If you took the normal path without AI and the things we were building and knew we needed to build, we're taking the 2030s and 2040s and packing them into the next two or three years because so much is happening right now. And that means we're taking the 2050s and 2060s and going to put them into the end of the decade and the start of the next decade. It's just like so much cool stuff.
All right, Joe, I've been really looking forward to doing this with you. You're one of the most interesting people to me and I love a lot of the work that you've done and so really excited to chat and learn with you.
Thanks for having me, Jack.
One of the things I want to start with is I kind of put you in this category of people who are both investors and founders like the real way. You know, a lot of people say that, but you really have done it. You've started many companies. You've started companies that you then become chairman, some you work for a long time on. Obviously there's Palantir, but there's many others. Then you also have your own VC firm. I'm just sort of curious like what's the mindset that you approach company building, company investing with? Like where do you start when you think about doing your work?
Well, I think starting a second or a third company is fundamentally an insane act after you've been successful because you already have the money you need. So what are you doing? Because these things are so hard. You almost have to convince yourself every time that it's not going to be as hard as last time even though you know it is.
I mean I did it once then I remembered. I'm like that's too hard.
It's really, really hard. I really... Well, I did like two where I was CEO of Palantir and Addepar. And then I did a few where I was chairman and really in the weeds with OpenGov. I was with Zach last night who was my CEO at OpenGov who did very, very well. We sold it for $1.8 billion after like 14 years. And it was like paddling through the dirt. I was saying it's like so freaking hard especially compared to these things today.
Like an AI founder who's just at 100 million in no time. We were next to this kid last night who I'm backing who's like, 'Joe, I'm not just doubling, I'm tripling revenue this quarter.' And it was so hard getting used to be like eight years of hard work where basically everything went right.
Yeah. Now like one of these kids who work for me has a billion in revenue. You're like screw these kids. I love it. I'm investing in them. They're great kids. But it's like they're great. You're adults. I should call them. But no, listen. I think building is really, really hard. There are just gaps in the world that I see. It's just how my mind works and there's like this is a gap and there's like people who are hurting because of this gap. There's things that are broken because of this gap and it's so broken and no one else is fixing it and I have all these really smart friends. Let's go fix it. And so I just can't help that instinct to just go fix things. We say at 8VC like the world is broken. Let's fix it. That's just my energy. So if no one else is going to do it, fine. Let's get people together and do it. But it's like the last option. You much rather back a really competent person who they themselves are already all in.
Yeah. Do you think not enough investors have this instinct? Like do you think it's right that most people don't do it or do you think there's actually more investors who can and should understand this is an interesting setup just like get some balls moving and then you can attract people to it?
I think if you haven't yourself gone all in and built a startup and gone through the pain of that over years, then you probably have no idea how to actually build a startup and you just shouldn't be doing it. Yeah, I think the first time you do it and probably the second time you should just be 100% all in and you shouldn't be distracted. So I think I'm like a really bad example for people like you actually shouldn't try to do these things.
But it's funny because even among investors who were former founders or who have done it before I think many just don't take that as one of the options of actually I could just get this thing started, work with somebody, build around it.
Takes a lot of energy. It's like you have to pour like a piece of your soul into something to really make it. Not that I do this right every time, like I make new mistakes all the time, of course, but like every once in a while. And frankly, like the last several times I've been involved, it's not really mine. It's really like I'm helping assemble the founding team. I'm helping bring together lots of talent. I'm helping bring together the mission. I'm giving them enough upside that it's theirs and then so that's a little bit different than doing it yourself, obviously. So, it's going to be a spectrum.
Yeah, totally. I'd be curious to hear a little bit about how you're feeling in this general moment. You know, we were just talking about how hard it used to be to build companies to 100 million of revenue. Now these AI companies are going so fast. It's like so fast. You strike me as like kind of in the sort of Peter Thiel contrarian-esque mindset of things. And one of the things that I wonder is like is it uncomfortable in any way when the whole world sees something and you're like we might be in that quadrant of the 2 by 2 that is both like contrarian and right? Like are you kind of in the mindset of like AI is actually the thing most people have it right and so I just need to go do it better than everybody else or do you sort of take the view of well when everybody's here I should actually be spending my time here because there's going to be a new wave by 2030? You could still make like a bunch of like sub quadrants and they could be wrong in some things and you could be right about some things but you're all inside the big AI is right thing, right? It's not like you just have this like super simple model.
