B (1:54)

It's a bigger crowd than I realized.

C (1:56)

It is a big crowd. We're very proud of that and we're glad that you're here. Thank you for making the time. I'm going to put up on the screen just so that everybody has a little bit of history. I'm not going to read you all of this.

B (2:06)

Sorry about this.

C (2:07)

Okay. No, that's to you. You should be very proud of it. But it would take way too long to introduce him. But as a co founder of Palantir, obviously a lot of you know that company as being involved in some of the most innovative companies that exist on the planet. We wanted to put them up here, as well as some other things we'll talk about, including the University of Texas of Austin. I love it. She is a co founder.

B (2:29)

That's true.

C (2:30)

University of Austin. I said it again, didn't I?

B (2:32)

University of Austin. I actually started all of these except for Era Bar Palmer started Airborne. But I'm back at him.

C (2:37)

Got it. You're with him.

B (2:38)

They did a big thing where they put my face next to his. On probably thought that I did, but it was not mine.

C (2:42)

That is okay. So Joe has got an obviously incredibly varied background, so I could take this conversation like a thousand different directions, but I'm not going to do that. We're going to try to stay really refined on a couple of things, the first of which is AI you have a very unique perspective, which we're going to talk about in just a second. But what I'd love for you to do is let's talk about what the myths are. And I. Let's talk about, you know, kind of what's real and what's not. Go with that one.

B (3:14)

What's real and what's not. I was on, you know, you know, I suggest. I was on the Zoom with Dario, who runs Anthropic today. If you guys haven't read his latest piece, I think it's actually a very good piece. He and I are. We differ politically, but he's a really obviously smart guy. It's one of the top three labs. You know, he's raising at $350 billion. And he's really very worried. And I agree with him about some of the things he's worried about. I think what's not real. What's not real, at least to me in my view of the world, is I don't think this is like existential thing that changes our entire civilization where we've created. And a lot of people think you've created the God, basically, or we will create a God in the next several years. And to me that's very unlikely. I think it does change a lot. I think it's very, very important how it changes the whole world. But I think a lot of people Especially in Silicon Valley, they don't really have the same maybe religious tradition that we have here more often in Texas. And so a lot of them, this has become their religion where they think this is like a whole new change in the entire world and everything's gonna be different. You've created that God is gonna run things. And I think that's not the right way to look at it. I think the right way to look at it is as investors and as citizens is this is a very important industrial revolution. So I think productivity is definitely going to go up a lot. The last time productivity grew this fast was probably the second industrial revolution, which is 1870 to 1900. If you take a typical person in America, the median wealth, median income, it more than doubled in that less than one generation, basically. So you basically have people, I think almost two and a half times as wealthy. It was very, very good for America. It was really awesome. It was a lot of change. Right. A lot of jobs that existed in the 1860s and 70s did not exist by 1900. You don't have Coopers anymore. We have our last names as Coopers or Smiths or Bowyers or whatever. But you don't have the jobs and you have a ton of these jobs go away and that's going to be similar. Now there's about $5 trillion today in wages in the services industries and already $2 trillion worth of wages. So already 40% of our entire service industry, we've seen examples of being able to double the productivity. So, so, so it's in some cases triple or quadruple. In some cases, we've already bought companies two years ago that we've already tripled their cash flow in certain areas. Right. So there's a lot of stuff already happening where productivity is 100% real. It's going to change a lot of things in a lot of positive ways, hopefully. But. But I don't, I don't buy the whole, the whole existential thing.

C (5:34)

Well, it's interesting because, you know, by the way, you thought I talked fast. He talks pretty fast. Right? So that's okay. You do you be you and they'll keep up. And you know, we're going to cover a lot of ground here. So what I will say is that the ability to think through what it is that truly is investable and what is not and what AI is going to be. I mean, it's funny because Christopher and I were working on an agent that I've been working with chat GPT on and I was being pretty firm with it. Because you're trying to train your agent, right?

B (6:13)

Yeah.

C (6:14)

So I was correcting it and Christopher thought maybe I was being a little bit harsh. He goes, they're going to kill you first.

B (6:19)

They come back and get you.

C (6:20)

I don't know, maybe I might want that.

B (6:22)

These things don't have very good memory yet, so. And I think they delete the logs. It's. It's. When we think about, if I may, on AI, I think, I think for how to invest in it, we're computer scientists. We start counting from zero. And so they think zero through five, there's six levels. Level zero is energy, which is great for Texas, by the way. I think there's going to be a lot of energy that we need. Whenever you hang out with these guys who run these companies, they were admitting to me they're not even telling how much they think they're going to need because they don't want to scare their investors. So you're going to need a lot of energy. And then level one is the chips companies. Right now, Nvidia dominates that with intel and AMD. That's over 99%. But there's a lot of new chips companies too. They're interesting. Level two is data centers. All of our friends have been investing in. That's been a great alternative asset class for a very long time. It's very hard to measure how much you need, but that's obviously a big thing. Now level three is the model companies themselves. That's anthropic, OpenAI Xai, Gemini, Meta Chinese models. And those are obviously very interesting companies growing faster than anything we've seen before. A lot of people say, what part's going to be commoditized? What part is going to not? That's a big question. Level four is the software infrastructure. These are the companies that help you deploy AI, no matter who you are. So it's something like databricks for. For organizing your database and infrastructure. Something like Palantir, of course, which has been playing a key role there. Ollama, which helps you deploy models. So lots of stuff there. And then finally level 5, the last level, is the actual apps and services themselves. It's the companies actually creating value directly in the economy, actually enabling things to happen. And so there's all these levels are all interesting to invest in. I happen to think level five is the best risk reward, but they're all pretty good right now.

C (7:56)

Okay, so just because you went through them pretty quickly. So five is basically the companies that are helping companies do better or Even

B (8:03)

doing it themselves or even doing it themselves. And this is a big difference in this wave. Historically you would like build a software company to help a services company deliver value. Most of the time now we're just building a services company ourselves because there's a lot of really deep integration with the operations and the software for who does what. And over time you just get better at it. It's much harder if you just give that to someone else. It's much easier to own it. So for example, in logistics on the back end, when you send something, there's about $70 billion of revenue a year paid to companies that handle all of the kind of back end payments and reconciliation.

C (8:39)

Right.

