America's aerospace rebirth

Speaker A: You don't try to bring back what left China. What you do is invent the generation after it and have that be here with Supersonic. There will be similar stories. We just don't know what they are yet. Imagine if you could get anywhere in the world in four hours for $100. But think of how differently you live your life. The fact that we're building our own engine changed the seats on the airplane. Uh, how?

Speaker B: In 1947, Chuck Yeager climbed into a Bell X1 over the Mojave Desert and broke the sound barrier for the first time in human history. It was a classified government program backed by the Air Force, unknown to almost everyone.

Speaker C: When it happened on January 28, 2025, in the same airspace, a startup called Boom Supersonic did it again. No government commission, no nation state behind it. The man who made that happen grew up watching Cessnas take off from a suburban Cincinnati airport. He dropped out of high school early. He taught himself aerospace engineering from textbooks and started the whole thing in his basement in Denver with 10 people.

Speaker B: What supersonic flight reveals about what's possible, what technology can do when ambition is commercially disciplined. What AI unlocks when applied to the hardest physical problems, and what it means when a founder in a basement can do what used to require a superpower. That's what we're here to talk about today.

Speaker C: Blake also runs the same diagnostic on everything around him. Airports, roads, hotel, check in, regulatory paperwork that he ran on the aviation industry. We want to get into how he actually thinks, not just what he's built.

Speaker B: Blake Shull, welcome to Possible Blake. Great to have you here. I'm extremely proud to have backed Boom, um, and to have, you know, a front row seat to this journey. Most people ask you about January 28th, the first time that, you know, XB1, your supersonic demonstrator, broke the sound barrier. But the moment I come Back to is February 10th, the final flight of the, you know, demonstrator. Um, you got choked up on the live stream, saying goodbye. This plane carried a decade of your life. What does it mean to close this chapter of your life and move to the next?

Speaker A: It was definitely a bittersweet moment. Is that there is, um. Man, I might tear up again. Uh, there is nothing more exciting than to pour a huge piece of your life energy into something and then see it actually work. And there is just so much drama in flight test. Um, uh, just the intricate teamwork, the excitement of it, the safety criticality of it. Um, uh, and knowing that that thrill had come to an end was very Sad. It was still the right decision not to fly the airplane again. Um, but in the end, it gave me all that much more commitment to build Overture. It's like, this is not the end of the story. This is actually, uh, this is actually getting to base camp, and we're going to take a very brief rest and then we're going to climb again. Uh, but it's a low base camp. Actually, the hardest work is still ahead.

Speaker C: Can you, for those of our listeners who might not know, can you just say a beat more on why was it the right decision for this plane not to fly again?

Speaker A: So let's put the why in the heck do we even build this airplane? Sometimes people say, uh, XB1 looks very little like our Overture production airplane. So why did we even build it? Wasn't that a waste? We've gone straight to production. And I think the actual reverse is true. If our prototype looked exactly like our production airplane, it would imply that we learned nothing. And, uh, the whole point of doing XB1 was to learn 100% of what it takes to build a supersonic jet safe enough to put a human on board. And we knew that there were mistakes, and the only way to find them was to go, design, build, fly, learn. And so, uh, across 13 flights of the airplane, um, we did 100% of, actually, 200% of what we intended to do was a story we can come back to. We solved Sonic Boom, actually without trying. That was an accidental discovery. But every flight of an experimental airplane carries risk. And this was an airplane that was not designed to be as safe as an airline or it was designed to be safe enough for a flight test program. So you never know how close you are to a mishap. From just a safety perspective. You don't do more than you actually have to do. The other piece is, um, it took a team to keep the airplane alive. Um, it was roughly 50 people full time on the program. Uh, every flight requires a control room. Like think NASA mission control. Um, uh, and it was costing something like $1.4 million a month just to keep the program alive. Uh, and those were also all of our best people. So we wanted to. One, they wanted to get out of the Mojave Desert as well. Uh, but we wanted to go put them on the next. And we could have paraded XB1 around the country very expensively with, uh, some risk of a mishap, and then also just delayed the actual supersonic future. So it just was not for a bunch of different reasons. It wasn't the right call to keep flying.

Speaker C: So we're going to dig into all of those details, but let's zoom out for a second. Yes, this is all about flight for you. It seems like you've been watching Cessnas at the airport with your parents since you were a kid. Why has Flight Light captured your imagination?

Speaker A: I think we don't really get to control our passions. I think they kind of choose us. And, um, I remember being fascinated with airplanes as long as I can remember. I remember as a kid just sketching things I'd love to build, but it never occurred to me that I'd actually touch it professionally. Um, I think because I was very drawn to things that were, uh, innovative. And um, a Cessna is like 100-year-old design almost at this point. And what was Boeing doing as I was growing up? Well, they were redesigning, um, yet another subsonic airliner that looks pretty much like the last subsonic airliner, which also looks pretty much like the one that came before that. And so there's nothing going on. Uh, and so I went to school for computer science. Um, I sort of built my first company in my parents basement in high school.

Speaker B: Uh,

Speaker A: it feels too much to call it that. We managed to host people's websites out of my basement. Uh, it broke even. But I loved flight. Started flying for fun when, uh, I was in college. And there was a key moment in my mid-20s, um, I was in Seattle out of my time at Amazon, and they have a Concorde there, uh, in the flight museum. And I remember looking at that and thinking, how is it that the most amazing airliner ever made is, uh, in a museum when, uh, there's nothing better in our skies? And, uh, that day I set a lifetime goal of breaking the sound barrier, put a Google News alert on supersonic jet. So I'm like, I just want to be first to know when I can go buy a ticket. So I thought I would accomplish the goal by means of buying a ticket, not by building the jet. Another 15 years kind of went by, um, uh, and I had my first career in tech. I sort of built a, um, baby startup that got Acqua hired to Groupon. And, uh, I'm like going to work every day, basically running the world's largest scale spam system and thinking, is this why I'm on Earth? Uh, and so I sort of fired myself from Groupon and I knew I wanted to do another startup, uh, but I wanted to do one where no matter how hard it got, I would never ask was it worth it? Um, and so, uh, my most ambitious ideas went to the top of the list. And I thought, okay, I will need to understand why supersonic is a bad idea. I'm sure it is. Because if it was a good idea, everyone says in the Valley, if it's a good idea, there's already a team on it. Right. And it was one of the worst pieces of advice because it causes some of the most important problems to be left untouched. It creates a bystander effect. Uh, uh, but I thought I would get too excited the research and move past it. And instead I just couldn't find the reason it was impossible. So, um, I felt like I had to go do it.

Speaker B: Well, and, you know, obviously the title of the podcast is possible, and it's been amazing for you to demonstrate the possibility. One of the things I think it's important to start at is people tend to think, oh, it's about building a plane. But actually, in fact, you're not just building a plane. You're reconstructing an entire industrial base, um, that is basically atrophied. So what does it mean to think at the industry level and not just the product level, and also might be a good time to demonstrate.

Speaker A: Yes. Um, let me tell a little bit the backdrop to this. When we started, it didn't occur to me how much of the industry would have to be reinvented to do this. America's aerospace industry is incredibly broken. Um, not only has it all calcified down to the Boeing Airbus duopoly that, uh, basically stopped shipping new products. I don't know if people appreciate that the last time Boeing introduced a new airliner was the 787 in 2004. In tech, we're like, this year's iPhone is too much like last year's iPhone. Right. This year's airplane is more than 20 years old. Um, the latest and greatest. It's absolutely nuts. And worse than that, so much of the industrial base behind aviation, um, is really driven by defense and, uh, the, you know. So, uh, a thing I would start to notice as we started to build the engine, we just kind of went to the first, um, we went to like, the default aerospace supply chain. We started getting these quotes, like, nine months for a turbine blade that was 3D printed. Like, why does it print for nine months? Like, why does it take nine months? Well, actually, there are a few different operations. You have to wait your turn on the printer and then it has to go to like, another factory where they inspect it, and then it has to. Which is in a different state. Uh, and then it has to go to another factory where they braze it, another state where they, uh, polish it. And it's just like each production process is in its own different location. And the thing spends more time on a truck than it does on a machine. And I'm like, why in the world would there be one factor in every congressional. Oh, right. It's like what we have is a congressionally optimized supply chain. So programs like the F35 actually drive the industrial architecture. And what it means is we have one supplier for every little bit of thing in every congressional district. And parts bop around. That way you get jobs everywhere. That's where you get votes everywhere. And all of these contracts tend to be cost plus, uh, contracts. And what that means is if your margin is basically fixed, um, then what you want to do is just jack up the price, because that's the only way to maximize profit, is just increase the cost. And we have this thing we call slacker index, which is how long it takes to get a thing divided by how long it actually takes to make it. As an example, the thing that's currently driving the schedule for our engine, the low pressure, um, turbine shaft, which must be inertially welded. And the quote we got out of the traditional aerospace supply chain is 88 weeks for inertial welding. Do you know how long it takes to do inertial welding?

