so, I'm not worried that they're going to find an uninspiring academic environment at home, right? We have a house full of books—that's going to be fine. So why do we send them to school? To get the stuff they can't get at home, like how do you deal with people who aren't like your immediate relatives and as obsessed with space as you are? That said, it seems not entirely unlikely to me that at some point the schooling system will, I don’t know, not be quite sufficient in some way. John (00:49:11): But in particular, the Bloom 2 Sigma effect is one of the biggest effects in education. The idea that you get way better outcomes with one-on-one tutoring focused on mastery learning, where you only progress through topics once the student has mastered the prior topic, and you move your way sequentially through things. And schools almost definitionally can't do that. Casey (00:49:38): They can't do that. Yeah, I actually think that maybe the Bloom 2 Sigma effect, you can achieve it on your own if you're able to follow your areas of curiosity and have access to resources and learn. So actually in my case, probably the reason I'm not a chemist for example, is that I didn't have access to chemistry textbooks until I was in late high school. What I want for my children, especially in a world with AI and all kinds of nonsense coming, I don't really care what they do as long as they love it and they're as good as they can be. They have to love it and they have to try. They have to sweat, they have to contend with the material because incredible privileges have come to them and to me and to some of my more recent ancestors. And I feel like you do have a social obligation to not waste that opportunity. John (00:50:24): Do you buy the explanation on the falling fertility rates that we've just made it too expensive to have kids? Have you seen the “booster seats as contraception” paper and all these? Casey (00:50:35): I just forwarded that to one of my Congress people actually, when they were saying, “Come to a California Highway Patrol session on fitting your car seats.” And I was like, “How about we just tweak the law? So if you have three or more kids, you don't get them taken away for putting them in a sedan.” For reference, something like 58 children's lives are saved in the United States every year, statistically, by car seats and something like 10,000 births are averted by families who cannot afford to upgrade their cars, which is a bit of a problem. Even if you add to lose 10,000 additional births, it still wouldn't reverse the demographic decline. John (00:51:07): Well, but that's one example of where, just I think people don't fully internalize the barriers to having kids. So car seats is, again, it sounds like almost an oddly trivial one but it does have a measurable impact. And that's before you get to housing, education— Casey (00:51:22): I think you could certainly say, well, what if we took all these things that have some marginal effect and we just turned them all up to 11 and saw what happened. And I think you'd make a plausible argument that you'd see a bump in fertility, especially if you had a pro-natalism law package that did all these things at once. And so you start a national conversation around how kids are great when you have more kids and we specifically focused on existing families and got them to go to N plus one or N plus two. Make a big difference. And that's what Australia actually did when I was in, I guess my teens, so let's say around 2000, was like, “Have a third child for the treasurer.” But the same problems kind of stick there. But this is obviously a very complicated problem because there are many countries on earth where these counterfactuals exist, at least in part, and it hasn't miraculously turned around the birth rate. And actually the number of developed Western liberal democracies where the birth rate is not either already catastrophically low or falling fast is very, very small and probably will be zero before long. But it did occur to me today, having spoken to a friend of mine about children, whether or not to have children and so on, that our evolution did not depend upon us in the past having a deep obsession with the need to have children. Because we didn't have contraception and we had sex drive. So children would just happen then. But once you have kids, then you're all about kids. As soon as you hold the baby in your hands, you're like, “Kids are fucking great! This is great. I might have six more of these.” But until you have that first one, you don't realize. You don't know. It hasn't come in yet. There's no evolved necessity to go out and have kids. Whereas with food, it's obvious. People get excited about food. So it sucks to be evolution in that case, I guess. But actually if you run the math on our demographics right now and kind of how dire it is, even if as a result of listening to this podcast, every single