B (8:39)

There's 18 systems you got to plug it into. Everyone has to talk to each other. It's very manual. These companies tend to have 15 or 20% margin. We have a company competing with them that has 60% margin now and that has actually bought and tripled the margin of things. Or I'll give you another one. Healthcare billing. There's $280 billion spent in the economy each year. That's a lot of money on healthcare billing. 120 billion of that is actually paid as revenue to companies. Doing it for you, 160 of it is internal. The systems here, they have all things themselves internally and that's also a 15 to 25% margin business. There's also at least two businesses I know of there at scale doing over 50% margins and they're doing it with AI and to break in. So you have this all over the economy where you're actually competing directly.

C (9:19)

Now. We had a great interview with Robert Smith, the Holy Relative Investing podcast. Tony and I did, and we talked about how software companies are actually being disintermediated by software. So even the companies that they have, they're concerned on some of them that literally they are not adjusting fast enough to what the customer's needs are. So the customer is actually building it themselves and that obviously disintermediates them.

B (9:43)

So that's, I think, I think that's right. I think the really simple SaaS companies you're just going to redo with the AI. I think the more advanced SaaS companies that are going to benefit from AI because they own infrastructure they can use to create AI on top of. So it's going to be a bipartite distribution. Some of these are going to get crushed. Some of these are going to do much better because of AI.

C (10:00)

And that's somebody earlier today, I can't remember who it was but they talked about how many professional services companies and other things are going to be really, really changed and threatened by AI simply because there's just so many things that can do what they do.

B (10:13)

It's a big question for the Indian economy, actually, of the five super giant, obviously service companies, Infosys, et cetera, those guys. And the question is like, how much are you going to need hundreds of thousands of bodies for some of these things? In some cases the high end ones might even be able to do more. But it's going to be tough, I think, for the long tail of those. That's right.

C (10:31)

So you've emphasized, obviously power and energy is the core building block for it. But why is that constraint so underappreciated? How important is it that we win this AI race that we obviously have to win?

B (10:45)

And this is where I think you can go back to Dario, that I think there is something true here where even I don't believe it's the whole existential thing. I don't believe it's a new God or whatever, but I do think you're going to have the equivalent of thousands of very high IQ actors that you have access to if you get there. And if you get there first, it's geopolitically extremely relevant. Right. So if you can imagine the bad guys having that versus using it against you, that's probably not very good for us. I mean, I don't know how much we want to go into warfare here. So a lot of the. I do a lot. But a lot of defense companies we talked about, and we talk about sea and land and air and cyber. And those are like the traditional one, cyber being newer. Like the fifth one you might call is cognitive. And there's all these ways in which you can basically strategically map out all the information about people in a country and figure out how to influence them and how to influence their culture and influence decisions they make. And if you're really, really smart, it turns out you could do that very effectively against another culture. China has an official branch of their military that does this and tries to figure it out if that's going to be powered by a huge number of the smartest entities. That's actually really scary if they get it first. And there's a lot of other versions of this. So there's probably lots of ways of conducting warfare, whether it's in the biological world, whether it's other things that they could do to us, if they have so much more intelligence before we do. So we probably want to have this before Them, that's the thing.

C (12:04)

Yeah, we need to win this race, that's for sure. But so when we try to separate substance from storytelling, which obviously is really hard, but what are the clearest signals that AI is really, truly core to a business and a high ROI rather than just hype. So when somebody, business owners in the audience, they're trying to decide, okay, what do I invest truly in in order to get a good roi, how do they go about doing that?

B (12:28)

You know, So I think so Palantir has had a very strong few years, obviously, and I think a lot of the problems we solved are particularly relevant for this world. And the thing that was, I think the framework that was really helpful there that we're applying is we have what's called our dynamic ontology. And it's an ontology of data, but it's an ontology of processes. And so the lesson from Palantir has been deployed now at a lot of the Fortune 500. And what they do is they'll come in and map out all the data in your organization, which making sure it talks to each other, make sure you know what it is, and you'll map out all the processes that exist in different parts of your organization. So for any company, whether it's a wealth management or alternative investment firm or whatever else you're doing, you're going to have a certain number of processes for everything. And it's going to be a pretty big tree depending if you're working with a client or making a decision or hiring someone or firing someone. And so what you do is you map out all these processes and there's certain things where you still definitely want people involved and you need people to be involved. But there's a lot of these things where the computer can now, the agents can now do most of the work for you, or could do most of the work for you with a quick check for a person or whatever it is. And so what you do is, for me, it's all about productivity and it's all about applying the agents to the areas where you need more productivity, where that speed or where that accuracy is valuable to you, or we're just not having to hire people's viable to you. And I think you're just seeing this productivity go off the charts in a bunch of areas. It really is. It was a question mark two or three years ago. Now, if you talk to anyone who analyzes these things, like, it's definitely real.

C (13:52)

Well, we, you know, you probably don't know this. Joe Lyle is here, obviously, but Most of you probably remember when we talked about it, but we did a five year strategic partnership with Palantir to effectively overlay machine learning and AI across all of our business. And we did it in anticipation of the book launch because we knew that there was going to be probably a lot of leads come through the door as a result of the book. And it was either going to be a trickle or a tsunami. And you can plan for a trickle, it's hard to plan for a tsunami. So we had to be prepared for that. So I've seen actually this back end, if you will work to where and some of you in this room, I'm going to give you a little secret. You were actually communicating for the first three or four emails actually with Palantir's system.

B (14:35)

Not supposed to tell them.

C (14:36)

I know that's okay. They're in the room, they're in the know. But so, you know, for the first three or four times you actually were getting a personalized response tied to you based on the information that we got from you until such time as it made sense to connect with the team. And then it actually recommended based on everything we could find out, you know, about what you told us, we would be able to put you with the right person.

B (14:57)

I'm trying to have it happen with my email. I don't trust it,

C (15:01)

understandably.

B (15:02)

It helps me route it a little bit better.

C (15:04)

Yes, indeed. But what that does though is you talk about efficiency and we talk about it so much in our investment management department. We talk about it so much in all of our departments about you have to try to identify where those efficiencies are. But as Ron, our chief operating officer, wherever he is sitting, will tell you, if you don't know what the process is, you can't automate it.

B (15:21)

Yeah, exactly. So the process discovery stuff's pretty fun actually for when you look at your own org to be, you know, there

C (15:25)

are very few people that would say process discovery recovery is a lot of

B (15:28)

fun, honest about what you're actually doing. Right?