But yes, I mean obviously this stuff is working in lots of ways in AI. I think people are probably not focused enough on productivity as like the kind of grail you should be going after. I think they're not focused enough on the moats created just by the very most talented technologists. So I think going after just the very, very best teams and going after problems that are going to lead to higher productivity and then there's just all these nuances for how big industries work and I've been obsessed with a lot of these large industries for a long time. So I think we have advantages there. But yeah, I mean listen, it is non-consensus that I'm just really bullish on this stuff right now. It's working. What are you going to say? It's all working and let's go after the very best of it and help it.
I can't remember where I heard this, but somebody was saying like there is a form of contrarianism even among good ideas, which is that you can like believe harder than everybody else.
You believe more, you can understand more of the nuances of what's working. So, listen, I've got really lucky the last few years because I get no credit for like Palantir figuring out what it did the last five years. The only credit I get is that I actually originally built the company. I did hire a bunch of those people who were in charge, but it's them now who've done it recently. But like the stuff that we figured out at Palantir and the stuff they figured out at Palantir turned out to be relevant to a ton of other AI companies. So it's like kind of the center of the AI world in terms of what's working in many ways was right around Palantir. And so that's just a huge advantage for us and I think we understand a lot of the nuances like how do you apply a workflow ontology to this area? How do you apply like the FTE motion to this thing and go deeper on this with a SaaS plus FTE?
It is interesting. The FTE thing is just everywhere now. I mean it's almost mimetic but it's true.
And Palantir figured out that the way you do these things in the AI world and you just have to use that.
When you say you know Palantir figured out a lot of like the moats of like how AI is working, like what else besides that FTE model? What about like the technology itself?
Well, it's like how do you extract like the ontology of the workflows and how do you figure out where to apply LLMs and where not to apply LLMs to and then what do you have to build for it to work and there's just a lot of stuff that people have got right.
When you talk about like the productivity stuff being like sort of underexplored and people don't appreciate how important it is. Do you mean like areas of like productivity software? Are you talking about like making individual professional workers more productive every day? Is it both? Like what are you thinking about?
No, it's definitely both. It's also just like how should this industry work in 2035? Like what does this industry look like today and what does it look like if it's doing everything that's possible that we know is likely to become possible? What matters 2035 seems like so far away now.
I can't even think about like May.
No, I know. It's like we're pulling if you took the normal path without AI and the things we were building and knew we needed to build. We're like taking the 2030s and 2040s and we're like packing them into the next two or three years because so much is happening right now and that means we're taking the 2050s and 2060s and going to put them into the end of the decade and the start of the next decade. It's just like there's so much cool stuff. Like my favorite one this week, every week there's something. My favorite one this week is I was with the guy who runs Joby Aviation. I was actually doing an episode with him and it's just like it turns out there's just going to be all these ways that jets and airplanes are going to be like 10 or 20 times as efficient in the next decade. And like he's launching in the next year obviously these current flying cars but actually there's like massive things that would take him three months before you do it in an afternoon now with the aerodynamics. It's crazy.
Yeah. By the way, it's funny because I feel like before AI took off or like right at the beginning like right after ChatGPT, so much of the conversation felt like it was around like AGI, ASI and like is this like sci-fi Terminator taking over. And then as everything has like become super commercialized and these are great businesses and it's proliferating the economy in a weird way feels less sci-fi now and more just like technology things to build. There's so much and people underestimate like really smart people like he was saying like his right hand had just like a list of like 50 things he wish he could work on and build and he just wasn't the resources wasn't the ability. Now we're going to get through it all. There's just so much to build, it's working. What's interesting is that that mindset though I don't think everybody has it, you know, and like you can see this actually in high schoolers or something where some percentage of high schoolers I think are like becoming like putting on an Iron Man suit with AI, but you can also use AI to just sort of like skate by and so you kind of see these like big skill gaps.
Well, it's like so sci-fi, it's a heroin, right? It's like it's really scary. Like you can use AI to beat your heroin. You can use AI to not have to do much at all. You can use AI to lose all your agency. There's like a very negative version of it and then there's like this extremely positive version of it. And you know, this is a cultural thing. We got to figure out how do we get people to be more agentic themselves. How do we get people to engage and to choose to get dopamine from iterating and winning and not from sitting there and attaching it to their brains. It's like TikTok is evil. And I hate that X is having to copy it because the financial incentive for social media is evil right now.
Yeah. No, I mean you're saying X have, you know, you on X, you go into a video, you start you scroll, you're like, you're on TikTok again.
I've blocked TikTok for years. Partially because of the CCP and stuff, but partially because also I don't have the willpower not to freaking scroll these things. So I block it. I block Instagram. I can't deal with it. And like I'm an investor with Elon, X, I guess I mean SpaceX now, so maybe it doesn't matter. But it's like so I have it, you know, and also I'm addicted to like commenting my stupid views online. I'm arguing with Balaji today. Whatever. He's a good guy, but I got angry at him. But you go on this thing and it gets addicted to you late at night, too. It's horrible. I don't know. I'm gonna have to turn it off. It's terrible.