person in the West just immediately started having babies, it still wouldn't really for many decades… This hole we've dug for ourselves, there's basically no way out of that. There's no way to avoid it. I think the news that came out yesterday was the triple lock in the UK is going to go up by 4.7% against a backdrop of inflation of two-point-something percent and economic growth of like 0%. John (00:53:29): What's the triple lock? Casey (00:53:30): The triple lock is the process by which the pension in the United Kingdom is increased by the biggest of three separate factors. One being the employment rate, one being—I don’t remember exactly what they are but I think one is inflation, one's GDP. So it's like this insane ratchet. And the dependency ratio is of course increasing because people are getting older and older, because we're able to keep people alive for longer. And these are actually all good things, but we need to have some way of thinking rationally about paying for these things or making sure we're able to provide for people adequately because otherwise we'll burn the seed corn. And I saw this insane stat in France recently that the average pension is now higher than the average income for working age adults in France. This could be a problem. And I just don't think there's any way out other than Arc Institute gets its act together, invents a pill that keeps me young and everyone else young and lifts this burden of non-negotiable disease from every single person who's lucky enough to live through their youth. John (00:54:25): Don’t we also maybe get bailed out by AI productivity growth? Casey (00:54:29): Maybe. But do you really think that 10% year-over-year productivity growth is going to localize itself in the UK? John (00:54:35): No. But just if you— Casey (00:54:39): Is the United States going to be like, “Oh, here's a big transfer payment for all your old people.” I don't think so. So it's certainly the case that AI seems likely to produce enormous additional wealth. Where that wealth will find its way to alleviating the tax burden on the working class in the 200 countries on earth that don't have their own AI industry seems somewhat unlikely to me. Whereas a pill that—actually Ozempic probably does this—slows down your aging by some factor would be enormously important. No one could have seen this coming. No one could have possibly anticipated that Social Security and other stuff would result in ever increasing dependency ratios. And we're just fortunate that we got rich so fast that it took this long to bite. It's a big problem. John (00:55:19): I realize I'm asking you for ideas for hardware companies. You're always banging on about desalination, presumably that's a good hardware company. Casey (00:55:25): I'm working on it. I already have a hardware company doing it. John (00:55:27): Oh, you do. Tell me about your desalination company. Casey (00:55:29): So desalination is not strictly ripping oxygen off stuff to make fuel or metals, but it does involve ripping water off salt. So the oxygen and water is what makes it polar. Yeah, so Terraform Desalination is the project. It's very early stages now, but it's kind of insane to me that we live in this state in California. John (00:55:47): And can you also just do it with electricity and much else? Casey (00:55:50): Yeah. John (00:55:51): What's the reaction? Casey (00:55:52): The reaction is that the salt which is dissolved in water, gets separated somehow, either thermally or with filters. I mean there's probably 30 different ways of doing desalination. John (00:56:00): Okay. But what’s the Terraform approach? Casey (00:56:01): The Terraform approach is thermal. Yeah. So basically distillation on steroids. John (00:56:05): I see. So you're just boiling off the water and recapturing it? Casey (00:56:08): Yeah, essentially. John (00:56:09): Okay. Casey (00:56:09): Yeah. John (00:56:10): Do you have to boil it? Casey (00:56:12): There has to be a phase change. John (00:56:13): Okay. Yeah. Casey (00:56:15): When people think boiling, they think hot, but you can actually boil at low temperatures if you reduce pressure. So we're getting into trade secrets here and there's nothing new under the sun. The challenge is to Terraform-ize this technology and make it cheap enough that you can run it intermittently off solar arrays, or solar power in general and still make money. That's quite challenging. But it is insane to me that we live in this state in California where they're talking about doing this project under the delta in the Sacramento River Delta. You spend $20 billion on it and it's going to potentially recover on the order of a hundred thousand acre feet of water a year. If you spent $20 billion on desalination, you could literally double the flow of the Colorado—5 million acre feet. John (00:56:52): So why is California ignoring desalination? Casey (00:56:55): Again, I think it's anchoring to