C (15:31)

Yeah, it's scary actually, because you go, we had take, you know, 42 steps to do something that really should take four.

B (15:37)

That's the thing. Sometimes you do change the process after you map.

C (15:40)

Well, exactly. No, we absolutely do. And it's something to where I'm actually going to talk about this because of the fact that you have an amazing podcast that I enjoyed watching in preparation for this called the American Optimist.

B (15:54)

Thank you.

C (15:54)

And he has some, you know, tremendous guests on there. But I have to Say the last one and I'm hoping I pronounced it correctly. Joe Gabbia.

B (16:01)

Yeah.

C (16:02)

He was co founder of Airbnb. Right.

B (16:03)

Airbnb.

C (16:04)

Start Airbnb. He's now full time in the. In the government working for the administration he was part of effectively Doge. And he stayed on after Elon stopped. Right. To continue to make efficient. Just briefly tell the story because most of the people in the room have heard about the mine. Right? They've heard about the mine. When people go to retire from the government, how long it takes and take all the paperwork, literally today. Paperwork. And they put it underground in a mine.

B (16:34)

Yes. Yeah. Speaking of mapping processes, that's one of the more comical ones is the giant under giant underground.

C (16:40)

But to walk through what he told you on the podcast, just because I think it's fascinating.

B (16:43)

It's funny because I'm probably going to remember it about as well as you, but basically it took six months to retire and you didn't really know what you were going to get paid until you got your first check because they couldn't tell you. And the process involved, I mean, by the way, lots of fax machines and other things like that that probably shouldn't be used for this anymore, but they actually had to get the papers printed and stored at the right part of the mine. And I think they even moved them around down there checking a few things too, before you could bring it back up. It just doesn't make any sense at all. It was something that maybe made sense that these things had to be really secure for some reason in the 1950s. But it was like is a totally insane thing. And he's got it all digital now. But the thing he really told me, which makes sense, but I hadn't really thought as much about, is pretty much every one of these processes. We were spending hundreds of millions or billions of dollars on consultants who would just always go years over and way over budget. And their incentive was to go over budget. Incentives are very powerful. So you have a whole consulting class in D.C. where I'm sure some of them are good people, but the incentive is to get it wrong and it'd take a long time. So fired a lot of these places. And a lot of what we did is we created the nonprofit for the government because you're not allowed to hire people for more than 200 something thousand. And there's a lot of great people who make that, but there's a lot of really talented people who'd add something who make a lot more than that, especially in the Alt world. And so these companies, and so a lot of us, Peter Thiel, Elon Musk, Joe and I and others donated a bunch of money to the nonprofit. And there's a law where it lets you do this, where Joe hires these people for the normal amount. The nonprofit pays them extra. And he has about 30 or 40 top tier designers. They're going through and just fixing all of the US Sites, fixing just a ton of things in government right now, making it all work better, which is kind of neat.

C (18:13)

How long did it take them to come up with a solution to digitize that process? That normally took six months.

B (18:19)

I think they figured that in less than like two or three months. Right. It was pretty fast. I mean, the solution they came up with right away, I think they took them a few months to get the people to let them do it. Probably most of the delay is the people allowing you to do things, getting

C (18:30)

the permissions going off memory as well, getting the permissions and getting everything approved and all that took two or three months. The actual work it took basically to design the system and to implement it and turn it live was four days.

B (18:42)

Yeah, that sounds right. Yeah.

C (18:44)

Or days.

B (18:45)

Yeah.

C (18:46)

Okay.

B (18:46)

It's a very down incentive. Yeah, yeah.

C (18:48)

It's a different incentive.

B (18:49)

People don't realize that a lot of things happening in our government these days is off by like 90 or 99% just because it's the incentive that's all wrong. It's. It's a big thing we got to

C (18:58)

confront and be honest about it is shocking. And so now people can actually know what they're going to get paid before they retire. They can actually do a digital process for them to be able to retire. And it just obviously is so much more efficient and, and saves taxpayer money, et cetera, et cetera. Those are some of the things that AI machine learning, in some cases are able to do and solve for. All right, I'm going to take a breath for a second and we're going to switch gears. Okay. And we'll come back to anything that, you know, that we need to cover there again. But you're not just investing in companies. You helped start a university. What compelled you to build the University of Austin from scratch?

B (19:33)

All right, let me offend. Fortunately, it's Houston, so I guess I'm safe here. It's a safe place for me.

C (19:38)

There's people all over the world.

B (19:39)

I know.

C (19:39)

Offend somebody, it's okay.

B (19:41)

Offend some of you. Listen, obviously we hire a lot of people the last 20 years from the top. Ivy League schools and from all sorts of great schools. And I work with a lot of people in government and I think our culture, especially for a lot of our best and brightest, is to teach them to be very, very careful not to have opinions, to be very careful of virtue, signal to not be interested in the truth if the truth might bother some people. And what that means is that in our cities and in our government like we're talking about, you have all these things that are broken and that are wrong, but people just don't think very hard about them because they don't want to offend anyone. And so I think that's actually really dangerous. And to me it's important. Classical virtue is what our civilization was originally built on. The four classical virtues, and all of them are important. I think courage is especially important, especially lacking. So the goal is, can we have one of our top universities in this country to not be run by communists and to teach courage? So.

C (20:41)

So for those of you that don't know, we're very big about freedom, right? We respect everybody's right to believe what they want to believe. Lisa and I have the Freedom foundation of Texas. It's a very important part of our family. So we have the right to agree or disagree. But the most important thing is we have the right and the freedom to be able to talk about it in all settings.

B (21:01)

And that's why the communist thing is important, because there's people on the left and the right and you disagree and it's healthy. But when the communists take over these departments, they don't allow people to disagree with them. They lock people out. If you go to, if you want to study education in the top 100 school, if you want to study sociology, if you want to study, the list goes on. You are going to be getting one point of view. You're going to learn to hate markets, you're going to learn to be suspicious of business people. You're going to learn to be skeptical of America's history. And by the way, I think there's some things we should learn in debate for sure. But I think you need to be able to give them both views and be allowed to debate these things. And that's not at all how Harvard works. It's not how Stanford works, it's not how Princeton, Yale work, which is a big problem. So we want to have Neil Ferguson, my co founder, he's one of the great historians in the world. He taught at Oxford and Harvard and Stanford. And Barry Weiss, my other co founder, created the Free Press, is doing really important work in our country as well. And when we announced this, we had 5,000 professors apply. 5,000 professors apply from a lot of top schools from everywhere. And people don't realize a lot of these professors, who are mostly creative, mostly various sorts of moderate people, they're afraid to speak up where they are. They're terrified. They're attacked by the crazy people at their school. So there's really a need and really a gap here, I think. And it's just tough for people to admit they're all that matters being run by people that are a little bit crazy, but they really are in most cases. So we wanted to have another option.

