Transcription
We started Founders Fund in 2005. So, in different ways, I've been at this, and I'd done some angel investing before that. But I've been doing this for 20-plus years.
And the area where we've probably had the least success from a returns basis is software. The two worst areas are, um, are educational software-related companies and healthcare IT companies. And, um, and, like education, we felt was so hard, we, we didn't even, uh, we did very little. Healthcare, we tried a few more things, but, um, it's, and, and, and the rough version of what's gone wrong in the healthcare sector, we've done so little in education, I feel I can't even comment on it. But, but healthcare, what's, what's roughly gone wrong is, um, you have all sorts of things that sound like a pretty good plan for reforming the system, and it's, it's this one angle on the pricing model you need to change. But then there are 10 different ways you can do it, and it's these complicated, uh, ways you have to get the payments to shift from the insurance companies or the government or, uh, the hospitals, which are very clunky. If you can just get consumers or doctors to change, that might be easier, but they, they typically are not the people who pay. And so there sort of are, there's this, you know, very clunky, messed-up system. Lots of people try, try to do it.
My, my sociological model is always that I'm, I'm deeply skeptical of, um, of investments where a lot of VCs would go to some silly VC networking event and, uh, want to brag about it. And I, I feel there's a lot of that with the healthcare education startups. Um, this is like the green, this like the green tech stuff, too. It was too popular. The green, green tech stuff. Yeah. I mean, at some point, it, it loses money, and then it's not popular anymore. But, but while it lasts, it's very popular because there's, there's no fast-term accountability in venture capital. You know, sometimes you have momentum, and the valuations go up, and that, that does give you a feel whether things are working. But in practice, uh, you often, you know, it's 5, 10 years later you get liquidity on, on these businesses, or even longer. It, it takes incredibly long, and so people tend to overindex on these sociological factors and what's, what's popular. And that's, and so there was, there was a, there's a way that, uh, the healthcare education sectors, even though they're very broken, um, people have these often too conventional ideas on how to fix them. And, um, and so, yeah, it's, I know, you know, another trend where I've, I've tried investing some over, over the last decade is, you know, is, are all these marijuana-related businesses. And it's like, yeah, it's obviously like a big trend, um, but it's, it's an obvious trend, and then it turned out to be extremely hard to figure out what, what to do with. Yeah, I said some of the things that are too popular and too obvious, it's just you have to have a totally different take on it, or it's not going to work. Uh, yeah, you, like if you have an unpopular take on it, that, that, that might work.
So at the time, I was working, I was working for, you were working on the hedge fund. You were invested in Facebook. You're doing other investing, and we decided to really go all in and spend a lot of time on Palantir. Why, why was that the right choice then? As tech, but go more specifically, like what, what was, what was your high level around that at that time? How were you thinking about it? What, what gap were we going after? It was very broadly government, then it was within that, it was defense, uh, more narrowly, it was sort of the, the, um, intelligence community, that sort of adjacent, you know, defense adjacent. Uh, I, you know, it was, it was a couple of years after 9/11. We're sort of starting it in 2003, 2004. And, uh, there were all these ways that, uh, I felt that we didn't, you know, that we, that we these very low-tech and very privacy-violating solutions for the terrorism problem. You know, it was, it was annoying that you had to take off your shoes at every airport, which was like, just not security, but it was this incredibly intrusive security theater. And, uh, and then, yeah, the question was, could you do something where you'd have, you know, way more security with way fewer, um, you know, technology is doing more with less. And so the, the terrorism context was, could we have more security against terrorists with less intrusive, um, civil liberties violations? And then, um, and then, and then there was broadly this question, you know, was, was there some way to, to reset all these things in the government? And, uh, it was a black box. And, you know, we did not know whether places like the CIA and NSA and places like that had, you know, some super-duper computer working in the background that was catching all the terrorists. Um, we sort of suspected they didn't. Um, and, um, and then it took, it took a while to figure that out, but, uh, I, I think our suspicions were probably correct.