Speaker B: I would guess three hours, seconds.

Speaker A: Like, you spin this thing up, you put the parts next to each other, and they bond. Literally, this is one of the fastest things you can do. But it takes 88 weeks to get a supplier to actually do it. Um, and so what we realized pretty quickly is, uh, we needed to just take all that manufacturing in house. And, uh, in particular, it was important to put every production process, starting from raw materials, under one roof in the same building where the engineers work. And in the industry, this is, uh, like I tell people this, to me it seems like the most obvious thing you would do. Um, you wouldn't, like, spread this out and, like, have all these trucks. But like I tell people we're going to go from raw materials to completed engines in one building, and people are like, whoa, like, what an insane idea. Uh, but, uh, it allows us to move faster and do things that otherwise would just not be possible.

Speaker B: Yeah, because it's actually functioning as a tech startup versus a defunct industry.

Speaker A: Right.

Speaker B: So, um, you brought a couple of parts that you have been, uh, manufacturing. You want to care to explain what they are and the relevance?

Speaker A: Yeah, absolutely. So, uh, a Year ago we had no factory, uh, we had no manufacturing team. We had nothing. And we had to go build the team, build the factory and build the know how. And uh, we basically have been just climbing up the learning curve doing on our own, progressively more difficult parts. Uh, uh, so these are uh, engine blades. Um, and these are amongst the most difficult to manufacture things, uh, in the world. And we just started with the easiest one. So this is a compressor, uh, stator vane. Um, and uh, it looks like a very funny shape, but uh, actually those contours are very, very important. And this must be accurate to within about a thousandth of an inch on that leading profile. So the smallest ones are the easiest ones to make. And then we went kind of up the difficulty curve. Uh, this is a stage two variable guide vane. So this sits um, between the first stage compressor and the second stage compressor. Um, there are uh, something like 60 of these kind of circumferentially around the engine and they actually rotate to redirect the airflow. So depending on the state of the engine, these blades actually pivot. So there's a thing outside the engine that grabs onto this and it twists. This also has to be accurate, uh, within a thou. And then the hardest thing is these two nubs have to have uh, what's called high concentricity, meaning they have to be aligned with each other. Otherwise when you put the part in the engine, it'll jam. And these are manufactured on a um, computer controlled milling machine, starting with a gigantic block of stainless steel. And the thing that's tricky is you're basically manufacturing these in a cantilevered fashion. 1 this machine is held and there's a little drill head over here. And um, it will very easily deform. So you have to get really good at reducing the internal stress in the material along the way and then controlling the manufacturing process such that uh, you don't end up slightly out of tolerance from here to here. This is the next hardest thing. And then this is a high pressure turbine blade. So this is the kind of our first generation high pressure turbine blade. So this goes to the back of the engine. Um, again there's um, uh, something like 60 of these circumferentially around the engine. And this is uh, it sits right after the combustor. So the hottest air comes in and basically hits this blade, causes the whole thing to spin. And that's how you get energy out of the engine. And, and what's kind of neat is these blades operate in an environment, um, actually hotter than starship. On reentry and it's actually above the melting point of the metal. And so the only way to have the turbine not melt is these blades are hollow on the inside. So there's these cooling holes at the bottom and there are 248 individually cut holes in the blade. So air comes in the blade and, and out all these holes and that blade actually gets enveloped in a layer of cooling air and the cooling air is nice and cool. It's only 1,000 degrees. Um, and so this is our first generation, uh, high pressure turbine blade. This was built out of a nickel super alloy on a 3D print process. And we brought that in house because we didn't want to wait nine months. And we can print one engine's worth of these blades in about three days.

Speaker B: That's awesome.

Speaker A: Right? And by the way, when you do that, it completely changes the way engineering works. So the thing that we've really learned is what you got to do is enable iteration both in the design world digitally and in the manufacturing world. So uh, we've done a lot of work with digital, uh, design tools to allow engineers to allow small numbers of engineers to go through design cycles much faster. And then if you 3D print a blade in a few days also you can iterate physically much faster. What otherwise happens if you get these out of the supply chain, it's called 912 months. The engineers can't try anything. Right. They have to get it right the first time or else you've got like a nine month setback.

Speaker B: Yes.

Speaker A: Uh, so you get a lot of conservatism, you get a lot of analysis, paralysis, etc. Now people who um, have been following Turbine Twitter will point out that the 3D printed blades are not going to give you the strength and life you need, uh, for a production engine. And that's true. There's a technology called single crystal that only uh, three companies in the world, again according to Turbine Twitter, can do. And so I can't talk more about single crystal yet, but we'll have something to share later on.

Speaker B: That's awesome.

Speaker C: Amazing.

Speaker B: It's just back to the earlier thing. Although this was like we're geeks and nerds, so we love this stuff. Um, um, when did you kind of realize that the challenge wouldn't be capital, wouldn't be regulatory? Both serious, but a knowledge and capability.

Speaker A: Yeah, um, reality keeps teaching the lesson until the student learns it.

Speaker B: Yes.

Speaker A: Uh, but, uh, the least difficult thing in all of this is the incredibly difficult engineering. Um, the next even more difficult thing is Financing, this is a pre revenue company that requires a lot of capital for a long time. Uh, but the actual most difficult thing is building the team and building the culture. And the reason is that this industry has had no entrepreneurship for an extremely long time. Prior to boom, the last new commercial aircraft startup, uh, founded by an entrepreneur, it was Douglas Aircraft, 1921. Right. So we went from 1919, 21, otherwise

Speaker B: known as McDonnell Douglas.

Speaker A: Right? Yeah, it didn't, didn't go so great in the end. Um, uh, but in its heyday Douglas was great.

Speaker B: Yes.

Speaker A: Uh, but yeah, so from 1921 to 2014 there were no entrepreneur funded commercial aircraft startups. And so, and worse, what's happened in the industry? Um, it goes back to my own story. As a kid I loved airplanes, but I wasn't going to go work at Boeing where I could redesign the same thing over and over again. I was going to go to Amazon where I could do something new. Um, there's an entire generation of kids that did not go into this industry. Uh, so not only is there nobody familiar with what the heck a startup would look like, um, uh, a bunch of the best and brightest didn't go into the field. And so it's extremely difficult to find, find senior talent. And basically everybody we hired out of the industry at a senior level failed. Every single executive was a failure. Some of them just abysmal failures. I caught people literally cooking the books on engineering data.

Speaker C: Right.

Speaker A: This industry is not just calcified in many places, it's corrupt. That's how you end up with stuff like the 737 Max. Um, and so we had to learn how to mint the team and ultimately came up with a concept we now call the talent distillery. And the idea of the talent distillery is you get some young spirits and you get just a little bit of oak. Uh, and then the oak will infuse into the spirits. And if it's the right oak, uh, actually the spirits will also infuse back into the oak and then the oak is kind of there to keep the spirits from just kind of like going everywhere. They have to contain them somehow. And so we've just tried to accept, accelerate development of our own talent in house that way. And the non obvious thing is, uh, you don't put the oak in the organizationally senior position. The oak can also just be advisory. Uh, so our rule is, I'll tell the story of how we kind of discovered this. In the early days we were all spirit, no oak. And so we had a very young, bright team that built the landing gear on XB1. And they had never done a landing gear before. Um, they were sufficiently self confident. They also didn't call anybody who had ever built a landing gear before. And so they made a whole bunch of first timer mistakes. And we ended up having to completely rip and replace the landing gear. Iterative engineering is good, but we probably could have started in a much better place than where we actually were. Um, uh, so the lesson we learned is it's totally fine to do something you haven't done before. Uh, but if somebody in the world has done it before, the requirement is at least to call them. You don't even have to take their advice. You just have to hear their advice. Uh, and so we've made that a rule in the culture and it's gone super well.

