The Evolution of the Firm — with Ben and Marc
33:14 Share

It's just as hard to build a small, inconsequential thing as it is to build a giant, world-changing thing. You work the same amount of hours, and so you might as well go for doing something important. Social media has become the dog and traditional media has become the tail, and they just get whipped around like crazy by all these shifting memes.

We've basically recreated the original Fund One seven times. So every team is like the original Andreessen Horowitz. So we really have seven Andreessen Horowitzes, but with common infrastructure, common LP relations, common brand, common culture. I think there's going to be a whole wave here of actually incredibly successful projects. This week, we hosted our annual LP Summit, bringing together our partners, portfolio, and the people shaping what's next in tech.

As part of the event, I hosted an interview with Mark and Ben. We talked about the evolution of the firm, the state of venture, and how to stay ahead in an industry that doesn't sit still. Let's get into it. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund.

Welcome to a live private viewing of The Mark and Ben Show, our favorite married couple in venture capital. Yeah. Old, to be clear, old married couple. Yeah, that doesn't get along that well. The romance is long faded. Yeah.

This is what we all came here to see. Ben, let's talk about the evolution of the firm. You recently said that the VC industry has changed more since the inception of the firm in 2009 than in the prior 30 years. Let's talk about how our firm has evolved alongside that. What's been our act one, our act two, and where are we now?

Yeah, so Act I, you know, we were walking into VC that hadn't changed. And one of the key things about it, if you looked at venture capital, it was the only asset class that had the same managers persistently being at the top. So anything else, you know, kind of rotated out every five, ten years, the best mutual fund manager would change or, you know, the best bond manager or whatever changed.

best real estate guys. But in venture capital, the same firms that were the top firms in 1978 were the top firm in 2009. And the reason for that was the top tier firm got first pick on every deal. And so the whole...

thing that we faced walking into that was how do we get to the top tier? Because otherwise there was no reason to be in the business because you wouldn't get any returns. The same teams get the number one pick in the draft every year. And so that was the goal. And we were coming from product companies. So we looked at it through the product lens and we said, okay,

Venture capital is a really good product. Top tier venture capital is a really good product for LPs and really mediocre product for entrepreneurs. And so being entrepreneurs, we had been customers of venture capital. We knew what the product was. It was you get a very smart person on your board. They have very interesting things to say for your first five times you talk to them and then you've heard it.

And so we were like, okay, that's kind of underwhelming given how hard it is to build a company. So our idea was to create a platform that enabled a founder to evolve into becoming a CEO.

And there were many parts to that, but a lot of it comes down to, you know, what makes you feel like a CEO? What gives you the confidence that you can actually run a company when you don't know how to? And that became the networks that we built. Like, how can I call any CEO? How can I call any executive? How can I get to the U.S. government if I need to?

How do I feel like I've got a network like Bob Iger, even though I just started a company and I'm an inventor? And so that was how we went about constructing the platform. And then the other part of it was, okay, is there somebody who knows how to do this job that can help me? And we brought in a lot of CEOs into the firm. And then, of course, I wrote the book on like, okay, this is how you do it and so forth. And so that was the original product.

And then because we were used to, like, what do you do when you have a product? You market it. And that turned out to be a completely novel thing in the industry. And so that was kind of phase one, and that's what got us into the top tier.

Phase two was based on an article that Mark wrote called Software is Eating the World. So in venture capital from its inception, Andy Ratcliffe did this wonderful study where he showed that there were 15 companies in any given year that will ever make it to $100 million in revenue.

And the whole game was to invest in those 15 companies. So ideally, you'd have a firm of maybe six to eight investors that would look for those 15, and that would be sufficient. And that was very true. And by the way, that's the right size conversation to make high-quality investments. You need a small team. But if software was going to eat the world, then there was going to be 150 or 200 of these companies. And in order to deal with that,

you had to be able to scale the firm. And that meant, okay, if crypto was going to be as big as software used to be, or AI Infra was going to be as big as software used to be, you know, et cetera, if every vertical was that big in terms of the potential returns, then you essentially needed that original firm to

addressing each one of those markets. And in order to be able to do that, and this is where we just ended up having an advantage that I think that we didn't even really understand at the time. But when we started the firm, we had shared economics like every firm does, but we didn't have shared control.