the past where the way you got water was you went and built a big dam in the mountains and then built canals and stuff. Henry Kaiser, eat your heart out. And desalination. I mean there's the odd desal plant here and there, but they're very small. They're mostly set up to help cities out. John (00:57:10): Oh, there is desal in California already? Casey (00:57:11): Yeah, there's a plant in Santa Barbara. There's a plant I think in San Diego, maybe. John (00:57:15): And they supply water for the city just— Casey (00:57:18): Normally they don't actually. It's only if the city runs short on water than they do because they run off the grid and the reverse osmosis plants, they're quite expensive to operate. They also have wear parts, so as you run them, it costs money, John (00:57:30): But yours will not have significant wear parts? Casey (00:57:32): Hopefully not. I mean, it's mostly spreadsheets and some doodling in the kitchen at this point, which is how Terraform started really. And then the garage and then in the castle. But California has more coastline on the West coast than any other state. Maybe Alaska has a bit more, I don’t know. So why are we short on water? It's right there. John (00:57:50): Are there private entities you can contract with? Just because government procurement cycles are very long and politically involved in everything like that. Can you sell to the city of Irvine? Are there ways? Casey (00:57:57): Yeah, I mean that would be the way to do it or to sell directly to a consortium of farmers or something like that. The challenge in California is the California Coastal Commission, which I believe you've probably encountered. John (00:58:08): No sense of humor. Casey (00:58:09): For a variety of interesting, and I think mostly valid reasons, they sought to try and slow down development of the California coast, which is fine. I don't want to develop the California coast. I want to bury a pipeline under the beach that goes out to the ocean and sucks up some water and then puts back some salty water and it's fine because the ocean's made of salt water. John (00:58:25): Are you putting back salty water? Casey (00:58:26): Yeah. So what happens is you take off salt water. So you take 10 gallons of salt water and you pull out one gallon of fresh and then you have nine gallons of slightly saltier water that goes out and mixes. And it's fine because that's how rain has formed in the first place. John (00:58:40): It’s the ocean, it'll be fine. Casey (00:58:41): Now, if you were to take out nine-and-a-half gallons of fresh water and put back half a gallon of supersaturated brine, even that would be fine if you mixed it in, but there's no need to. So that's a general plan and just do that at massive scale and at the lowest possible cost. So I believe that it should be possible to produce water, desalinated water for 99 bucks an acre foot, which would be transformational. That means it's no longer like, oh, well if we're really short on water, we'll turn on the desal system and you can flush your toilets and drink water. It's like the rain came in a bit weak this year in the Sierras. That's fine. California Central Valley agriculture will be fine. We need about 14 or 15 million more acre feet to saturate demand. That's about double what we currently have. Bring all this additional agricultural production online, reduce costs, improve the ecological impact of farming on places that are short on water anyway and pay for itself. John (00:59:29): Can you do stuff in the delta where it is brackish without running into the California Coastal Commission? Casey (00:59:35): I honestly don't know. That could be tricky because— John (00:59:37): I know they’re sensitive about the delta and the hydrology there— Casey (00:59:40): Yeah. You might want to run a line for instance along the floor of the bay all the way out to the coastal current. Where you put the— John (00:59:47): You’d be increasing the salinity of the— Casey (00:59:49): Something like that. Yeah. The best way to pay for this actually is land value appreciation. So my hobby horse is the Salton Sea, which is pretty relevant to your Northern California liberal elite crowd, but us true Southern Californians. The Salton Sea was formed as a result of an irrigation accident in the early 1900s, and it's about the same size as the entire San Francisco Bay. It has 110 miles of coastline and you could not give it away. It's owned by a patchwork— John (01:00:12): The Salton Sea. Is there water there right now? Casey (01:00:15): Yes. John (01:00:16): Okay. Casey (01:00:16): It's like three and a half million feet in there. It's too saline for fish to survive now. Yes. About 15 years ago, the last fish died. It was pretty smelly. But when it first formed and through the sixties it was actually pretty much fresh and then it got saltier over time. John (01:00:29): How did it get saltier? Casey (01:00:30): Evaporation. John (01:00:32): But how was there water there to begin with, to evaporate? What was the manmade intervention? Casey (01:00:37): Irrigation runoff. So a lot of water is diverted out of the Colorado River through the All-American Canal about one point— John (01:00:43): And it's accumulated— Casey (01:00:44): 1.2 million acre-feet a year. And then most of that is used in irrigation and some of it runs off into the Salton Sea. John (01:00:51): I see. Casey (01:00:53): It originally formed when the Colorado River burst its banks in 1905 or something like that. And it flowed down there for a couple of years before the railroad was like, “Enough of that.” They cut it off. But yeah, I think it's quite a starkly beautiful place. John (01:01:06): And how does the salt get into the Salton Sea in the first place? Casey (01:01:09): The Colorado River is quite saline itself. John (01:01:12): Why? How? Casey (01:01:12): Mexico sued the United States for the Colorado River being too salty. John (01:01:14): Do I not understand how rivers work? Casey (01:01:16): Yeah, because they erode rocks, and rocks have salts in them. John (01:01:18): Ah, okay. Casey (01:01:19): That's why the ocean is salty. John (01:01:21): Really? Casey (01:01:22): Yes. John (01:01:22): That's where the salt comes from? Casey (01:01:23): Yeah, from dissolved rocks. Why the salinity of the ocean is so stable is another question. But my grand plan is to put another 3 million acre-feet of water in the Salton Sea, bring it back up to its historical 1990-era level. Stabilize the level, stabilize the salinity, build a canal that comes in from either Mexico or maybe under the mountains from maybe both, maybe from San Diego. Put some whales and stuff in there, like SeaWorld but bigger, and start building gigantic hyperdrive cities around the edge to do industrialization and retirees and a bit of everything. 110 miles of beautiful coastline with a climate like Phoenix and Tahoe, but warm. And we could do that and it would pay for itself because you just didn't run the math on 110 miles of coastline on a nice river. John (01:02:08): Have you pitched the Trump administration on this? This feels like the kind of thing they’d be very into. Casey (01:02:11): I did write a one-pager for the Secretary of the Interior, but I think they've got a lot on their plate. But actually, the United States has an agency called Bureau of Reclamation, which is part of the Department of the Interior whose job it is to figure stuff like this out. And it had one extremely powerful, charismatic leader who basically led all the big projects. If you've read Cadillac Desert, it goes into some detail here. John (01:02:28): That's a great book. Casey (01:02:29): And then—that's generous. John (1:02:30): You didn’t like it? Casey (1:02:31): No, it’s so scold-y. It’s like, “Humans should not live in the west. It's too dry.” I'm like— John (01:02:36): But if you're judging, there's so many… The Kaiser book was bad in a way, but there's really good takeaways from it. Casey (01:02:44): Which one? One of them was written by this Mormon fellow who was his 2IC for many, many years at the Kaiser-Frazer car company, amongst other things. And the other one was some independent researcher. John (01:02:52): I don’t know— we’ll put them in the show notes. But— Casey (01:02:54): They're both fine. They're both great books. John (01:02:56): But they're great because of the subject matter. Casey (01:02:58): Some of them are a bit long-winded. John (01:02:59): Yeah, you have to be able to read past all this stuff. Casey (01:03:01): But, the Bureau of Reclamation hasn't built a major dam in 50 years, so they've pretty much forgotten how at this point, I think. John (01:03:05): But your Salton Sea development, part of the reason it feels plausible is because there's no conservation argument for keeping it the way it is, because only conserving prior human errors. Casey (01:03:14): Well, you’d make it much better. You could actually make it have— John (01:03:16): No, but my point is, even if you wanted to develop the California coast and you say, “We'll make it much better.” People will say, “No, sorry, that's not how it is. And we need to keep it in a state of nature.” But the Salton Sea is not in a state of nature, it’s in the state of altered effluence. Casey (01:03:30): It's in a state of neglect, really. Profound neglect. And various studies have happened over the years. They want to spend $100 million dollars there, $200 million there to try and stabilize it to an extent. But I think that the mistake is assuming that it's always got to be this festering pustule environmental catastrophe—it doesn't. We could restore it and then unlock enormous economic value for people who need a place to live and could live around it. And build it into a new city in the United States—hasn't been built for a long time and they can be so cool. I'd go there, build that