C (22:16)

That's great. Well, you obviously get to start from zero, which means you get the rare opportunity to design intentionally. So how do you go about building a university? I mean, you've got governance, faculty, curriculum incentives from the ground up. And you have to be thoughtful because you don't want to potentially have it become something that's not consistent with what you're found.

B (22:34)

Yeah, no, no. You need to be able to pursue truth and have an open, open culture and debate. We have a constitution, which is rare through university. We thought that was a good model to copy from the US to have something to appeal to there. And it's actually kind of nice because a lot of times the board's afraid to go along with something controversial. But then you can say, even though this is controversial, we have free speech in our constitution. We're not going to. Not going to say no. We recruited some really, really top students. We have great professors in all sorts of areas. We have a great STEM side. We have a great side of the intellectual foundation of the West. You basically want students to appreciate the great debates of the last thousand years and where our traditions come from. If you ask, I think what is the foundation of our exceptional civilization here? I think there's really three things you have to have that you have to all learn if you want to understand our civilization. One is the classical world and classical virtues in the history there with Rome and Greece and everything. I think two is our Judeo Christian heritage and how Christianity shaped the West. I think that's really critical to appreciate that and a lot of the wisdom from that and how it differentiated us from Rome and from other civilizations. And then three is the Enlightenment and understanding the values of the Enlightenment and understanding how Locke and Hume and all these other people kind of shaped a fertile ground that kind of led to our founders being able to create the constitution. I think if you don't have all Three of those, you don't get America. And I think if you don't appreciate all Those, I think AI is going to change the world dramatically. The next 20, 30 years, a lot of things are gonna be different, but I don't think that means philosophically we don't need to have the same understanding of these values that made us great. Cause if we don't have those, then we're not gonna know how to make sure we keep the things we need to keep.

C (24:09)

As the world's changing, you're a business person. So obviously anytime we get involved in something, you get involved in something. You know, there has to be a definition of success. So, you know, when you look 10 or 20 years out, what does success look like? How do you know that you've declared victory?

B (24:28)

You know, there's, there's. I think the, the longer term goal is to have, is to have massive influence and force the other schools to be more like us. I think the way you do this is you steal away all the very top talent from the other top 20 Ivy Leagues and make it a place where they're much more successful. Yeah,

C (24:47)

we like big, hairy, audacious goals.

B (24:50)

This couldn't have been done 15 years ago. But I have over 100 friends who are like in our talent network who've started the Big Unicorns and who are helping us. And a lot of these guys are already hiring. Peter Till already stole away three of my students. Actually he's hired full time. So even though we only have two classes, which is good for the students, but we have a lot of people helping us and I think we'll see what we can do. But it gets better every year, you guys. If you go to Austin, please come, please come by and check it out.

C (25:11)

Very, very good. We're going to zoom back out to focus on a theme that I mentioned earlier that we're excited about for our personal money. How do AI, defense, space, energy and critical minerals fit together in a strategic landscape? And why do they matter more today than they have in the past?

B (25:28)

I feel like this is like a quiz here. It's tough and to study ahead of time. You know, it's funny, I personally don't do a lot of in the critical minerals area, but they feed a lot of our defense industrial base. And so originally I was a computer scientist and we started Palantir and we weren't really working in manufacturing at all. It wasn't the thing you did in Silicon Valley. It's very expensive. It seems like a weird thing to have to do. And it just turned out that obviously our industrial base has really rotted away. Unfortunately, the last 15 years, we've outsourced a lot of it. And this is especially problematic for defenses, especially even more problematic. So thanks to AI, you have all these autonomous systems. And the best way to fight now in general involves if you watch Ukraine, of course, like millions and millions of drones that coordinate together and things that drive in the land that coordinate. And Sauronics doing giant autonomous ships is also doing thousands of smaller autonomous ships that are weaponized. And so you have to be able to mass produce and then coordinate things with AI. And it turns out we can't mass produce. China has 230 times our shipbuilding capacity. Not 2 times, not 10 times. 230 times. This is a big problem. If there was ever a war right now in the Pacific, you'd exchange a bunch of things, a lot of stuff would all blow up, and then they would build a fleet, and we'd build like a dozen ships, and they'd have a whole few hundred again, that's bad, by the way. It was the opposite in World War II. In World War II, the German battleships were much better than our battleships. They're actually much better engineering. It oftentimes took multiple of our ships to hunt down and destroy one of their ships, but every time they destroy one of our ships, we put five more in the water. We built 18,000 ships in 1943. Right. So. So that's a whole different world. And there's a lot of us kind of going back and studying from the great men, actually, who lived 80 years ago, none of them around anymore, and really taking those lessons. Right.

C (27:13)

Well, another episode of your podcast, and I'm drawing a blank on his first name, but Mikhail or Michael. Last name Michael.

B (27:22)

Emil. Michael built Uber with my friend Travis, and he's now the Undersecretary of Research and Engineering under Dep. Sec. Feinberg. So we were fixing the Pentagon. He let me interview inside the Pentagon.

C (27:31)

It was very cool.

B (27:32)

We made a lot of inappropriate jokes and they didn't keep me.

C (27:36)

But again, fascinating episode. I'd encourage you to watch. But one of the things he talked about so much, obviously, you know, again, you have technology people that understand how to build businesses now working in government to try to fix things and make them much more.

B (27:50)

And sorry to go back to your question, there's lots of things we have to build. Drones is a very big part of it. The engines for drones especially require magnets, require rare earths, as did lots of other things. It's not actually that expensive to do all the rare earths. It just has to do it. So they're finally doing it. It's probably a problem you can solve with 20 or 30 billion dollars. You just have to actually solve it. And the reason why, by the way, we haven't done it is the refining of these things is just really environmentally nasty. So we outsourced it everywhere else and we're just not willing to approve it. I think they're finally going to start approving that again. They just do it on military land. Just get it done.