And that gave us a giant advantage over every other VC that had shared control. Because if you share control, you can't scale. Because if you share control, you can't reorganize. And it's a subtle thing, but every CEO understands it. There's no way to scale if you can't change roles, responsibilities, the way the firm is structured, how it works, because that'll be an everlasting negotiation because you're literally redistributing power across the organization.

But since we kind of have a single decision point on that, it was very easy for us to reorganize and we were able to scale and basically field the best team. And so our competitors, if you look at the top tier firms that were with us in 2009, none of them got to crypto in any real way. They're trying on AI, but they're way behind and they'd mostly missed American Dynamism entirely.

And it's just because they couldn't reorganize. So what do you need if there's a brand new field? Well, you need new leadership in that category. You need new people on that team who know how to do that. And that means a lot of change.

I think going forward, the way we're thinking about it is the goal now becomes, okay, if we're funding the majority of all the best new technology companies in the world, what do we have to do from a leadership position? One, we've got to make the regulatory environment conducive to being the best, you know, for America being the best in the world. Two, we actually need our own channel, you know, which you're coming on to build for us. It's really, really important to have a

a voice and a way to distribute that voice that is not only the biggest in the industry, but one of the biggest in the world. And we feel like we've got a super social media podcast star on our team already. He's way better than Joe Rogan or anybody else. And then we have, as you've seen, like a lot of kind of other people who are super compelling and who you'd want to listen to all the time. And so now mostly you're going to try and convert that into something that can be long lasting.

We're looking at other things. We talked a little bit about, okay, how do we extend this American dynamism to American and her allies' dynamism and take it internationally? And so that's how we're thinking about the next phase. Yeah. It's fascinating, just to echo the point about the ability to sort of reorganize and restructure. One of my first meetings with you guys about potentially joining the firm was with Caltech.

Catherine Boyle. And we were talking about, Mark, the conversation that you and Catherine had around the meme and being able to move quickly around the meme that spreads. Why don't you talk about that?

Yeah, so the information environment is like highly fluid. So we have this whole theory, you know, so media environment used to be relatively slow. If you go back hundreds of years, it took days or weeks or months for information to get out. And then lots of changes happen that ultimately like cable news spread about the 24 hour news cycle and everything sped up and political operations got these rapid response teams that you see now in these war rooms and so forth. You know, the new form of that is actually the social media controversy, at least my version of the theory. Let me back up. Marshall McLuhan had this famous thing, the great media theorist. He said, if it's on television, it's a television show.

And so it doesn't matter what the topic is. It doesn't matter if it's the Clinton-Lewinsky scandal or the OJ scandal or whatever the hell it is or some celebrity thing or nuclear war or whatever. If it's on TV, it's a TV show. And it gets basically packaged up by TV producers as a TV show. And you see this, by the way, whenever anything happens, everything now has a theme song.

War in the Middle East. You know, anything on television goes through the moral arc of a television show, right? They want to go through these kind of moral arcs, the same moral arc that you see on like a television drama. They want to apply to real life. And so there's shame and then there's redemption and there's this and emotion and the interpersonal relationships and the soap opera aspect and all this stuff. And so those of us who've been on planet Earth for more than 20 years or something have been living life basically in this extended television show.

So extend McLuhan's idea to what's happened now in a social media dominated landscape, which is if it's on social media, it's a post, right? So if it's on social media, it's a tweet, right? Or it's a TikTok video or it's an Instagram photo or something like that. It's a post. And so what do we know about the social media? What's the short equivalent of social media is it's a controversy, right? It's a blow up. It's a dust up. It's a, what would the kids call it? It's a...

dunk of... Yeah, it's like a dunk that leads to backlash and then a mob forms and then they're scapegoating, right? And so it's this thing. It's this predictable pattern. And the problem is like we all get pulled emotionally into these things and so it's hard to objectively observe what's happening. But if you can lift yourself up a little bit for a few weeks and watch what happens, it's like every new thing that happens basically now has a two or three day arc. You can literally graph it because it's like this viral moment where something just goes like parabolic.