Yeah, we saw them doing a lot of things that made us suspect that they had actually gotten way behind Silicon Valley. And there were, I mean, there were details that, you know, it was hard to say, you know, with our, I remember our first trip to Langley in, uh, in 2005, we got Intel to invest, and you and I were there. And I think it took us about a day or two to process, but, you know, at the end of the trip, it was like, do you want to come? Do you want to buy some stuff from the CIA gift shop? And it was like, you can get a windbreaker and a CIA cap, coins. They were kind of cool. I think they were, half of them were made in China, which is, you know. But, but, um, and then of course, there was probably some, you know, there's some steelman version where the, it was just a decoy, and there was really a lot of stuff going on in the background. But, do you remember the doors all had these like eye detectors on each of the doors that you're supposed to scan your eye to get in, but none of them worked. So, there, so wherever you go around, there'd be like the eye detector, but they were broken. They'd turn them all off, which I thought was kind of like a, kind of a typical, I don't know, you know. There were, yeah, there all these, all all these plays. And so in that context, you, you, you'd think, okay, there's, there's some very big problem that can be, can be solved. It, you know, in some ways, it turned out to be more difficult than, uh, than one would have thought. You know, there obviously are all these pro, all these sectors where you have that are very, very broken, where, um, you know, education, healthcare, all, all these sort of deeply broken sectors where the solution, um, you know, there's some obvious solutions, and the fact they have not been implemented tells you there's some big barriers to they're set up not to allow you to break them.
This is kind of what I think of the Palantir version. You know, there was, there was this, I forget what the acronym stands for. There was this, uh, software program. This was once we moved from intelligence community into the broader military called DIAS, which was just sort of this, um, in theory, was a system we were supposed to figure out ways to integrate the data across all these, uh, units in the field and headquarters and, you know, integrate what was going on, which is incredibly important thing to do. In practice, uh, it was a sort of a clunky consultant billion-dollar boondoggle that didn't work. And, uh, and we naturally thought we would replace it in six months. And it took, it took something like 10 years. We finally replaced it in 2019. And, you know, we had to sue the army, which, and, you know, we had all our, all our consultants quit and said, you know, um, this is just the most ridiculous thing. You'll never get a, you can't sue your customer. And, um, you definitely can't sue the army. And, uh, not only will you not get them as a customer, you will not get anybody else in DC. You'll just be blackballed by everybody. And then, but then, you know, um, and it was under, we, we, it was in 2016 that we sued them. It was under a 1994 law where, um, you know, the military was supposed to have an open procurement process. And then when you, uh, and then what you can do when you sue somebody is you, you get discovery, and you discover, and you sort of discover what's going on in the sausage-making factory that no one's investigated for 22 years. And there's a lot of stuff. And there were sort of, you know, there were internal reports that Palantir worked better, which didn't matter because obviously nobody can assess the software. But then more scandalously, there were, um, you know, there were, um, orders to suppress the reports. There were emails that people had to hide the findings, and there was a cover-up. And this, this did not sound like the open procurement process. And, uh, and then, you know, and you, you won at the court level, then you won at the district, at the DC circuit level. And, uh, you know, you were forced to, they were forced to reopen the process. And then, you know, after a decade, we, we sort of got the contract. So, yeah, so there was something that was doable, but it was, it was not as straightforward as it, as it sounded.
It feels like there's these, and SpaceX had to sue as well, of course. It feels like there's these parts of our society that are set up to just maintain their broken nature, and you have to just punch really, really, really hard through them. And it's very rare people do that, you know. Yeah. SpaceX was the first successful template to do it. Uh, and, and there's, you know, there's obviously something like this. You know, the university system seems as broken or or more than the military procurement process, and then especially with the, you know, crediting agencies and all these things where where they just, uh, are designed to prevent newcomers. And the, it's, it's like if you actually had a newcomer, they would just eat the field. They try to crush you. They try to crush you. It took us thousands of pages of regulation. But I think there was an Elon metaphor he told me once, and this is like, I don't remember when this was, like maybe 2010, 2011. And it was, it was sort of like, uh, there were, there were these three or four aerospace companies in the US that were like really badly managed and badly run. And it was sort of they were these very fat cows on this grassy enclosed field. And if you ever got in there, there was just going to be a bloodbath. Oh, yeah. So, it's kind of, it's like he's like this, this carnivore, and there's a big fence, and you got to climb this really big fence. Eventually, you got in there.