Speaker B: Yeah. And the proof is in the results.

Speaker A: Yes, absolutely.

Speaker C: Well, I'm about to ask you about one of those results because like you said, you need the right team, you need the right culture. It's not just the young spirits. It's like how does that, uh, interact with the world? So you, your innovation was boomless cruise, and it's a striking example because it was flying supersonic without an audible ground level boom. And that turned out not to be a hardware problem. That was a software problem. It was the right weather data and an algorithm that tells you what speed and altitude to fly. And NASA had spent decades trying to engineer the boom away, and you figured out you could just root around it. And so first of all, tell us more about that. And is that a pattern dissolving these

Speaker A: hard physical limits with AI Uh, it's absolutely a pattern. There's all kinds of interesting things we can extract out of this. But I should tell the story because.

Speaker C: Please do.

Speaker A: Uh, it is a non obvious story. So this was not the plan. So Reid will remember the original boom pitch was you don't have to solve sonic boom. Uh, you can just fly supersonic over water, subsonic over land. The longest, most painful flights are over water. Anyway, uh, a sonic boom solution is not part of minimum viable product. And so we kind of ignored it. Uh, we knew along the way that there were these physics called Mach cutoff. And what mach cutoff means really is if you break the sound barrier at a sufficiently high altitude, um, and you're controlling speed, um, uh, which basically lets you control the angle at which the boom comes off the airplane. The boom will refract as it goes through the atmosphere and make a big U turn in the sky and never touch the ground. And this Was this was, you know, we didn't invent this physics. Like we didn't invent the idea of applying it to a supersonic jet. It was all out there. But at the time we had made the, um, terrible decision to outsource our engine to Rolls Royce. Uh, uh, this was one of the most awful things I ever tried to do. Uh, and that the, all Rolls Royce was like, maybe willing to do is offer us like a warmed over, modified subsonic engine. And, um, it didn't have the performance to break the sound barrier to sufficiently high altitude. Uh, you need what's called good transonic thrust drag margin. Uh, so right around the speed of sound is the hardest place to fly. Uh, and so if you don't have a good airframe and a good engine, uh, you have to break the. You have to punch through lower. And the key threshold is that you have to be somewhere above 30,000ft in order to have enough refractive power in the atmosphere that the boom goes away. And so we looked at it and we're like, can we get roles to change the engine? We're like, that's not going to happen. We'll be lucky if they even make us an engine, uh, let alone customize it. So we kind of put the idea in the back shelf and forgot about it. And now it's, um, late December, early January, uh, 2024. And, uh, we're gearing up for the first supersonic flight aboard XB1. And the world believes, no matter how many times I say that you don't have to solve sonic boom, the world believes that solving sonic boom is incredible, incredibly critical for minimum viable product. And they believe that we called the company Boom, so therefore we must have solved it. Uh, and so I was like, okay, guys, I need a sonic boom forecast for this flight. Because if it's going to be a real wallop, I need to set everyone's expectations. Also, I kind of want to know where to stand. Having spent 10 years on this thing. I want to hear the sonic boom. Uh, and the team comes back and they're like, no, man, XB1 has great transonic performance. It's going to be up 30,000ft. There will be no boom. And at first I was disappointed. I was like, darn. I really, really wanted to hear the sonic boom. I wanted to call the flight the boom heard around the world. That was like the marketing idea. And I was like, okay, shoot, it's going to be boomless. All right, fine. Uh, and the team had the insight to have our friends at NASA turn on the microphone away so that we could actually confirm this really happened. So we go and we do the flight, and sure enough, three times through the sound barrier. Look, Ma. No. Boom. And the Internet is rejoicing. Twitter blew up. Uh, and everyone's like, boom. Solve the sonic boom. And I'm, like, on Twitter, trying to convince people that we haven't. Um, and it occurs to me, like, wait a minute. Uh, we had our famous breakup with Rolls Royce. Uh, we're doing our own engine now. But when we're doing our own engine, we can customize it. And a thing I keep hearing as we go through these engineering status meetings is, oh, the transonic performance just keeps getting better. And that means the acceleration altitude can increase. And every time we're in one of these reviews, it's like, oh, there's another thousand feet. Feet. There's another thousand feet. And I was like, what if we actually crossed the threshold of viability without realizing it? And so I called the team together on a Saturday, and I was like, all right, guys, we got to rerun the math. Can our own engine in production do Boomless Cruise? Or if it can't, is there, like, a small gap and we could actually choose to go after it? And it took about 90 minutes for the team to run the simulations. We've, uh, gotten very, very good at running simulations quickly. And the answer came back, no, man, it just works.

Speaker B: We got it already.

Speaker A: Someday we'll write a book, and we'll call it how to Solve Sonic Boom without really trying M. And so I called the marketing team. I was like, guys, we're going to relaunch the entire company around Boomless Cruise. And at first, my team pushed back. They're like, boomless is off brand for boom. What are you talking about? Everyone's going to love this. And so we did it. In the second supersonic flight, we announced it was going to be Boomless once again. We did three times through this island barrier. Once again, nobody heard anything. And so there we are.

Speaker C: That's awesome. And I mean, it's certainly key. Like, who are the people building this? And you've talked about sort of the types of people you're bringing in, but you've also talked about the size of teams and how that with, like, small, creative teams, you do not have to be an aerospace company with thousands who are doing this. Like, talk about what that looks like.