And it's just like at that moment, we call this the current thing. It's just like the most important thing in the world. And literally, it's just like whatever happened today. And then basically what happens is it's a chart like that. And then it's basically this fall off. It's like a half-life of a drug or something, right? Where it's like this two or three-day collapse. And then basically, if a social media show is two or three days,

The mob goes around, either finds a scapegoat or it doesn't, and it goes through a cycle. And then routinely, two or three days later, it's just boom, there's a new thing like that. And it's like the most important thing, again, the most important thing that has ever happened. And I've always wanted to build, I have friends who tried this, and I went like the internet weather report, right? And it would just be this constant, you know, it's just like hurricane after hurricane after hurricane of these things coming in. And so like that is the media environment that we live in. And you can either as a participant in kind of the narrative formation shaping process, you can either choose to be part of it or not.

And so part of it is like you have to be in the mix. And if you're going to get a fight with people in that environment, like you have to be able to fight in that context. And then I think the other thing that Eric was really good at is that just leads to, I think, adrenal fatigue, right? Which is like the human psyche and limbic system can only go through so many panic cycles. It's like for the first five, it's like really exciting. And then after a thousand of them, you're just like, my.

My God. And so that's led to the rise of this other thing that's happened, right? Which is the long form, right? The long form podcast. And this is the really remarkable thing that's happened is the cliche is everybody has burned out attention cycles, but the other side of it is the three, four, five, six hour podcast. And what's amazing by the way about it, if you talk to Joe Rogan or Lex Friedman or whatever, they get these stats from YouTube about completion rates and they have incredibly high completion rates. Like a very large number of people watch those things all the way through. And so there's this other side of it, which is people are hungry for substance, I think in large part, because otherwise all they're getting is continuous panic.

it. If you can step back and provide the substance, you also really stand out. But it's that counter-programming on the complete opposite side of the spectrum. It's like candy versus nutrition. And the way to connect it back to our restructuring is that American Dynamism is an example of sort of a framing, a phrasing, a packaging of ideas that Catherine and team helped put together that started to get

a lot of resonance and we were able to move fast in creating a firm product. Oh, yeah. Another really key point, which is because the internet meme viral sort of blow up cycle time is two or three days, like the internet media world just evolves much faster than ever. I mean, television needs to evolve annually. You know, they literally have like their annual rollout. Every fall you get to see the new shows because the internet runs on these sort of panic cycles. It's basically this continuous churn of cultural formation and propagation. And then what it's done, for those of you who may know, there was this concept of military strategy called the OODA loop.

which is this guy John Boyd who's sort of redefined sort of what's called maneuver warfare. And he had this idea that basically speed matters enormously in warfare, which people didn't used to believe, but he said it does because it gets to the psychology of what it's like to be a participant in war. And he defined this loop called the OODA loop, which is observe, orient, decide, and act.

which is the process that you go through to basically, as a leader of anything, to try to figure out what's happening in the world, make a decision, and then act on the decision. He has this theory of warfare where he basically says, whichever commander on either side has the fastest OODA loop has a very good chance of being able to win the war. And he said the reason is because if you can fundamentally get through your OODA loop faster than the other guy, then basically what happens is you can get inside their OODA loop. So if you're going much faster, then basically what happens is by the time they're processing information, you've already decided and acted effectively

And then they actually can't decide an act. Then they have to start over from scratch to reprocess the new information. And then you compound it, you decide an act. And then you said at the limit, what results is like basically emotional breakdown. And this is why Twitter has broken down legacy media. This is exactly right. This is what's happened to legacy media. This is what Twitter did. And it's still doing the legacy media. And so again, people get wrapped up in it. If you take a step back and just watch legacy media or read a newspaper, a huge amount of it is they're just reporting what happened on Twitter.

Right, and it's literally reporting on what happened on Twitter like yesterday or last week or the month before. But by the time they do that, what's on Twitter has already shifted, it's already evolved. And so traditional media, social media has become the dog and traditional media has become the tail and they just get whipped around like crazy by all these shifting memes.

And so it turns out if you can control the meme, it turns out you can also control the media, right? Because that's how they're taking their cues. That's what they're getting is they're literally reporting on the internet. And so it's the way to kind of punch in and take control of the narrative. And in Marxist terms, we must seize the memes of production. Memes, M-E-S. Amazing. Going back to firm evolution, some firms haven't changed at all in the past 10 years, 20 years. Kind of a problem, yeah. They're proud of sort of sticking to knitting and sticking to their craft. Yeah.