city, let's do it. And essentially infinite solar power, lots of critical minerals just like bubbling out of the ground, very close to the Mexican border, which has its advantages obviously. And you can produce plenty of fresh water for Mexicans as well. Part of a deal, everyone wins. Something like 50 to one—just back of the envelope calculation—is like 50 to one economic leverage. It would generate so much value that you'd actually… The way you would do it is you say, “Well, whoever comes and develops all the desalination technology could be asked, could be a consortium, could be the Israelis, whatever. We'll give you some fraction of the coastline and then you get to either sell that off or keep it and hold that money.” A bit like, I think Irvine did it that way and that's how you fund it, right? I don't want to box myself in here, but I think you could give 10% of that coastline away to the developing entity and 90% could be to the people of California in general, and you would still be making money, hand over fist. That's pretty unusual. You don't ordinarily find mega projects like this where there's so much value on the table that you can be that generous with it. John (01:04:59): Part of what feels attractive to me about desalination for you is, currently Terraform sells commodities, be they natural gas or methanol or things like that. Whereas someone who wants water in a location, they need to buy water in that location, which has a price. Getting water to the city of San Francisco or getting water to Los Angeles is not a commodity. It's a very specific project. Getting water across the delta in this specific context costs $20 billion with that capital project that you were mentioning. And so you presumably, even at the early stages of the project, even when it's more expensive than it will be at large volumes, you can find somewhere where they really value having an incremental water supply and they're willing to pay for it. Casey (01:05:48): Yeah, I mean the value to be unlocked by bringing even small amounts of water onto this land is enormous. So again, why is Nevada 90% federal land? Couldn't give it away. Could you pump significant quantities of water into Nevada? Yes, of course you could. It's actually downhill like three quarters of the way. John (01:06:05): Sorry, three quarters of the way to Nevada? Casey (01:06:07): Yeah. So for example, there's already an existing aqueduct that runs from the delta here all the way down to the Cajon Pass, which is the pass that takes you from LA up to Vegas. And from there, if you were to release—so you need to upgrade that canal to carry a little more water— you could release water and flow downhill down the Mojave River down towards Death Valley. You can divert it at the Amargosa Dunes and then pump it up the Amargosa Valley there. And then basically— John (01:06:32): I see you're doing very little pumping. And it’s not net downhill most of the way— Casey (01:06:36): It's not downhill most of the way. The pumping is actually not that expensive, because solar array, pump, it's relatively straightforward. John (01:06:42): I'm sensing a pattern here. Casey (01:06:43): It's like anything that you want from SpaceX, it's like Starship will do it for you. And anything you want from Terraform, it's like just solar panels plus other stuff. And then you say, well, Nevada is a hundred valleys and we've got cities and a few of them and Burning Man in one of them, and you've got nuclear testing in one of them and Area 51 in one of them. And that's fine. And then most of them can stay as desert and maybe half a dozen of them turn into solar for synthetic fuel and half a dozen of them, you run a river the size of the Colorado River through them with a few pumping stations here and there. But mostly natural flows down into the Humboldt River, which will now become a perennial stream down into the Carson Sink where you scoop it up and then pump it back to the beginning or drop it into Lake Mead, past Vegas, where it can find its way back out to the ocean. So not very much like scribbling on a map is required to make this work. Nevada doesn't even have a significant river at all. John (01:07:29): You could be the first. Casey (01:07:30): It could be the first, but it's an entire state worth of land that would have the climate of Switzerland if it just had water. John (01:07:35): Do the Israelis do a lot of desal? Casey (01:07:37): Yeah, reverse osmosis was largely developed there. Actually, I think they just announced that they're using coastal desalination in the Mediterranean to refill the Sea of Galilee, which is in essence the reservoir that they use to feed downstream irrigation on the Jordan Valley. John (01:07:52): Why not produce your Terraform desal edition and sell it to the Israelis for— Casey (01:07:59): They're already very good at it. John (01:08:01): But this wouldn't be cost-competitive or it wouldn't be