Yeah. I feel like tech probably lost a decent amount of trust with the public as a result of social media and people being like, you know, they don't care. Like the tech industry is just here to like monetize eyeballs and make money off of us.
You know what's ironic? It's like generally market forces are really good in our society. Like they cause us to have to work hard to solve problems. They're going to bring down the cost of so many areas if we can get rid of crony capitalism. In the media itself, like with newspapers and in social media, market forces are terrible.
Yeah. Because it's an attention economy. Your incentive is to like distract people and just like polarize them. So, it's just like that is like the worst area of society for tech. And unfortunately, that's what we're known by.
It's really bad. It's also, you know, I'm not like some I'm not anti-journalist. I'm not against mainstream media or social media particularly, but you know, it's like even with journalists who I really like, it's not their fault always, but they might write a great article and then somebody, some editor puts a crazy headline on top of it because they got to grow a business. It's a very bad incentive and it's actually ripping apart our society right now. These bubbles that it creates in social media. There's probably half your listeners if you have typical listeners who are like, 'Oh, that Joe Lonsdale guy, he's like I don't like him. He has different views than me and he offends me.' And the other one's like, 'Yeah, you go.' And it's like, it's so unhealthy.
Well, and the problem is, the middle of the bell curve who is neutral on Joe Lonsdale is not going to get shown this because that's not how the algorithm is going to work. And so,
I got to show everyone who either loves me or hates me.
Probably, you know, it's like in general, I think the... win over some haters, maybe.
That's right. Yeah, we'll try. That'll be our goal here. We'll run a survey afterwards.
Thank you. I feel like that is probably the problem with social media though is it just amplifies extremes and like a middle-of-the-road neutral agreeable take is not incentivized.
Yeah, I know. Most people are boring anyway.
That's true. No, but here's the thing. I was on a walk with one of my mentors at the Dish just before this. And she comes from the moderate left and I don't obviously, but I actually think like most issues there's a dialectic where there really is truth on both sides. And what I'm trying to do lately is like actually step back and just like where are the dumb people on each side, but where are the smart people on each side? And I think social media does not do that. Like I wish there was a way for it to do that.
So relating this to AI, I feel like the tech industry has some currently missed opportunity to do a better job with our like, you know, relations to... we need people to love AI more because otherwise we're going to like ban it and it's going to hurt us and they don't love all societies.
Like I don't think America loves AI right now. I think not enough people.
Yeah. I think not nearly enough people and I think they're scared, right?
It's understandable. Like I think are they scared? Why don't they like it? What do you think?
There's a threat for jobs, you know, obviously and so if you're in an industry where it looks like, you know, you're seeing all these headlines.
Do you know anyone personally who's worried about their job being lost? Just to be totally honest. Like oh my life's going to be tougher because everyone talks about it's not coming for me. I think everybody believes it's going to happen to everybody else.
What are we afraid for? I'm actually really curious like who's the archetype because you walk through it, it's hard to get to the person.
You know what's interesting is we should be afraid, you know, software engineering. Let's take as a good example of like it's probably the best application of AI right now. I just saw a chart that like job postings for engineers are going up and are like as high as ever right now.
So much to build right now.
And so you know then you have the question of are we basically just amplifying the value of that function and so there's going to be more demand because it's more valuable.
There will be people, this Oracle did cut I think 30 or 40,000 people because there were things that they had to do with 30 people that they have to do with three people now and they just don't have as many of those things to do. Right. So there definitely are already things where you're seeing some cuts for sure. The question is can those people now in this economy do better? And I think most of them can. But that's been another agent question we have to ask about.
But so then why do you think we're sort of like failing to make AI look like a positive force for the world?
Because everyone hates tech from social media. I think social media has had a terrible impact on our democracy. It's ripped things apart. It's like very distracting for people. It's just not good for kids. It's like terrible for young girls especially. I have five daughters and a son and it's like I'm not letting my young girls on social media just from everything I've seen. I think it's poison for them. And so this is like their last consumer experience with us. And now that this is like the next social media to them, of course they're afraid of it. Like I think that's the energy.
Yeah. You were mentioning to me before that China likes AI more than America.
Yeah. Some surveys show me it's like 75 or 80% positive and we're like 30% positive now. They may just take better direction from their leaders than we do, which is a cultural question. But I also think they're just more optimistic about how they could build with it and create wealth with it.