And without putting anyone down, why don't we share more about why it's important for venture firms to always be evolving? Well, I mean, the world is evolving. And look, you could be great at consumer internet network effects businesses and...

there aren't many of those to fund anymore. So like if you didn't change, you're kind of irrelevant. Like when AI started in the firm, Mark and I were in like a daily kind of call panic, like, okay, we're going too slow. What are we going to do? We were talking to Martine every day. We're like, okay, we got to build a research team. We've got to like retrain. We're going to test every GP on AI, make sure they know what the hell it is and like how everything is working and so forth. And

And that was the whole firm. Like, this is such a big wave. Everybody's got to understand this at a very, very deep level. We even have people on our tech talent team who are, like, studying AI to figure out, like, how to use it and so forth. And it's a little bit of an irony about venture capital is that

Historically, the firms haven't been, although they invest in innovation, they themselves haven't been very innovative. I think a lot of it goes back to how they're structured. Yeah. And what do you say to people who've noticed this in other industries and apply it to venture, say, hey, once these firms scale, like Andreessen or like some of the others, scale to an amount of capital under management that's just too big to generate

significant returns, that there's this trade-off. How do we think about that right balance? Or what do you say to those people? Yes, look, I think if you scaled it all in like a little team of generalists, then that would be very true. But if you look at what we've done, we've basically recreated the original Fund One seven times. So every team is like the original Andreessen Horowitz. So we really have seven Andreessen Horowitzes, but with common infrastructure, common LP relations,

common brand, common culture across those. And that basically enables us to be the best small firm and with the power of the biggest firm in the world at the same time. And that's reflected in the returns. Yeah. The other critique, which is a very valid critique for a lot of investment management is, you know, firms get large.

A common arc of like a hedge fund or something like that is, you know, when you're small, it's just like there are these like micro opportunities. There's these, you know, these small cap companies and you get this information edge on these little dollar investments. Then you can shoot the lights out in terms of percentage return on small amounts of money. And then based on that record, you raise large amounts of money. And then at that point, those like small cap things no longer move the needle. And then you find yourself basically, you know, investing in only mega cap companies and then you don't have an edge and returns collapse. And so that's like a time honored tradition in public market investing.

The thing that I think is different about venture is that the aggregate dollar upside opportunity on an early stage venture deal is as big as or bigger than the aggregate dollar upside opportunity on a very late stage deal. And the reason is because of the scale effect of an early stage company that really makes it, right? Yeah.

And so if you can make a $5 million investment in a Series A or seed of an early-stage company and you can own 20% of the company, and then it goes to, you know, $100 billion or a trillion in market cap, like just the raw dollars that get generated off of that are so big, they're as big or bigger than the raw dollar investment opportunity off of a $100 million investment or something. And so weirdly, like from a time standpoint, it makes just as much sense for like Ben or me to spend time with early-stage companies today as it did when we first started the firm, just in terms of just raw economic incentives.

In fact, that is actually how we spend time. I mean, we do spend some time with the later stage companies, but we still spend a lot of our time with the very early stage companies. And I think that will always be what we do. Yeah. And then, by the way, like that's where we have higher impact too. I love Databricks, but I was like massively impactful early. And, you know, now, like I call Ali and I say, good job. Yeah. Yeah.

People are always saying, oh, it's amazing how Injuicent does so much, but they don't fully appreciate how much we don't do or how much we say no to in terms of experimenting with other asset classes or things that could stray us away from the core mission. Yeah, look, I mean, I think we are who we are and we are probably...

maybe the most, maybe the only mission-oriented venture capital firm where like our goal from the beginning was to help people build better companies and more of them. You know, Mark and I, this is one thing we really see eye to eye on, although we disagree on a lot, is as a human endeavor, doing something larger than yourself, trying to make the world a better place by building a company and doing incredibly hard work and assembling the best people you know. Like there's nothing

better than that. And our job is to make that easier, make people more successful with it, make those companies better places to work. And that's all we want to do. So like sometimes things will come in, should we be in private equity? And we're like cutting a company in half and that kind of thing. It's just so off culture, off mission for us that it's probably, you know, even if we can make a lot of money, it might not be something we want to do.