C (28:22)

Get it done and then hopefully find a way to do it cleaner and more efficient, of course. But he made a shocking comment to me that if we actually had to go to a battle, we would run out of certain things in like four days.

B (28:34)

Yeah, I think, I think there's some stuff like these probably reserves more than that, but I think there's, there's definitely, there's definitely not enough production capacity to stay in the battle over time unless we put it back. And so we are putting it back as fast as we can. It's ironic, in Texas already has delivered hundreds of these smaller vessels. We built two ships that are 180ft long on the Gulf last year. The fastest ships that size built since World War II. We're gonna do 10 this year, 20 or 30 next year. And then on top of that, we're building a whole new facility in Texas, hopefully to 100 exits. So we're going as hard as we can. And the government, listen, this Department of War is running very well. They're very serious. They want to scale and they're no nonsense. They want to have competition and only have the best things win. So it's very exciting. We have seven or eight kind of things that are becoming new defense primes that are, that are crushing it. It's really fun.

C (29:19)

You know, a common question that I get and would love to get your thoughts on it is there's obviously so much of this that ties to AI but are there actual really good opportunities in the world of defense, space, et cetera, that does not require AI in order to be successful?

B (29:35)

Well, it's like.

C (29:36)

Or is everything AI it's kind of

B (29:37)

like everything you do to do it better and smarter. You're going to want to have more dynamic systems. Right. So for example, we have like a new radar system that our friends. We put it in Jordan with the King of Jordan and it was detecting the Iranian shaheeds twice as far away as Lockheed's. And it Turns out that the new AI chips can sync in feptoseconds, which is a thousand nanoseconds, and allows it, even though these antenna are far apart, to be as if they're one giant dome. And it is anyways. But it's like an AI derived thing from the chips and from how they sync and listen, there's stuff that you could do with what SpaceX could do with stuff in space that's not necessarily AI, I guess, but the sensors of course are going to use the AI's stuff to do it.

C (30:14)

So effectively it's going to be part of the engine that makes it more effective.

B (30:17)

Yeah, everything's going to have AI in the engine for sure. And this is why the Primes are so far behind is because they lost a lot of their computer science talent to Silicon Valley, to the innovation world when they kind of all because, you know what happened is you had about 80 top companies in the US, one of the best defense companies in the world. And the Cold War ended and we kind of forced them all to merge because there's going to be less budget and they merged into about eight or nine Primes. And then you had Silicon Valley translate

C (30:40)

for some of the audience going to get it A lot won't.

B (30:43)

Primes would be prime contractors like Lockheed and Raytheon and Northrop and L3, et cetera. And by the way, there's some parts of those companies, they're still very, very good, but there's other things they just can't do relative to the innovation world right now in which. And it's kind of sad because they're still not fully aware of it, like they're losing to us in a bunch of things now. But the people inside are kind of lying to the CEO and saying we're just as good, we're just as good. And it's like they're not aware. No, you're actually not even close in these areas.

C (31:08)

So Anduril, obviously we were early investors alongside Peter and the team at Founders Fund. Trey, they have completely flipped the model to where it used to be, that you go get a contract and you take as long as you possibly can, go over as much as you can, you get paid once you have the contract for a long time. What Anduril did effectively is flipped it around, said, okay, we're going to build it because we know you're going to need it, but we're going to use our money, we're going to go faster, we're going to get it to you and then if you want it, you can have it and you'll have it.

B (31:38)

Now this is the, this is what Palantir and SpaceX did to break through and then Anduril did it was the first defense company because Palantir is Mars software and obviously SpaceX or SpaceX. And then with Anduril's path, there's been about seven or eight others coming through doing very, very well. And the three guys who built Anduril with Palmer all used to work for us at Palantir. So it's all kind of the same DNA, you kind of break through. And similarly, some of the talented surroundings coming from Tesla and SpaceX, because Elon really kind of recreated how you do advanced manufacturing and now we know how to. And we could help those guys help

C (32:08)

us and scale it up and so the primes can catch up. Or are they just.

B (32:11)

No. So the thing that's really hard in general in the innovation world is building like a really top computer science or AI culture. And the reason that's really hard is the best people there, if you're going to hire them, they cost millions of dollars a year. So you really don't want to hire them as a pay. You want them to be your co founders, you want them to be people who own the company with you. And then those people are the ones who kind of know what the other talent is and how good it is and how to nurture it and how to build it. And so to kickstart that from scratch inside of a Lockheed or Aretha where they don't have shares to give, they don't even know what the talent looks like. I think they're lost in that area and I still don't think they've fully realized that, although they're slowly coming to realize this, that they're just going to keep losing in those areas dramatically. And which is why it's so important we do build all the new ones, because China does have a lot of its best engineers working on these things and so it's important we do too.

C (33:05)

That's a fascinating thing because it is a totally different culture than the businesses that have been built for the last 20 or 30 years.

B (33:11)

Yeah, they don't know what they don't know and they don't know how to kickstart it inside. It's just not, I feel badly for them, but it's, it's kind of also ridiculous because it's not that they're corrupt, that they're set up to stop anything else from getting in. So for years, like Both Palantir and SpaceX had to sue in order after we won things and weren't given them because they were so corrupt. How they set it up. I'll give you a quick story. Epirus. We started the year after a underworld. So it's 2018 and we were talking about it earlier. It turns out the new AI chips move really, really fast, like on nanoseconds. And you can get all the power to hit an emitter at once. And you could shoot emp, you shoot a cone of microwave energy and like a 10,000ths of a second all packed in the energy in one and it'll turn off electronics miles away. And this is really useful for swarms of drones that are going to attack you. You can turn them off like a force field type thing. And we figured out how to do this much, much better. So we broke into the contest. They wouldn't let us in the first year we got into the contest an extra L3 and Raytheon and Northrop and these guys in this desert with all the generals and the hardened drones start coming. And long story short, we shot down the hardened military drones nine and a half times farther away for same size, same power. And the generals were applauding and it was a big deal. And then the next year, we basically got almost no revenue. And I found out these guys all went to Congress and they convinced them this is the worst area to fund and turn it off in the ndaa. So I'm like, okay, now I have to play the game back. So what happened was Anduril and Palantir and a bunch of my companies and I spent my time on it too. We built the top team on the Hill and we've now teaching the Senate, we're teaching the Pentagon. Here's how it works. And I'm not saying give me money, I'm saying don't let the bad guys defund the things they're losing on, right? Because it's very, very corrupt how this stuff works. But fortunately it's going the right way.