Which people should be here too.
I think this is going to, the last time we had productivity go up a lot over 30 years it was like the working class person, the median was like 2.4 times wealthier. There was 1870 to 1900 so more than twice as wealthy real terms for working class. I think it's going to be more than that the next 30 years. I think it's going to be like four or five times wealthier the next 30 years if we allow this to work. This is amazing. Like people said I think people will start to see hopefully if we don't make it illegal like just what it could do the next 5 years. I think it's going to be really positive.
It's almost like we need it to proliferate quickly though because to that point there's, you know, we were just talking about these founders who are having these explosive businesses. You have a lot of people who are resistant to AI. That sounds like the beginning of wealth inequality to me.
Well this is why we have to actually make sure that the applications of AI that help the middle class that help the working class win faster and make sure they're legal. So right now, for example, like the number one thing I think AI could do for 350 million people in the US is we can make healthcare like half cost, like really fast. Like healthcare should cost half as much and it should be way better. It should be way more convenient.
Well, there's a bunch of different levels of this, but we need healthcare services and AI. We need to actually have primary care work with deterministic AI where you're allowed to start a business that has one doctor and has a bunch of nurses and a bunch of others and it has the input through AI and they optimize it and they do things through AI that we know are safe and we prove out here's a 100 different workflows. So Utah just allowed one workflow. They allowed if you have a chronic disease and you've been prescribed medicine by a doctor, your represcription can involve AI. It doesn't have to involve a doctor and they're testing it out. But so what I'm pushing right now is a healthcare AI sandbox legislation. You know, we have Cicero Institute which gets things done in like 20 states. So, we're trying to get a few states to create an AI sandbox where we can very safely prove out dozens and then hundreds of kind of primary care workflows and then that would allow huge cuts in cost for these things. And then there's a bunch of other areas of healthcare of course it goes after as well. But this is you generally can have AI practice more safely for a bunch of things where it's actually better.
Yeah. Yeah. I mean I think there's going to be so many of these. We just have like there's like a marketing problem right now. It feels like to me.
Well if you're, you guys are the better marketers than me. I need help on that.
Well, I don't know either. I mean you know like... let's fight to actually bring down cost for people. That's a good marketing thing better.
One thing I don't think helps for, you know, you were close to this obviously, you know, if we can talk about a little bit, you know, at a high level at least about like the Department of War thing that just happened.
I mean I'm actually interested to hear you think it doesn't help because the war looked bad or Anthropic looked bad or what?
No, I don't think that. I think what it is is when you have two companies or other where you see them at odds and you see everybody fighting and there's this dynamic of like when the pigs roll around in the mud everybody's messy kind of thing and it's just like... just a negative energy around it.
It's like everybody's like telling each everybody is trying to project that like the other side's terrible.
You know what's funny is that the positive is that it is still being used right now with Palantir and Anthropic and it's working like really, really well. This is like one of the most accurately fought wars with least civilian deaths and the most bad guys falling so fast, which is amazing, you know. So, it's like all this positive for the US. If you...
It's like a nasty primary election before the general election.
You know, and it's like, hey, like there's this sense I have where I'm like on some level like if we all really care about AI getting there, you don't want a really nasty primary before the general. It doesn't help.
And I So, I think there's a little bit of that. I'd be curious like you know because you're obviously very close to the Department of War and so like you know with...
My friend Emil and then Depber and all these guys in the middle of this.
Yeah. And so it's like you know with most people in tech it's like people have got sort of their tribe to some extent or whatever and that's fine.
We were talking about this ahead of time. That is my tribe but I was trying to tell you ahead of time that even though I am on and I can explain why you know they think they're right and Palmer's right. Like I actually introduced Dario and Palmer and was trying to get them to get...
And it's like how you're saying you appreciate the moderate left and you want to hear the good side of the other side of...
This, I think it's really important in this state like if you've heard about the DO and Anthropic battle and you only think there's like stupid people on one of those sides. You have to assume it's smart people who want good outcomes for the world. Smart people want good outcomes who have their good point of view. So my main point of view supporting the Department of War ultimately is that you can't have the gray complicated nuanced decisions about what's allowed being made by the company or having the company involved in those. You have to be able to make those nuanced decisions at the Pentagon at the end of the day and I think that's a fair totally correct viewpoint. But stepping back like listen originally Anthropic came from like a much more effective altruist I'd call it very left very different perspective and originally they had like I don't know like 50 pages of demands for the Pentagon which were ridiculous and then to Dario's credit a he went and worked with the Pentagon ahead of time more flexibly he was trying to help America he does believe America's better than CCP b he got all that 50 page of nonsense after lots of iteration with his team and lots of work down to just two requests.