anything competitive? Casey (01:08:03): I'm honestly not sure. I think they can produce at 400 bucks per acre-foot. And to be honest, if I was able to produce a 400 bucks per acre-foot, I'd be very happy right now. Long term, $99. At $99, I think I can steamroll the California Coast Commission if I need to, but even $400 would be pretty extraordinary. John (01:08:20): What have you learned about leadership running Terraform? Casey (01:08:24): That's a good question. I went to a high school where there was a leadership program in a way, and it was kind of a Judeo-Christian based type thing. And so this idea of leadership through service was taught, which I think is super valuable and something I've always thought about. And our org chart is drawn upside down because ultimately that's the way the—can I swear on this podcast? Yeah, the problems flow downhill. Yeah, exactly. So that's an important one. I think that I've made plenty of mistakes at this point, but having made those mistakes, I can now read Leslie Groves’ work about the Manhattan Project or Rickover’s work about the nuclear Navy project and interpret those texts and understand what they really mean or have a better insight into the sorts of problems they're actually dealing with because they don't necessarily translate that well into a narrative form. Just learning to hold myself accountable, hold other people accountable about how to titrate how demanding you're being versus other things about how to understand what people want rather than what they say they want and then make sure they get what they need. Stuff like that. John (01:09:27): Dial up your demandingness? Dial it down? Casey (01:09:29): Dial it up. John (01:09:30): Okay. Casey (01:09:30): Dial it up. Yeah, absolutely. John (01:09:32): And is it about finding constructive forms of it that lead to better performance? Casey (01:09:37): Yes, it's a balance. One of the mistakes I've seen a lot of first-time founders make is they assume that an hour of coaching is as good as an hour of firing. Because at the end of the day, you can't be some other founder, you can only be yourself. And so you want to build an organization that essentially is like an extended mind for you and for your hands and arms and the ways that you think about a problem. And I think Stripe is very much a reflection of you and Patrick. I'm not sure which one the most, actually. I have a question for you about that next up. And so to an extent, the parts of Terraform that weren't working with the parts that I was trying to build, not my way against my better judgment. And when I just gave up on that and said, “Screw it, I'll just do it my way.” It suddenly started working becauseI could understand and intuit how it worked.Yeah, had to become pretty demanding there. And obviously no one's perfect at recruiting. And so the success to that is you've got to give people extremely direct feedback and coaching and mentoring and ultimately invite them to build their career elsewhere. It's been really tough. Some of the people I had to let go, people I really like, people I’m friends with, for many years in some cases. And obviously I wish them the best. But again, having read resume after resume after resume, nothing is worse for your career than sitting in a job failing for you because you don't have the guts to quit and your manager won't fire you. And I've just seen this so many times because it's not that you're not growing, it's that you're growing backwards in that time. For a couple of years, success is mandatory. I don't really care that much about how you go about succeeding. I'm happy to help you, but at the end of the day, you have to succeed. We're going to force you to succeed. Success is mandatory, no matter what you're doing. That did not come naturally to me, but it's very intuitively obvious. So my question for you is, in the counterfactual where you and Patrick founded YC companies together but they weren't Stripe, they were two separate companies, what would the world now look like? John (01:11:24): Oh, I have no idea. Casey (01:11:26): Would he have founded Stripe or would you have founded Stripe and then the other one do something different? John (01:11:30): Stripe was much more his idea than mine. Casey (01:11:32): Would you be in the United States? John (01:11:34): Plausibly not. He came here first. I mean, we both came here for college. Casey (01:11:37): What were you studying in school? John (01:11:39): I did not stick around long enough to declare a major, but— Casey (01:11:42): You can make something up. John (01:11:43): Yeah— Casey (01:11:44): Physics. You would've done physics. John (01:11:45): I would have done physics. Casey (01:11:46): MIT or Harvard? John I was at Harvard. Casey Harvard. Maybe you would've wound up at Draper Lab or something. John (01:11:50): Or a JPL, you know? Casey (01:11:51): Yeah, I wouldn't curse you. There's a sad story. John (01:11:58): Is there any hope for it? Casey (01:12:00): I would like to think so, but much to my consternation, people who I knew who were there, who I always thought were like JPL boosters, were like, “We're going to succeed no matter what.” And now it's beyond hope. And I'm like, “Well, if you think it's beyond hope, then that's a bit dismal,” which is a real problem because it is really the jewel of space science in the United States and much of our technology besides, but it has been allowed to wither. There’s a lesson there. John (01:12:26): Do we need more private research organizations in this area, where obviously there's many tech companies building research organizations model after the Bell Labs type thing, Xerox Park, Microsoft Research, many parts of Google. Do we need that in this line of business? Casey (01:12:44): I think a better model, I don’t know how to get there from here, but a better model might be that every time a lab or an agency or something is set up, it's automatically set up with its destructor, right? And after some fixed period it sunsets. John (01:12:59): People always say this. I don't know about sunsetting regulations. Casey (01:13:02): It's not like you fire everyone, but you have to reconstitute it and you have an opportunity to do a reorg. So after 10 years, say NASA gets summarily executed and then either you bring up another copy of it, and this time maybe you sell off two of the labs to DOE or you buy up another lab that you wanted but you couldn't because you had to support these two other labs that weren't doing things that were useful for your current mission anymore. John (01:13:25): But people probably don't want the offshoots. Like, does DOE want to take on the budgetary expense of the offshoots? Casey (01:13:31): Well, that's kind of part of the problem because, well, I mean the idea is any existing agency could bid for it. And actually the Antarctic stations, for example, are kind of run this way in that the contractor who runs them is cycled from time to time. But the alternative is you have the core of some function that is ostensibly important enough to US national security and soft power and so on, that it has to be government funded. Screw Milton Friedman. It has to be nationally funded. But if its mission is that important, then it must be, you must insist that it continue to succeed. And we have this conversation every few years about the Air Force and the nuclear missiles and someone finds something's gone wrong and it's like, “We'd better make sure this works!” But that kind of applies across the board. And if you don't do that, then what happens is it just kind of withers and festers and then when you need it, it's not there. And when we relaxed the Outer Space Treaty in 2017 with the outer space accord, so the Artemis Accords, so we can go and make a land grab on the moon. And then it turns out China's going to get there first. Oops. Maybe we shouldn't have done that if we weren't sure we could win that one. Why can't we win it? Because NASA, the organization whose entire job is like, no more Sputnik moments, please, has kind of lost the plot because it wasn't forced to reconstitute itself and so on and so forth. And you say the same thing here, I hate to say it, but in Silicon Valley most of the companies don't last that long or they get acquired or spun off. And the ones that do, in many cases, you'd really rather they hadn’t. John (01:14:55): The IBM problem. Casey (01:14:56): Well, IBM, it's still around, but there are plenty of other companies that have become quite large and successful and in some ways victims that are in success. And I think, again, one of the anomalous things about Elon's companies is that they're now quite old and quite large and they've got other organizational pathologies, but they haven't had that problem. People should figure out how and why. John (01:15:15): Last question. What's the shirt? Casey (01:15:18): This is the Hyper America shirt, which I wore today because it was kind of a theme with if people like you don't strike out for the hills and try and do insanely ambitious hardware projects, it's just like Elon and a few other wannabes. We're not going to get Hyper America with a trillion Americans and a trillion planets. John (01:15:37): Maybe Stripe should take on a hardware line of business. If we find one that's good enough, we wouldn't be opposed to it. I think we just haven't had a good idea yet. Desal is taken. Casey (01:15:45): Stripe banner? No, I mean I welcome competition. I'm kind of sad actually that at Terraform it's been four years and there's only one other company that has come in, which of course, and there should be many more. And I think actually it would help us if there was competition, but it is really hard. I can understand why people look at it and me like, “No thank you.” John (1:16:02): On that note, thank you. Casey (1:16:03): Thank you.