C (34:48)

There's a lot of moving parts there. And obviously anytime you're dealing with the government, it becomes really interesting. And along those lines, there's this perception that it's tough to invest in defense and space necessarily because of the fact that you don't know that the revenue is going to be there. So how, as an investor, do you get comfortable investing on something that could be five to 10 years before you exit, knowing that administrations can change? How do you get comfortable with that?

B (35:14)

Well, listen, so defense is probably about 15% of what we do. And listen, you're taking a risk when you're building these things. What you're saying is that I believe this should exist. I believe it's really important for the nation this exists. There's a giant gap here we know is necessary and we're going to put enough money to prove that it works and prove we could build it and enough money to go make our argument and make our case in D.C. and you gotta hope for the best. And that's why when these things work, I think Anduril, our first investment in the first round is, I don't know, it's like what, 300x400x. And they work, they work really well. But there's a lot of risk to them as well. So that's the nature of it. But I think if America is going to be successful, it has to have these things. And so you kind of have to take a jump.

C (35:59)

And that's where the longer term view is. Administrations will change as I showed on the chart there. You know, if you look at it over a long term horizon.

B (36:07)

But defense, defense is, defense is the least partisan thing in Congress. Like when I go to speak to the House and Senate Armed Services Committee and I happen you guys can tell I'm complaining about commies, I'm obviously on the right a little bit, but there's staffers, you know, there's, there's staffers on the left that are taking notes and that are getting along and there's, and there's, and there's senators and there's people. Adam Smith, who's you know, from Seattle, very left place and he's a Democrat and he's the chairman on the House side of the minority and he's a very rational guy. Fortunately, America, not everything is partisan. We all want our country to be functional and not to waste money on things and things that work. So yeah, I'm sure some of them will want to attack some of us at some point. But I don't think it's partisan to have defense be functional.

C (36:47)

No, the reality is that all of them, all of us, everyone are unified on one thing is that we want to keep our freedom and we want

B (36:55)

to be able to make, we have to keep this in nonpartisan. One of my mentors was Secretary Shultz, who the secretary, four different, different administrations. And he always tell me, say Joe, don't say bipartisan, say nonpartisan because it really, it really is not political like this. Some of these things just shouldn't be Political.

C (37:09)

That's a great way to put it. So from your experience with Palantir and now as an investor, what actually distinguishes a defense or a space grade company from a commercial business other than the fact that they simply sell at the government? How do you look at them as being truly world class in what they do?

B (37:24)

I love all the defense. It's, it's a lot harder. It's a lot harder. So to build a great company in defense, which is funny, we call it defense because it's Department of War now, but it is defense also. You have to basically build two companies. You're building an A plus company that's this way ahead as like a 10x product that's better than everyone else that's needed and then you're building a great team in D.C. and it's just very annoying. But this is this, you just have to be a realist. Like it turns out that you just need to be able to have a team that could talk to the right people in Congress, talk to the right people in the Pentagon, make your case, defend yourself against the existing guys who are there, who are going to be jealous and are going to attack you. And so you basically need to build two things well at once. And I didn't know this when we started Palantir. I think our ground game in D.C. was probably terrible for the first several years and they obviously learned over time because we had the best product but we didn't have the best ground game. So you have to build both. So it's a more expensive company to build. It's a higher bar. One of the things Peter Thiel and Palmer pointed out that pretty much all of these companies that have been successful so far in a big way have been started by billionaires because it just, just having the access and the wealth to do it is pretty tough. So it's not, it's not something that I think everyone should just be investing tons of things and it's really tough and mostly invest other areas, but it's really important so we do it well.

C (38:34)

And that goes back to the point we've talked about in the last day and a half. So much of you have to understand what is actually underneath the hood as to what success looks like, why it's going to be successful or what it's not, and why tourists should not apply in many, many of these areas. They're not going to be successful.

B (38:49)

There's not going to be more than a dozen new really big defense defense companies is the nature of it. And so it's not, it's a, it's a non trivial thing to do. You can't just have a good idea.

C (38:56)

Got it. You can't just have smart tech people.

B (38:58)

They also, even, even that's not enough. It can't just be a bunch of smart 21 year olds. Like, like if I didn't have Peter Thiel as a co founder of Palantir, we didn't weren't able to bring in a bunch of other people with Alex and Peter, it wouldn't have worked.

C (39:08)

Got it. So different path back to Palantir a little bit. But first I want to ask this. You know, when you look at an AI native company, when you look at, you know, what you described in every business, in order to manage it, you have to be able to measure it. And if you're going to measure it, you have to be able to quantify it. So what is a metric that our, you know, that our folks can really use to analyze success or failure of a company and know if they're on track or off track?

B (39:34)

What are metrics? You know, I think, I think it's a really good question. But I think the problem with some of these emerging technologies is that it's not beyond just a very basic traditional business metrics. It's not really clear how else to measure these things. So I mean, I'll give you an example of a metric. We led the series A in a company called Bedrock last year. And these are a bunch of guys from Waymo. Waymo is that company that drives cars from Google.

A (40:00)

Right.

B (40:01)

With no one in on autonomous driving. And these guys were building autonomous excavators and it didn't work yet when we invested, but they'd gotten a bunch of groups to give them data and they hired a bunch of talented people. And the things I cared about for the next year were pretty simple. It was one, the talent they're getting because we map out talent, are they getting the very, very best people? And two, could they actually get this to work? Could it do certain tasks? We defined and we defined the task with them and lo and behold, six months later it was doing those tasks. They made a bunch of these hires and a bunch of other top funds paid up, you know, seven times the valuation, put a few hundred million more in and so there's no exact metric they were looking for, but it worked. And the people are going there. So it's like it's a weird art in that sense.

C (40:44)

Well, but I mean, you know, for anything that's a developing technology, there's Got to be steps along the process in order to make it fully fun.