And he really has boundaries of why they created Anthropic and what their boundaries should be and that's in his mind and he's managing a bunch of co-founders who have different views, you know, and they got them all the way there. They were working with them. They tried really hard. But at the end of the day, like when they're in the room with Emil, there's 20 people and the two demands were like no autonomy and no surveillance. And they'll say, 'Okay, well for autonomy, let's say we're going to be using one of Palmer's autonomous fighter jets and something. We're planning and Anthropic's involved in planning the whole thing. Is that not allowed?' Or let's say we're using some autonomy in how we respond to attack and oh that's offense. It's okay. Well, what if we're doing an attack? But then as part of the attack we're using autonomy to respond to things at time there's all these nuances and Dario's response was call me. We'll work it out. Just call me. And the secretary's like well I can't just call you. I'm in charge not you. I can't have a CEO make the decision. So I respect where Dario is coming from but I think the part before I had to make the decision I had to make. I'm very sad I got into all the mudslinging. I'm very sad about that.
Yeah. No totally. And that's where I'm just like, you know, to me, you know, going back to the AI thing, it's where I hope there's some way for the industry to just like want to think of it as, you know, all on a team of, you know, this is going to be something that's good for the world.
Yeah, I think it's good for the world, but I think we have to have debates and it's fair to have debates and there's going to be really tough debates about how this is going to be used. And, you know, I think Dario is right. There are laws Congress does need to pass that it has not passed yet because it hasn't even anticipated some of the nuance of like how can you even do surveillance and things with these tools you could never have done before. So there's a lot of good points on both sides here. We're going to have to work together and hopefully our republic can figure out some good rules.
I hope it will, I think it will.
On the defense side you know obviously you've been involved in a bunch of companies like Palantir and Saronic.
Yeah I started Palantir, Saronic and I only invested in...
Only invested in every one of the early rounds. So those guys are crushing it, man. It's so impressive to see, you know, three of them used to work with us at Palantir and then Palmer is a good friend and it's just Yeah. And then Dino at Saronic just, you know, he's built more big ships than anyone else now. Fastest built ships since World War II that are big. It's so fun.
Yeah. I mean, I wasn't doing venture 10 years ago, obviously. But like seems like there was no interest in this stuff now. And now outside of AI, it's like the hottest.
It wasn't just that there was no interest. Like when I wrote the check into Andrew, I had multiple people tell me they would never work with me again and to not even talk to them about their next round because it's so evil to be working in defense and working on weapons. It's like it was like you had to basically be willing to just go completely against the cultural zeitgeist and be excommunicated.
Basically, it's like because which is like, but by the way, sometimes that's the most valuable things to do is when you do stand up to you. Well, it is. I mean, Trey wrote something, Trey Stevens even wrote something recently that's basically about like you want to be funding people who are, you know, saying the thing that's not popular and not exciting and it's going to be the creation of the next...
I love this. This is one of my favorite things is like, what does no one else want to do that is really important to do? Yeah.
And by the way, to me that's like a definition of leadership.
It also, like, it both means you're playing in a space that other people aren't playing, but it also reflects a personality type that has like a very...
You're willing to just say you guys are all wrong. It's important. Can I give you a crazy one right now that we're doing that's working?
So, China during COVID ended up cutting off a lot of things from us. And they had this whole big strategy of like going after our biotech sector. And one of the really key things you need in biotech is to do primate testing. And this is very controversial in the US because obviously you're hurting monkeys, of course. But like my 5-year-old daughter, by the way, gets it. I say, if mommy had a very bad disease and we were trying to work on a cure, do you want us to test the cure on mommy or do you want to test on the monkeys first in case it kills the monkey? Would you want to know that instead of killing mommy? She's like, 'Test the monkeys.' I'm like, 'How many?' And she's like, 'Hundreds of monkeys.' Hundreds of monkeys before mommy.
What's the ratio? It's high.
It's very high, right? Because it's mommy. And actually, but there's like wisdom there. Like I think human life is more important than even though I feel bad hurting a monkey, but like, so get this.
Well, you don't test the monkey right away. You obviously are...
No, of course you do all these other things first. You try to use AI to figure it out, but sometimes AI is not going to know as well as testing it. And so anyway, there's not nearly enough primates now. And like no one will touch it. So we started a company to provide safer human medicine. It's called... and we're providing like massive numbers of primates working with a bunch of biotech firms and we're breeding them here and bringing them over and no one will touch it because they're afraid because it sounds scary to do, right? But guess what? Would you rather save your sibling or the monkey?
That's a very taboo idea.