Speaker A: Yeah. So, um, communication overhead increases exponentially with team size. Right. I think this is. This is why when you zoom in into really historic efforts, you Often find it's like three people that did the key thing. Then you look at these multi, uh, tens of thousands or hundreds of thousands of size companies and they don't get nearly as much done as you would think. Uh, sometimes they get very, very little done. So there's a real magic in small teams. Uh, yet if you do things the classical way, these hardware companies require enormous teams. And what we've found is if we really embrace software enabled design, we can enable small teams to do, uh, big things. So I'll give an example. So if you're designing one of these blades, so when they operate, they're very hot, right? So there's thermal expansion in the blade, it gets bigger. The thing has to be structurally sound and aerodynamically sound. But when it's operating, it's actually a different shape than it is when it's made. So to do the design, you're basically going from cold, uh, shape to hot shape, uh, looking at the aerodynamics and the structural integrity, uh, of the hot shape. And then you're converting that back to a cold shape, which is what you have to manufacture, right? And so you're going in a loop on this. And classically there's one team of people who knows aero, there's another team of people who know structures. Um, and doing one, one set of analysis on one blade takes one guy one day. So if you ever want to do very many iterations, you just need these astronomically sized teams. Uh, and so the boom team, two people on the boom team, uh, built this thing that we call Blade Runner. They built it in something like three weeks with Claude code. One of these people is an, uh, engineer. The other one is a person who was an engine engineer but hated the classic big companies so much that she quit and became a software engineer because that was the only way to do anything interesting. And now she gets to go back and actually do the software stuff that she wished she could do at the engine company. So these two people built Bladerunner. And Bladerunner basically has three panels in the user interface. One panel is the blade geometry and you can literally drag sliders. Uh, we have our own custom CAD kernel and it will generate all this geometry in real time. So you can say, oh, what if you change the blade twist? Uh, or you change the cord length, you can m, change the fillet radius, you can change all these things and the blade just refreshes on your screen in real time. Uh, then it runs a real time structural simulation. It shows you that in pane number two, uh, and then pane number three, it has a real time aerodynamics simulation which otherwise wasn't possible to do if you weren't running on GPUs. And so it allows one engineer to basically do iterations on the blade design in moments. Um, and what it means is only a couple people will be required to actually design and optimize all the blades in the engine. And there are like a thousand blades in the engine, so there's incredible leverage. And, and so it means the team size can shrink, which means the hiring bar can be elite. And um, it's enabled by software. But now the fact that software is easier to create thanks to AI just incredibly supercharges all of this because what we find is what we used to do. Actually we still do this as we take software engineers and we embed them in hardware teams. And you don't have to have a hardware background, you just have to think hardware is cool, but you'll sit with the hardware engineers, realize what they're doing that's repetitive and automate it. This allows iteration cycles to improve, um, uh, significantly allows the hardware teams to become much less change averse than they are historically. But now because anybody can code, what happens is the software people play the role of the architect. Um, how should these systems be organized? How do you have clean uh, data structures, uh, ah, how do you actually make sure the system will scale and not hit a wall? So those are the things the software people do then basically they create the framework, they create the sandbox within uh, which uh, all the hardware people can contribute their code. Uh, there's a whole lot of classically all this is done in spreadsheets. Spreadsheets they're actually like software, but they're not treated like software. They have no automated testing, they have no continuous integration. Uh, they have no way of integrating one person's spreadsheet with another person's spreadsheet without a bunch of manual translate, et cetera. So what this does is enable really powerful small teams, uh, to move quickly. And I think where this goes writ large is to the first order effect. You might say, oh my gosh, this is the doom narrative around Jobs and AI. Here it is right there in front of you. AI took a thousand person engine team and reduced it to two. I think actually what will really happen is you have a larger number of small teams. Uh, we did this experiment, um, as I wanted to get the whole company AI pilled and I was like, how do I get. I would like go up to all hands and I would tell everybody to use AI but nothing was happening or very little was happening. How do I actually get everybody to do it? So we paused all project work for a week and we said, everybody work in small teams. Work only on what you think is the most important thing that, uh, you see to the company. You get to pick that you have to use AI and you have to demo it at the end of the week in front of the whole company. So I forced everyone to use AI for a week. And my expectation was we get two or three really creative things that would pay for the entire company being shut down for a week. And I thought mostly we would get dozens of not very good stupid projects. Uh, instead what happened is we had two or three stupid projects and dozens of really good ones. Uh, and also two or three that changed the trajectory. Blade Runner came out of this. So what I learned was that surprised me is even the lowest level employee in the company, literally the shipping and receiving person used Claude to build a shipping and receiving tool. Uh, I was like, wait a minute, everybody actually has an idea?

Speaker B: Yes.

Speaker A: Very few people have the skills or the resources to go realize that idea. But now they're enabled. So I think we're going to have an explosion of entrepreneurship because poor be the person who has no idea for a product or service that could exist. And now those people are empowered. So I think we're going to have an explosion of small teams doing creative things.

Speaker B: I completely agree. Um, in addition to Blade Runner, are there any other key projects? Because one of the things I think is amazing about this is showing how the AI software revolution plays from the world of bits to the world of atoms.

Speaker A: Yeah. So, um, there's another one called makeboom. Um, that's our airplane design software. We created this pre AI coding. Um, but I think it's emblematic of the kind of thing that will be very powerful. So let's make a software hardware contrast here for a second. In the software world, um, your modules of code should be independent of each other, right? If you change this bit of code over here, it should not affect what this code over here does. If it does, something's very wrong. But airplanes are very integrated, right? Um, and there are these intricate trade offs. So for example, if we say, oh, we want to add another row of seats to the airplane. Well, now the whole structure of the airplane has to get longer. It got heavier. The wings have to get stronger to hold up that extra weight. There will be more drag from the bigger wing. And so the engines have to be more powerful. You need to carry more Fuel, you have to find a place to put that fuel. And it all goes back in this loop. So you can't change anything without changing everything. And so classically, the way this engineering is done is there are these discipline owners. There's a structures team, there's an aerodynamics team, there's a propulsion team, there's a systems team. And they all have in their own laptop, some Excel spreadsheet, um, with all their calculations in it. And they have to all hand off to each other every time you want to do a design iteration. And so you can't do very many design iterations because of just the complexity of the handoffs and the fact that the last thing you want to do is tell the guy sizing the cabin air conditioner that you added more passengers. So then now they have to rethink all their assumptions about how much heat dissipation there has to be and resize the entire system. Um, it makes you very, very change averse. Um, what we said is actually we want to turn all of those spreadsheets into software, treat them like software and have all the components talk to each other so we can iterate very, very quickly. So that was the origin of what Make Boom was. Literally you can define a supersonic jet in a configuration file. Um, uh, and you can run this thing and it will evaluate that entire airplane, ah, in a low fidelity way, like in moments. And it can run, spin up a whole bunch of CPUs in the cloud, do high fidelity simulations overnight, uh, and give you a higher fidelity answer. And plus, it can do what you call trade studies, like, do we want a little bit more passengers, do we want a little bit more range? Or if we end up with a higher fidelity model of the engine, we can drop that engine model into Make Boom and see how it affects the rest of the airplane. So one of the. I'll tell a story about this. How this actually, the fact that we built this tool and the fact that we're building our own engine, changed the seats on the airplane. Uh, how one of the things we deeply believe is, uh, you can't just design the airplane technically and then shove the passengers in there and kind of hope it'll all be okay. You have to design the passenger experience together with the airplane itself. Um, so we want engineers to think like designers, designers that think like engineers. And so we've been trying to figure out how to make the cabin really great. Um, and so somebody came up with an idea. I can't say what it is yet, but, but, uh, it Is a breakthrough cabin, but it required the geometry of the airplane change. And so we took that initial geometry, we fed it through the simulation, and we lost 1,000 miles of range. So this airplane is supposed to fly 4,000 miles nonstop. If we put this new cabin in, it could only fly 3,000 miles nonstop, which means there's a whole lot of routes you can't fly. It's like, ugh, uh, I don't think we can give up 1,000 miles of range. And I remember sitting with the team and saying, guys, just think as passengers, do you want to fly in this airplane or this airplane? Everyone's like, that airplane. Like, we all want to be on that one. Like, okay, should we try to find the range? And everyone's like, we should try to find the range. So we had no idea whether it was even possible to go find 1,000 miles out of 3,000. That's a lot of range recovery. And, uh, um, what somebody had the great instinct to do was to generate, um, about a dozen different theoretical engines. And by theoretical engines, I mean, do you want a little bit more fan on the front of the engine or a little bit less fan? Do you want a little bit more core or a little bit less core? And each of these design parameters kind of optimizes for one phase of flight at the expense of others. Um, and it's not immediately obvious which combination will be best. So we generated a dozen combinations of engines, Dropped them all into make boom. Uh, and what make boom would do is take that theoretical airplane on top of a theoretical engine and. And fly it through an entire mission, Meaning it would take off, climb, accelerate through the sound barrier, cruise, descend, land, and then tell you how much fuel was in the tank at the end. Um, uh, and so out of those 12 engines, we found a counterintuitive, better one that gave back most of the thousand miles of range.

Speaker B: Wow.

Speaker A: And so we were able to change what engine we were building, change what airplane we were building, and the result is a meaningfully different passenger cabin. That would not have been possible had we not had the software tool. And by the boom is full of stories like this. I think we can fill three podcasts of stories like that.

Speaker C: Yes, well, so I feel like a lot of critics of sort of using AI using LLMs is like, yeah, you can do documentation, compliance, whatever, but you're doing hard engineering. And to do hard engineering, like, AI is not going to help. Like, what do you say to those critics?