B (40:50)

Yeah. And you kind of give and you kind of like give yourself challenges that you define each time for. Can it do this, can it do that? Is it working like, like another one? I'll give you. So Cognition, right, is a pretty interesting company. I don't know if you guys are, are involved there. You probably are directly through the funds obviously you invest in. And so this, we were talking about this one earlier too, that there's a guy who worked for me at Addepar, which is a technology powers, a lot of these things in wealth management that we use. And he had won the Global Programming Competition three years in a row in 17, 18, 19. So like the king of the nerds and really, really smart guy. And he three years ago gets a bunch of these other guys together for a new company and he has like 20 of these gold and silver medalists, like all the best programmers and math guys in the world all together. And they started figuring out how to use new AI to help enterprises code. Right. And it turns out this is one of the, I think it's like the sixth or seventh fastest growing company in the world. Now it's three years later. It'll do well over a billion dollars revenue this year. And the goal was to be very useful in creating a colleague. How you define that? We kept changing how to define it because we didn't know. But, but, but people are paying for

C (41:56)

it, you know, so let's take that and build upon it a little bit to where if somebody's going to apply AI in their business, you know, there's, it's, it's tough to measure sometimes. How do you measure it when you're looking at your companies that are utilizing AI for productivity?

B (42:10)

Yeah, I mean, I mean, I mean it really is like, are how much. Are you able to either create more revenue for the same amount of costs or for low costs, or are you able to save costs? It's, it's one of those. And it's, it's. And it's usually, actually, I think usually it's create more revenue. Surprisingly, you'd think it was usually just cutting things. In most of my experiences, it's about, it's about leveraging people to do more and create more because it turns out there's always more valuable things to do, but everyone's too busy. But in some cases there's also our processes. You define that you need a certain number of people for, and now you don't need as many people. And there's some areas obviously customer support. You probably do end up cutting people right now and that's fine. That's what productivity is. It's how everyone gets wealthier, is that you do more with less.

C (42:50)

So average revenue per team member is a metric that somebody could look at. I think it's actually improving from a, you know, a particular sector or particular.

B (42:57)

Yeah. Can it. Can I cut people here and do more or can I do more without hiring? One of our. I'll give you a cool story we mentioned. Some SaaS companies are replacing with software, but some SaaS companies like you're actually making better. So one of, one of the ones we helped build a decade ago is the leading enterprise software for title insurance. Turns out there's lots of title insurance offices Qualia. And right now, because rates are coming down, these title insurers are going to have to have a lot more business probably. And the CEO used to work for me. He just launched this thing there which is a new agent that does most of the work of a title insurance professional. And he's gone from 0 to 10 million revenue in three months, which never happens with a new product. And there's all these different title insurance companies that are realizing they don't have to hire as things ramp up, which is obviously very valuable to them.

C (43:42)

That's, that's an example of an industry that's ripe to be disrupted.

B (43:46)

Well, that whole industry probably shouldn't exist, but that's not my problem.

C (43:50)

That is, that is probably true as well. So let's talk about that maybe in a slightly different way. But you know what's the most overrated moat out there? People think there's this big moat.

B (44:00)

These policy moats are so annoying. They're so annoying. Policy moats, policy modes. When you pass a law to stop your competition. Right. And actually I do think this is like a new thing you're going to see and you are seeing a lot more, which is a very good thing is entrepreneurs back who are already been successful understanding an industry, knowing it should not work this way and then trying to build something that both solves it and tries to change the policy at the same time. And you're going to see a lot more of that. And I'll, I'll give you an obvious one is primary care. I think health care. AI should be able to. You should. If there's, there's like a bunch of types of things where you should have a doctor check it, but there's a bunch of types of things where it's very Clear what's going on and you should be able to get a very safe prescription without having to go wait for half an hour and do a copay three weeks from now. You should just be able to do it. Thank you. And by the way, I think this is really important for our civilization if we can bring down the cost of healthcare and make it better for all of us. And so there's things like that where we are going to have to go to war against scope of practice rules. We are going to have to work with our friends, you know, in hhs and a lot of people, a lot of us are working on it. And I think these moats that protect the way it works right now that are policy modes like we have to go against this.

C (45:09)

Got it. So on that basis, and it may be what you just said, but I'm curious if it would be something different, one regulation that you would rewrite tomorrow.

B (45:19)

Oh, there's so many. I actually have a company, my wife and I. My wife and I had a whole nonprofit dedicated to the regulatory state. And I think if I can be meta about it, like the thing you really need is you need a better garbage collection mechanism. So, so basically if I could like rewrite the regulation, it'd be the regulation on regulations, which is that regulations should expire and they should have to be data driven. And when they expire, you have to justify themselves.

C (45:46)

Well, it's funny because the fact that, you know, most people know I'm on the State of Texas Pension Review Board and you had the sunset provision, but

B (45:53)

it doesn't work because you never turn anything off fully because it's like you have to turn the whole department off.

C (45:57)

That's right.

B (45:57)

The rule in Texas is you had to turn the whole department off if you want, but you're never going to vote to turn the whole department off. So it's not. You should do the other. The sub things need to justify themselves.

C (46:06)

So effectively take that and apply it on specific roles and specific things.

B (46:10)

And technology can help us do this, by the way, because it would be. This would be totally impractical in the 1950s, probably anyway, but it'd be a lot more practical now with technology to help review all these things. And so I think garbage collection for regulation.

C (46:22)

Garbage collection for regulation.

B (46:23)

That's a cs. That's a CS term, but yeah.

C (46:26)

All right, I'm going to come back to Palantir in a second. But what is the best book that shaped your worldview?

B (46:33)

God. You know, I really like Matt Ridley's books and he wrote one called Evolution of Everything that was really fun about how evolutionary systems apply to everything along those lines, actually, since it's a similar idea. Carol Quigley was a great historian that wrote the evolution of civilizations, and he maps out all civilizations through seven different stages. And it's like the way things kind of mix and then form and then they grow, and then the interests that were helping it grow take over and serve their own ends, and it decay, and then they get invaded. And it's really interesting because the way civilizations kind of rise and fall is very similar to how institutions and companies rise and fall too. And I thought there was a lot of wisdom in that. And you can kind of see a lot in our parts of our civilization where it gets to, like, the part where the special interests are breaking it, and then there's. You can either fix that and grow again, or you can decay. And I think there's some really important battles tied to that that we're all facing right now.

C (47:22)

No, and it's the life cycle of business, life cycle of companies, life cycle of countries as well. You know, when you mentioned Palantir, obviously, it's one of the most important companies right now and kind of leading this.

B (47:35)

It's bigger than I thought it would be, to be honest. It's pretty cool.

C (47:37)

Well, talk to me about it. What was the motivation to start? What was the. The problem you were trying to solve? What was the why behind.