Speaker A: Hold my jet A. Uh, I mean, it's just false. AI turns every hardware Engineer into a software enabled hardware engineer. And by the way, I would also not be dismissive of the compliance implication of this. One of the first things we did with AI was automate regulatory paperwork. And to give you again, a theme that runs through all of this is reducing change aversion, increasing the ability to iterate. So you're gonna go certify an airplane with the faa? One of the very many things you have to go prove are right is the AB withstand lightning strikes. A lightning strike test certification plan might stretch for 200 pages. Um, and so the kind of person who will write this, frankly, not every engineer wants to write a 200 page test certification plan. Uh, the best and the brightest tend not to want to. So you have to have this army of less creative engineers to go crank these things out. And it takes maybe two months to build one. Now someone wants to change. Some crazy person comes along and wants to change the cabin of the airplane. Right? And what that means is all the lightning strike specs are now wrong and need to be redone. And so great. Do you now want to have the army of certification engineers rewrite all this technical documentation for the regulators. It will create this organizational resistance to change because the actual cost of change is high. Well, it turns out we built a rag with all the regulatory documentation it and you can prompt AI and get a like 90% correct lightning strike test plan. And so what that means is, and basically the airplane design is like embedded in the prompt. Ah. And so if you update the airplane design, in fact, you could probably auto update the prompt and then out comes the test plan. So what it does is, uh, it reduces the aversion to change.

Speaker B: Yes.

Speaker A: Uh, what this means is we'll be able to go certify way more iteratively than we can today. If you play all this out today. Um, one of the many reasons we don't get new airplanes, uh, is Boeing and Airbus think like patients going to bad doctors. Right? Like you go to a doctor, you say, doctor, it hurts when I do this. And the bad doctor says, well, don't do it. Uh, right. And the great doctor says, let's figure how to make it not hurt. But, um, basically Boeing was the patient that went to the bad doctor. Uh, it hurts when I certify an airplane. Okay, well, let's not certify any new airplanes was their response. This is why they're still making 73760 years later. But if you can, Um, I think the right mindset is to make certification not painful. Uh, and part of that you can do company side. Part of it requires working with regulators, but by the way, they will totally do it if you keep their trust. And so this will allow us to iteratively certify many m. Many iterations of the airplane. And so the whole improvement curve is going to be way faster than it ever has been before. And AI is part of that story.

Speaker C: Can you tell us for a second? I feel like everyone, when we're trying to innovate, we talk about regulation, slowing it down.

Speaker A: Yeah.

Speaker C: And I feel like a lot of Americans agree, but with air travel, I feel like a lot of people would say, ah, but don't mess with that because safety is so important. Can you tell us why your way is actually safer than the alternative? Because I think that's where some people would have a question.

Speaker A: Yeah. So, um, we're not trying to drop the safety bar. If anything, we're actually increasing it.

Speaker C: Exactly.

Speaker A: Uh, the question is really, how do you prove to yourself, to the public and to the regulator that you've met the safety bar? Um, this is about reducing the cost of proof of safety, not about moving the safety bar. And I'm very proud of the safety culture we built at Boom. We did this airplane program with zero incidents, zero mishaps, uh, uh, no major injuries, no OSHA reportables. It's about reducing the cost of demonstrating safety, not the other way around. I do think some of the most important regulatory stuff is actually not in the air, it's on the ground. So a thing I'm spending a lot of energy right now on is trying to figure out how to build physical infrastructure faster. We are a year behind on our engine test facility because a local fire bureaucrat who knows nothing about how to build anything safely is forcing us to apply the code in such a way that is far more difficult, far more expensive, and less safe. And this makes me want to set my hair on fire. And this is the entire way we build in this country, I think, is backwards. This country is found on the idea of innocent until proven guilty. Uh, but not if you build. If you build, you have to prove your innocence before you're allowed to do anything. And if we drove to work the way we build in this country, it'd be like, oh, can't get on the roads. Uh, you have to have a plan. What, uh, route are you going to take? What speed are you going to drive? When are you going to use your blinkers? Then some bureaucrat waits a few months and then gives you feedback and tells you you should really take 101, not the 280.

Speaker C: Ah.

Speaker A: And you're like, fine. And then eventually you can drive. But that's the way we do buildings today. Um, what instead we should do is make building work like roads, where you have to have a license. Uh, so you prove that you're competent to do this. Uh, and then there are rules, but the rules are enforced rather than pre approved. Um, uh, and so what you can then do is just build at your own speed. And if you break the rules trivially, you're fine and you have to fix it. If you break the rules, seriously, you lose your license. Um, if you break the rules in such a way that somebody gets hurt, that's criminal when you go to jail. Just like driving. And if we had that model, we'd be able to build so much faster. So much faster. And it would have zero compromise on safety. In fact, I would argue very strongly that you'd actually have a safety lift. A lot of times the people enforcing the rules, um, particularly for complex facilities, are not domain experts. And we're domain experts on how to do complex hardware safely. Uh, the local fire department is actually not.

Speaker B: Um, well, let's go to one of a really interesting product extension, which is Superpower.

Speaker A: Yes.

Speaker B: Um, so it's kind of, it's one of the most elegant moves I've seen in a long time. You're using the engine you built for a plane, uh, to fund the plane and in the process, potentially solving the power crisis for AI data centers. Um, that's not just a financing strategy. It's a systems insight. What did you see that made that connection obvious? And when did you see it?

Speaker A: The origin story on this. So we sort of, um, again, famously had our breakup a few years ago with Rolls Royce, uh, did the thing that everyone said we couldn't do, which is just take on building an engine ourselves from scratch. And I remember sitting in one of the first technical presentations, the lead engineer said, by the way, your engine looks a heck of a lot more like a power turbine than subsonic engines do. Because subsonic engines are about a burst of power to get off the ground. And then you hear it as a passenger. The engine kind of gets quieter after takeoff. Um, and the supersonic engine is actually partial power for takeoff, but full power the entire flight. Supersonic. Um, and that's much more like a power turbine. And so the engines, by the way, if you ever. So everybody who builds jet engines kind of has a side business selling them on the ground for power. Ah, G.E. pratt. Everybody does this. And the engineers are like, by the way, if you ever wanted a power turbine business, you'd have the most amazing thing, because there are no new power turbines. Your engine will be new in state of the art, and, and it's naturally aligned with what makes a great power turbine. And I was like, uh, uh huh, uh huh. We're an airplane company people. Next slide. Um, that was four years ago. Uh, and so fast forward to March of last year. So we've broken the sound barrier on XB1. My primary focus is lobbying to have the speed limit in the US repealed, which did happen successfully. And I see the things that are like, we all know now AI is massively power intensive. The grid is not growing. Takes five to 10 years to get a grid hookup. Everybody's starting to say they have to vertically integrate their own power generation. To do that, you need turbines. And now the turbines are sold out. And the turbine companies are not AI pilled. They got burned the last time there was a data center bubble. They don't want to invest in production capacity. And so everyone's stuck. And I was like, huh? I've been looking for 10 years to find what could be a way to pull revenue to the left. How do we building, uh, a supersonic airliner, certifying it for passengers, and then only when all of that is done do you get your first revenue dollar is like a financial Hail Mary. And so I've been trying to find what can we do that will get more cash sooner. And we said no to a thousand things that were shiny objects. But so I remember seeing those tweets. And I sent two texts. Uh, one was to Sam Altman, who backed us, uh, about the same time you did Reid back way in the early days. Uh, another one was to my head of proposals. And to Sam, I said, hey, if we had a 40 megawatt nat gas turbine, uh, would that be of interest? And the other ones, my team was like, what would it actually take to do this, guys? Is this feasible? And Sam wrote me back right away, was like, yes, please. Here's who you should call. My team was like, oh, yeah, we actually had a plan on a shelf. We just didn't know there was a market. And I was like, ah, so this is opportunity. And I think it was about 90 days from those text exchange to we actually had a billion dollar LOI for a launch customer. And, uh, then we went all in on this. Uh, so we realized this is the most important opportunity in front of us. It will allow us to generate the Cash to do the airplane. It will also allow us to test the engine on the ground. So we'll go to the FAA with the most tested new engine ever. Right now, everyone else builds these aero engines and then they convert them for the ground. I think it's backwards. Run out of the ground first, beat that thing up where if it blows up, nobody gets hurt. Right. Find all the bugs, iterate on the ground and then put it on the airplane.

Speaker B: Yes.

Speaker A: So that's what we're doing. And so, you know, as we speak, 90% of the team is on the part of the airplane called engine, and 10% of the team is on the part of the airplane that is not engine.

Speaker B: Yes.

Speaker A: Uh, and we're just very focused on speedrunning this development, getting as much power production online as we can as quickly as we can.

Speaker C: So everyone talks about bringing industry back to America. And instead of bringing the old industry back, I mean, you guys. So superpower is your engine powering data centers. This is a picture of American industry that is working and sort of for the future. Do you see this as a way to really be optimistic about American industry?

Speaker A: Oh, absolutely. If we just get out of our own way, great things are going to happen. Getting out of our own way regulatory is a big deal.

Speaker C: That's huge.

Speaker A: Um, but I think there's also another thread underneath this, which is about the, how we were talking earlier about how it takes a year to get a turbine blade. And so a single crystal investment cast turbine blade is actually 12 months plus out of the traditional supply chain. And the longest thing in that is actually the tooling production, meaning, uh, like the complex high precision mold that you need to go cast things in. And part of that is because the American tool and die industry, uh, basically died, uh, no pun intended, uh, and it all went to Asia. And there's a whole other story we could tell about how China sucked that out, but it did. There's very little tool and die, um, left in this country. So you might say, great, if we want to re industrialize America, we've got to go rebuild tool and die. Sometimes that is actually the right pattern and we will need some tool and die. But our first thought was, can we just do this without any tool? Can we just delete that part of the manufacturing process, uh, and can we go to more additively based approaches for this? And I'll be stealing my own thunder if I go too much further in this story. But what I will say is this is a great pattern. You don't try to bring back what left China? Uh, what you do is invent the generation after it and have that be here. And it tends to be digitally enabled. Um, it tends to delete hard tooling. It tends to be robotics heavy. Um, uh, so we should just don't try to bring back old school textiles as an example. Build next generation textiles.

Speaker B: M A natural thing for boom. Um, is comparing with Concorde and Apollo, which were impressive tech demos but bad products.

Speaker A: Uh, you said it very well. They were tech demos.

Speaker B: Exactly. So kind of go into the details of what the thinking, how the thinking was a mistake and what thinking we need.

Speaker A: Yeah. So I have a very. I get looked at askance at aerospace events because I actually think Apollo was a terrible idea and we shouldn't have done it. And, uh, before you slaughter me, m so imagine going back to 1969 and finding the most pessimistic person you could find on the street and saying, in the year 2026, what do you think space travel will be like? And what do you think, uh, air travel will be like? And of course, 1969 was the moon landing and also the first Concorde flight. I don't think you could have found anybody so pessimistic to have predicted what actually happened, which is we can't go back to the moon and we can't fly supersonic. And so if Apollo and Concorde were supposed to be these massive leaps forward, something went wrong, it didn't deliver right. And I think what happened was we got confused about motivation. The first half century of flyte was entrepreneurially driven innovation. Um, where Adam Smith's invisible hand forced product market fit. Nobody built products for tech demos. You had to build products that had markets. But instead JFK said, great, let's go to the moon so we can show up the Russians. And also let's build supersonic airplanes so we can also show up the Russians. And very ironically, in our attempt to defeat communism, um, we took a communist approach. We said, great, let's beat the Russians with our own centrally planned projects. Turns out communism doesn't work any better for Americans than it does for Russians. Central planning is a terrible idea. Uh, and so the centrally planned projects built useless things. And I think we all sort of recognized that Apollo had no commercial intent. Um, what is not broadly recognized behind Apollo is it contributed this entire supply chain that was cost insensitive. Right. And then after Apollo's done, there's no vision beyond it. So the, the Apollo program gets shut down, but now you've got all these suppliers used to these fat Happy government contracts that basically find other ways to keep themselves alive while doing nothing hard. Then an analogous thing happened on the supersonic side. Concorde at least looked like it was supposed to be commercial. Right? This is an airliner. And people start to buy tickets. But nobody thought too hard about this. Here's a 100 seat airplane adjusted with tiny seats. Adjusted for inflation, fares are about $20,000. Um, a 747 at the time would have maybe 40 seats in first class at a lower fare. So hang on, we're going to have two and a half times as many seats at a much larger fare, A much less comfortable experience. Uh, this doesn't add up at all. Obviously the seats will be empty a lot of the time. Even the most popular route, which is New York to London, it was half empty economically. It couldn't do anything else. It's a classic example of no product market fit. Um, and why? It's because the mechanism that drove product market fit, Adam Smith's invisible hand makes every entrepreneur focus on this. And we cut off Adam's hand here. He was not involved in the creation of Concorde. But then people didn't realize this and they drew all the wrong conclusions about supersonic. Like, oh, people won't pay more for speed. It's inherently more expensive. Blah, blah, blah, blah, blah. Supersonic airplanes are tiny and uncomfortable. None of these things are inherently true. It's as if we had built the univac. It had stopped iterating in computers. And then everyone says you couldn't possibly have an iPhone because all computers are the size of a room. But that's what we did in flight.

Speaker C: So thinking about supersonic flight, I'm sure you've never heard this critique before that this is just for wealthy people. You're building something that's a luxury good. Why should we care about your company if it's just 100 rich people? People who are going to fly to London faster.

Speaker A: Um, so first off, just to factually say where we're starting here, um, we've been able to take about three quarters of the cost out. So flying supersonic on Overture will have the cost profile of business class today, which is something tens of millions of people already do.

Speaker C: Incredible.

Speaker A: Uh, now that said, that's where we start, not where we stop. Um, uh, we're not done until. You can't convince kids that we didn't just always normally fly supersonic. Um, I have a kid's story. We can come back to on this if we want, but this will be like electric cars. Computers, cell phones, uh, where initially the market is high end, but with iteration and scale the cost comes down. And what a lot of folks don't realize is so when you go faster there is an energy intensity. So it matters to reduce, um, energy efficiency matters and choosing a low cost energy source matters. Um, but uh, when you go faster, you actually get more done with the same airplane and crew. Uh, so if you design the whole thing for high utilization, you get what we call the speed dividend. And so, um, there is a tipping point actually where the fastest flight can actually become the cheapest one. Uh, not because it had the lowest fuel usage, it probably won't, but the total operating cost thanks to utilization was better. So we're going to keep iterating towards that. And I think we end up somewhere in the vicinity of today's economy fairs, uh, but with a much better experience. Uh, it might be that the lowest supersonic is actually economy plus, I don't know, 20%. But the experience will be something where you're treated far better than you're treated in an economy today.

Speaker C: So to be clear, my kids will be upset about this because they're allowed to watch unlimited shows on planes. So they just want the flight to be the longest possible.

Speaker A: Uh, we found the one passenger that doesn't want it.

Speaker C: Exactly, exactly. But, uh, tell us your kid's story.

Speaker A: Yeah, so this is how innovation impacts the next generation. Um, my youngest, when she was about 18 months old, uh, I would just watch SpaceX launch videos with her. Um, she got fairly obsessed with them. She would ask me for red car rocket and it took me a while to figure out what that was. That was the Falcon Heavy test flight that had Elon's bright red roadster on it. So red car rocket meant Falcon Heavy. But eventually we get some. Now she, at this point in the story, she's maybe two and a half. I take her to go launch one of these Estes model rockets and so we go out and like put it on the pad and of course I'm, you know, I am who I am. So I got the smallest rocket, I put the biggest rocket motor in it. And so we launch this thing off and boy, it goes and it ends up in a tree like three fields over. And she's so upset. And at first I thought she was upset just because we lost the rocket. But it turned out the reason she was upset was she completely expected the rocket to come back to the pad. And I realized I'm just this dummy of a parent. I've only shown her videos of rockets landing, because I think they're so cool. And of course, she would have no idea that this is a recent innovation. Uh, so that's where we want to get with supersonic, where you have to explain to kids that flights used to be longer.

Speaker B: Absolutely. Exactly. Well, speaking of which, you personally led the campaign to legalize supersonic flight. Flight over the US we talked about that briefly earlier, testifying before Congress, pushing the White House. And the House just passed SMA unanimously. What did you learn about how American policy actually gets made?