B (47:43)

Yeah, no. So I was at PayPal as a kid while I was at Stanford, and the Chinese and Russian mafia were stealing all of our money. And there was like, eight competitors. And we're the only ones who survived because we figured out how to build tools. We actually took. We had a customer service department, and they took like, the top 100 people there to be investigators with tools. And we cut down the fraud by enough to make it profitable. Sold to eBay for 1.5 billion, which used to be a lot of money. At 2, 2002 or so used to

C (48:09)

be a lot of money.

B (48:10)

It's a different world. And Elon and Peter and everyone took the money and kept going. And 911 happened right around that time. And because we had been helping the FBI and the duh, it wasn't dhs, it was a secret service. And the FBI at the time were the ones because DHS didn't exist yet, were the ones that were kind of helping us investigate these things and try to arrest the bad guys. And after 9 11, these guys got a ton of money to fix their systems and go do Better things. And similar to what I talked about before, they spent tens of billions of dollars just on stuff that was 20, 30 years behind silicon Valley. This wasn't a unique opinion to me. It was obvious they just didn't have access to everything that was going on in the 1990s and 2000s in the Valley. And so we were pretty horrified that our country was under attack and there are these terrorists who wanted to hurt us. And not only were we not catching the bad guys, they were building the systems in such a way that it had no protection at all for government. People just go and see all of your data and spy on you. And I don't really trust the government to do that without checks and balances. And so there's really a double mission is one is to help make sure we killed all the bad guys that wanted to hurt us, and two is civil liberties. And to make a long story short, before I left the company, we actually did partner with obviously all the different branches of government. Their job was to do this. And we hunted down and eliminated 9,000 terrorists.

C (49:26)

Let that sink in for just a second. Wow. And obviously it's transitioned into significant amounts of activity not only with the government, but also with companies.

B (49:37)

Yeah, it's actually kind of neat. This has been true in America's past as well, that when you solve really hard mission driven problems for something you really care about for the country and you get people working really hard on it, a lot of those problems we solved were then very relevant and valuable outside of government as well. And so I started the commercial side of the company, first in finance, then in other areas, you know, a few years into it, partially because we wanted to also hedge in case government would never pay us enough that we could keep the company alive. So we were pretty nervous at the time. And then here's what really happened that I didn't expect. So that was enough to get it to be a successful company worth 20, 30 billion dollars, which is obviously a huge success. When I first started the company, we'd give people offers and give them shares. And you'd say, here's what your shares would be worth based on the outcome. Outcomes. And the highest outcome I wrote was like 5 billion. Here's what your shares might be worth. And everyone's like, Joe, you can't say it's going to be worth 5 billion. It's totally unrealistic. Which is fun. So it got 23 for me. That's a lot of money. And then it turns out this ontology stuff where we map out Processes and map out the data and organize it. Makes it really easy to apply AI on top. So Palantir had a whole nother kind of growth wave the last five years from AI that took it from, to being worth 10 or 15 times more, which I never expected is awesome.

C (50:51)

See it? No, it's, it's, it's been a moonshot for sure. So in most of what we've talked about, you know, it's technology and how it's used but you know, there's a core of people behind it that are creating it. So what types of teams and leaders consistently win in the domains we've talked about? What are the best operators actually look like? What, what differentiates them? And what do you look for when you're hiring?

B (51:14)

Yeah, gosh, well, the hardest part is just the really deep tech culture. So you have to have, it's like getting athletes for, for your, for your baseball team or football team or whatever is. You have to have the people who are the best in the world at this. And there's only a small number of those people. They're the kind of people who win math, chess and physics contests and who, you know, come from the top, usually top computer science departments, although there's something come from you never expect. They just came from. They'll know where they're really smart and you sometimes you'll find them from another company and pull them out. So first of all you have to have the top technical culture and then, and then you want. You mean courageous or strategic leadership. It's actually interesting. Most of the people I've worked for who've been the most successful or I partner with like whether it's Peter Thiel or Elon Musk or even Charles Koch, are actually philosophers. And I think you want people who are very thoughtful about the world and very often the best co founder pairs. You have some people who are just deeply, deeply technical and attract those best athletes alongside like a courageous kind of philosophically minded person. And I think that combination I've seen work really well. The courage is important. You have to be willing, you have to be a little bit like of a difficult, not difficult person overall, but difficult intellectually where you're constantly challenging things, you're constantly willing to think for yourself and not, you know, not go along with what other people think because you have to rethink it for yourself.

C (52:24)

I mean it's truly rethinking the way that everything is being done.

B (52:28)

Yeah, you have to, you have to be world. Exactly. And I think, I think that, I think that bias is good. There's probably, like, some things that Palantir we didn't need to recreate. We recreated like, everything's done. And then later we're like, oh, that's why sales is done this way. We didn't know what you were doing. So it's like there's some things you really didn't need to recreate. But I think it's a better bias to recreate as much as possible. So I think, I think it's a very anti business school bias where rather than learning how things are done, you have such smart people that they just try to recreate everything. And then, of course, you do want some mentorship to figure some things out.

C (52:54)

Solving the unsolvable, as we like to call it.

B (52:56)

Yeah.

C (52:57)

So I'm going to end with this. Your podcast is called the American Optimist.

B (53:00)

Yes.

C (53:01)

So I'll frame the closing question as this. Give us one sentence on what you're most optimistic about in the next decade and one sentence on what we cannot afford to get wrong.

B (53:14)

Well, I'm most optimistic about America being a successful, prosperous country because there's every one of our challenges that I see that we face. There's really good answers that come from the values that created America and from the innovation world to confront those. And the thing I would say, and there's so much we can't afford to get wrong in terms of respecting those values. But one big part of that politically, is we can't afford not to fix the cost of living and then fix healthcare, education, you know, housing, because those are areas we know the answers to. But if we don't fix those things, I think the populace are going to tear things down. So I really hope we get those right.

C (53:46)

Thank you for joining us today. And to learn more, you can head over to Tony's YouTube page where you can download previous episodes and see everything we've covered in this series. You can go to holygirlofinvesting.com as well to learn more about the book that Tony and I wrote together and learn a lot more about other sectors and things that you can expect to hear in future episodes of the Holy Grail Investing. And don't forget, you can also go to Apple Podcasts and Spotify to download as well. We look forward to seeing you next time.