Speaker A: So much. It was 115 days from, uh, announcing boomless cruise to an executive order to legalize supersonic flight. Yes. And as you said then, the House passed a bill to codify that unanimously. Uh, and we're working to make the same thing happen in the Senate now. Um, uh, American politics, in my view, has become very tribal. And the moment you pick, uh, the moment you take a position that's aligned with any human, now you're branded as on a team and the other team is evil. Um, and, uh, I think it's pretty bad for the country. Um, and it leads to a lot of unnecessarily polarized position taking. I call it values oriented communication. It sounds hokey, but I don't talk tribally. I just talk about what really matters. And so I can walk into a Republican's office and talk about why supersonic flight matters and why it should be legal. And I can say only things I passionately believe to be true. Um, and I can walk into a Democrat's office and I can also talk about supersonic flight and say only things I passionately believe to be true. They are different, but compatible sets of things. Um, like, you know, if I'm speaking to a more left audience, I talk about human connection. I talk about what happens when our kids grow up with friends from other continents. Um, what happens when we're far more culturally connected. If I'm in a Republican's office, I talk about the need for American leadership and technology, how a lot of defense technology is commercially derived. And if we don't have the next generation of commercial airplanes in the country, we're not going to have the next generation defense airplanes either. And that's really bad. And so we make different arguments in different places, but they're all compatible and they're all true. And what I find is, I think on many of these issues there's just a rational middle 80% that is normally not spoken to. One of the kind of neat full circle things here is the, uh, Opposition to supersonic flight was originally a leftist issue. Uh, uh, the ban on supersonic was one of the first victories of the environmental movement. Um, entire books were written about the evil of supersonic flight. All of this is basically forgotten. By the way, one of the guys that wrote one of the books wrote me an email. And, uh, it was one of the best emails I ever got. Because he's like, I helped lead the campaign to ban supersonic flight. I have listened to you and how you're doing it. I decided I was wrong, and I'm now behind it. I was like, what just happened? But just speak the truth passionately and explain the human value of this stuff. And I think if you do that, and my hypothesis is this actually would work on a number of political issues, issues that conversations aren't happening today, but you could unify people.

Speaker B: Um, what's the most important thing that still has to go right, even if it's not fully under your control?

Speaker A: We just have to not screw it up too badly. Uh, let's start with what fundamentals are good here. People want passenger supersonic flight. I don't know. Apparently, we just heard about the first humans that want longer flights. Your kids. But everybody else wants a shorter flight.

Speaker B: They might learn.

Speaker C: Yeah, they'll learn.

Speaker A: Everybody else wants shorter flights. And, uh, airlines want these so long as they can make money. Airlines are so frustrated with the duopoly, they're desperate for another differentiated product that would actually grow the market. The technology is all there. Uh, we demonstrated the regulations are there. And also, they're way more malleable than people thought. A year ago, the answer would have been. Capital is an enormous challenge, but we actually figure how to make our own capital. Uh, so, uh, we now have in the bank account enough money to get from here to actually delivering the engine. And the engine prints all the rest of the money to deliver the airplane. So what's left is really, do we make a mistake? Um, do we take too long? Do we screw up something critical? Do we have a setback we can't recover from? What I worry about every night is just, what mistake are we making, uh, that we don't know about that will be painful to fix later? And just how do we. How do we discover those quickly? And so I put a lot of my energy into building a culture where people, um, talk openly about problems. That's one of the most important things, the natural human tendency is when your thing isn't going right. Well, I don't want to tell anybody. It's embarrassing. Uh, I will hold onto it and fix it. And so what tends to happen is the problem metastasizes and the problem is shared only when it's, like, really quite big.

Speaker B: Yes.

Speaker A: And. And, uh, then it's very hard to recover from. So we really try to train our people to share their problems early.

Speaker B: So one of the things about these, you know, difficult projects, the ones that are long, um, is that they take an enormous amount of grit and conviction. So, you know, um, you know, as we've talked about, you spent a decade on something most people think was either impossible or relevant. Um, so what's your theory on why ambitious things take as long as they do? We've been talking about speed a little bit, um, and whether that's a bug or a feature.

Speaker A: Well, I have a couple thoughts on even the first part of what you said, too. Um, I think it does take a lot of grit, but I think you can engineer the grit. Um, the way I picked Boom as the thing I wanted to work on was by optimizing for how happy I would be if it succeeded. And, uh, that we nearly failed so many times. We got into seven days of cash once. Uh, we had to do a down round. It was very painful. I don't recommend it. But I got through those moments because I never lost conviction that this should exist. And so I think founders should be more selfish. They should engineer. They should pick something where they will never question, is it worth it, um, no matter how hard it gets. And I think that's very important. But, okay, so why do these things take so long? Boy, if we'd been wiser in day one, we could have done this much faster. Uh, part of what's going on here is we had to rediscover how to do a startup in an industry that didn't have startups.

Speaker B: Right.

Speaker A: Like, I had no background in this. We had to go figure out a team that would figure it out. And a lot of what we know today, we just didn't know on day one. Like, the single biggest mistake we made on XB1, uh, superficially, was I made it way too hard. Uh, we said the goal was to fly 10% faster than Concorde. It had to have a second sequence in it so we could call it a passenger airplane. Uh, it had to not have afterburners. Like, I imposed all these requirements that made it actually far more difficult to engineer. Some of those requirements were actually dropped along the way because they weren't actually necessary. But we went far enough with them that it made the airplane harder. Okay, so I made it harder. Um, well, why did I make it harder? I didn't know how hard it was going to be. I was pretty naive. Okay. Why did it take me so long to figure that out? Well, this is the real worst mistake on XB1. We only plan to build one. What we should have done is plan to build a series and made each one, the early ones in the series, far simpler. We should have been iterative, right? Had we been iterative, we would have learned, uh, the lesson about all the requirements being too hard and we would have changed it much faster. So the single biggest lesson learned was just to be iterative in hardware. And that was, um, in hindsight it feels really obvious, but it was, um, it certainly wasn't obvious to me in 2014 and I got the company on this very long path that was unnecessary. But now we've got a whole bunch of hardware startups, right? And we're like exchanging lessons learned. And now the next entrepreneur founding the next boom won't make that mistake.

Speaker C: Okay, so that's my question in hindsight is 20 20. So now you know all the mistakes that you shouldn't have made, but you think the next generation of founders can have what it know, all the things on day one that you didn't know.

Speaker A: I want way more hardware startups in America. So I'm not, I'm not secretive about any of these things. Uh, I write a substack about this. Um, I speak to every YC batch and I meet with the hardware founders and I just try to share, here are all the things that I got wrong that nearly killed Boom. Don't do those things, do different things. And one of the most essential things is you have to get to integrated hardware quickly. You have to plan to have several units along the way that you throw away that you iterate on and that forces you to keep it simple enough. Mhm. That you can actually go do the iteration and then that creates this learning loop where the mistakes, the mistakes get found and fixed inexpensively.

Speaker C: So I think most, well, who knows what most Americans think. But when you think about the difference in software and hardware, hardware just seems so hard. Like things are in the real world. It's hard to do. Like, how do we get more hardware startups? Like, what is that path?

Speaker A: We have to overcome that mythology, right? Like, uh, the notion that startups exist only in software is a relatively recent notion, right? Like Silicon Valley. Let's remember Silicon is hardware, right? Like we used to have semiconductor startups. Um, and then I think we got like A little drunk on SaaS and forgot that anything else could exist. And there's this mythology that hardware startups, uh, uh, consume more capital than software startups. But if you look at how much VC Uber consumed, how much VC Airbnb consumed, blah blah, blah blah, and you can compare it to Anduril, SpaceX, uh, actually the hardware startups consumed less venture capital than these asset light companies. So what do you make of that? I think the one thing that really is different, it is more expensive to build a prototype m and is more expensive to ship. So in the SaaS world, how do you prove you have a good idea? How M do you prove you have product market fit? You just ship and you prove it by actually being in the market. Right. I cannot build a supersonic jet in order to demonstrate that people want to fly supersonic. Right. You have to have other mechanisms. And frankly Kickstarter shows the way. You know, we got pre orders for our airplane with deposits. We did, we did Kickstarter for supersonic flight and ah, so you have to be different. There are different playbooks, um, for product market fit, for long duration expensive hardware projects, but they're not. This isn't rocket science. It really is just Kickstarter.

Speaker B: It's just jet science. But yes, um, so the jet age, um, produced second order effects that nobody forecast in 1960. Uh, global tourism, just in time, supply chains, the modern business trip, fresh produce the year round. Um, you said the thing you're most excited about isn't the travel itself, it's the effects you're not smart enough to predict yet.

Speaker A: All the most amazing stories about the jet age are the things that nobody would have predicted. Like shoes. Um, like Phil Knight, I only discovered this like reading Phil Knight's memoirs. He's the founder of Nike and he uh, fell in love with Japanese style running shoes after business school and a transfer to Japan. And then Nike actually got their start importing Japanese running shoes to America. This couldn't have happened without a jet baseball. There was no major league baseball west of the Mississippi before the speed of the jet. Um, and so I, uh, think with supersonic there will be similar stories. We just don't know what they are yet. And I think the most important innovations in the world tend to be platform innovations, innovations in which other innovators can build, like the smartphone, I don't think. When Steve Jobs stood on stage in 2007 introducing the iPhone, one of my favorite moments in the keynote is he's got this like, uh, it's a revolutionary mobile phone A widescreen ipod and a breakthrough Internet communicator all in one. Right? And he's going through this and there's like this like heavy applause at the phone line. There's this heavy applause at the ipod line. And he says, breakthrough Internet communicator. And it's kind of muted applause. Everyone's like, what is that? Well, it turns out it's actually the only thing that mattered. Right. But he had no idea that Uber, uh, would come out of the breakthrough Internet communicator. It was a platform innovation. And I think airplanes in a non obvious way are also platform innovations. Uh, they enable kind of more liquidity of movement on the planet and that liquidity creates non obvious, um, benefits. I think we can predict we'll have major league sports worldwide, uh, but there'll be other things that are less predictable.

Speaker C: Well, I also think I just read that the median American took 00 flights last year. And that's America. And that's compared to, you know, countries in Africa or Southeast Asia. Like no one is flying. And so there could be so many changes from a global perspective.

Speaker A: Totally. Like a thought experiment is imagine if you could get anywhere in the world in four hours for $100. Now, I'm not sure we'll actually literally get there or anybody would really want to be on the $100 flight. Um, but think of how differently you live your life. Um, and think about what if it was all available last minute so you could just decide this weekend to go to Mumbai and come back in time for work on Monday. Right. That is a not impossible future. Uh, and it will be one that has incredible second order effects.

Speaker C: So a lot of people have thought your ideas about the future were crazy throughout this career. Is there something about the future of flight right now that you believe that almost no one agrees with?

Speaker A: Probably a lot of people still disagree with everything I just said.

Speaker C: Right?

Speaker A: That might be the answer. Yeah, I think this is actually going to happen. Uh, and there's still a lot of people in the industry that just don't believe it.

Speaker B: Yep, fair. So we're about to move to rapid fire. So you can answer these at any length you want. They're just the same questions we ask every guest. Um, is there a movie, song or book that fills you with optimism for the future?

Speaker A: I love the Martian. Um, and I think so many. We are missing, uh, utopian sci fi and we have way too much dystopian sci fi. I think this is, ah, the Martian is great. And you know, one of my favorite things is Kind of the final scene where Matt Damon's in front of a college class and he's talking about almost dying. And he has this thing about, you know, uh, something's going to go wrong. And if you just solve one problem after another, if you just keep doing that in a row long enough, you get to live. And that resonates with me. And it makes me think of the life of an entrepreneur. Like, if you just keep solving problems and don't stop, then eventually you get to ipo.

Speaker C: Uh, what is a question that you wish people would ask you more often?

Speaker A: There's some technical ones that I think are underappreciated. Like, what does the speed of sound have to do with an airplane? Nobody ever asks that. It's non obvious, never talked about.

Speaker B: Should I ask you that question?

Speaker A: Uh, yes.

Speaker B: Okay.

Speaker A: So what does the speed of sound have to do with an airplane?

Speaker B: Yes.

Speaker A: Uh, the speed of sound is the speed at which press pressure waves propagate in air. Right. And a lot of what an airplane does is about manipulation of pressure. Uh, so that's kind of the fundamental connection. So what's supersonic? Um, supersonic is when, uh, the airplane outruns the pressure wave. Imagine there's like a group of people, a crowd, and we're running at them and they can see us coming. Uh, they're gonna kind of look at us weirdly and then they'll start to move out of the way. Right. So we don't just crash into them. That is subsonic. Like there is a pressure wave in front of the airplane, uh, running ahead of the airplane. That's basically saying, hey, I'm coming. Start getting out of the way. And, um, when you go supersonic, uh, it's like running at a crowd of people that can't see you coming. So what happens is you crash right into them and they start bumping into each other and falling over like dominoes. That is exactly what happens with air. When the airplane is going faster than the air's ability to get out of the window way, it compresses the air. Uh, so any part of the airplane that is moving supersonic relative to the local airflow gets compressed. And so you get all these little bands of compressed air coming off an airplane. That's what a shockwave actually is. And then those shock waves, as they come off the airplane, they coalesce. They constructively interfere as they get to the ground. And that's what causes, uh, a sonic boom. It's just these bands of compressed air that stack up on top of each Other. That's it.

Speaker B: Unless, of course, they reflect before they get there.

Speaker A: Exactly.

Speaker B: M. So where do you see progress or momentum outside of your field that inspires you?

Speaker A: Oh, there is so much to be excited about. I mean, uh, you know, we're in a golden age of space thanks to SpaceX. Um, the second order effects of that we're just starting to see, um, with global communication. The other examples I think are kind of widely known. Right. I think AI, it's going to be even bigger than people think it is. The second order, I don't think we're going to have a job apocalypse. Uh, I think we're going to have an explosion in entrepreneurship. Maybe the biggest beneficiaries are going to be YC and VCs, because we're going to have way more startups, uh, than we've ever had before. And I think that is incredibly, incredibly exciting.

Speaker C: Well, as always, our last question. Can you leave us with a final thought on what is possible to achieve if everything breaks humanity's way over the next 15 years? And what's the first step to get there?

Speaker A: A question I like to ask people sometimes is people get very doomerism about lots of things and they think there is a past that is somehow better. But I like to ask people, um, if you could pick any time and place in history to live, where would you pick? And what this tends to do is force people to be rational because they realize that they go back in time, they have to give air conditioners and cars and indoor plumbing if they go back far enough. But then still most people fail to realize that that trend probably, uh, progresses into the future, and that probably the future is much better than the present. But that doesn't happen automatically. Uh, people have to go make it happen. And I think if we can overcome, I think one of the most important things to overcome is the bystander effect in innovation. This notion that if it's a good idea, there's already a team on it, which causes people to ignore some of the most important unsolved problems. Like one I love to point out is traffic. There are, as best I can tell, zero teams actually working on the root cause of traffic. But it's very solvable and it actually could be a gigantic business. I think if we can overcome that culturally, uh, and we have this explosion of entrepreneurship, we will have solved problems that people don't even realize are problems today.

Speaker C: Awesome. Blake, thank you so much.

Speaker A: Thank you.

Speaker B: Always a pleasure.

Speaker A: Thank you.

Speaker B: Reed Possible is produced by Palette Media. It's hosted by Ari Finger and me, Reid Hoffman. Our showrunner is Sean Young. Possible is produced by Tenasi dlos, Katie Sanders, Spencer Strasmore, Emo Zhu, Aman, Suri, Lexi Kevin, Danny Garrison, Trent Barboza and Tafadwa Nimurundwe.

Speaker C: Special thanks to Surya Yalamanchili, Sayda Sepieva, Ian Alice, Greg Beato, Parth Patil and Ben Rellis. And a big thanks to Laura Arnold, Laura Haine and the team at Greylock.