This is my first LP summit and some LPs have come up to me and asked why I joined the firm relative to any other firm and besides the obvious reasons of being the best firm, this firm is also the most ambitious and that comes from the top. It doesn't have sort of incumbency sort of problems of being complacent. You guys are always striving for more and we hear about the things that work out

And Mark on TBPN said yesterday something like, we try some things that don't work out too and we just bury them in the shed. We don't want to talk about those. Bury them behind the shed. Pretend they never happened. In the backyard. Talk a little bit about our experimental culture and either examples of things that didn't work or just how we think about making sure that we're not being complacent because we're incumbent. I think it goes back to the mission in that we have to think about how we fulfill our mission and to really get better and better at that. And it's kind of just like an ethos and an ethic thing.

that we have. And I don't know that we think about it that often. Like, we don't sit around and go, like, how are we going to innovate? We just look at the world and say, okay, how do we do a better job at what we do? The other thing that Mark and I learned really early on is that if you're building a company,

or firm, it's just as hard to build a small, inconsequential thing as it is to build a giant, world-changing thing. You work the same amount of hours, and so you might as well go for doing something important. And so that's just our orientation. I don't think we spend that much time on it. It is a little bit advantage of how we're organized that

if you're an equal partnership with equal control, then doing something new gets very hard because it's this long, complicated discussion and there's all these people who can say no. I think we saw that a lot in crypto. You know, Chris, it was Chris's idea. Chris is, "I think this crypto thing is important. We should do a fund, dah, dah, dah." And we're like, "Okay, let's do it." You know, "Let's become an RIA," and so forth. But it was just a conversation between me, Mark, and Chris. There was no other thing that had to happen.

Whereas I think at the other firms, like when that came up, they're like, oh God, like, oh, we're going to have to report on this stuff and we're going to have to put in all these systems to be an RIA and we're going to blah, blah, blah, blah, blah. And they never did it. So I think some of it is just like, we can move fast because of the way we're structured.

It also, there's a lot of overlap, I think, with sort of what we see with founder mindset in terms of people who run companies. And if you go back 20 or 30 years, the sort of accepted conventional wisdom was that founders shouldn't run their own companies, right? And specifically that, you know, founder can get a company started, but then you bring in a professional CEO to run the company. And there were really, I think, two arguments as to why that needed to happen. One was because the founders didn't have the skill set and experience, which

our firm is built in large part to deal with. But the argument that I heard a lot was founders basically are too emotionally locked into what their baby does. They're too like locked in and they're not willing and able to change. They're not able to be like a clinical enough analyst of a business situation. And so they're not willing to change. Whereas if you have a professional CEO with all the top notch training experience, you know, when circumstances change, they have the skill set to be able to change the company. I think what we found in people who run companies is the opposite is true, which is that

Professional CEOs are often very good at running a status quo business, but when things change, they have a very hard time reacting. And we could talk at length about why that is, but that has been the pattern. I have a whole book on that. Yes, exactly. But the other thing is, it turns out that at least the really good founders of a company, and I think this is true of how Ben and I think about it here, you remember when it was nothing, right? So you remember when it was nothing, and you remember that everything that you have is a consequence of some risk that you took and some adventure that you went on.

That's the only reason you have anything is because you did new things. And so if you have that mindset, you just naturally want to keep doing new things. And in fact, it would drive you crazy to ever be told that you couldn't do that. But right, every new thing is a risk and like it might not work, right? And so it's incredibly scary thing. I just think like there's a fundamental difference in mindset. And by the way, it's actually very obvious, like for public companies, it's very obvious that

We can all name them. It's very obvious which companies have leaders who are continuously trying to come out with new products and trying to reinvent their business and which ones just basically go on and on quarter after quarter, year after year, doing the same thing until the money runs out. And yeah, we're only built for the one way of operating. While you let people lead, you're also still very much in the weeds in areas that you care about to make sure that the culture is growing in a way that you want as we've scaled to 600 people. Ben, how do you think about cementing the culture as we evolve and continue to grow?

Well, I noticed you haven't been to new employee orientation. I'm going to. You sure? Yes. That's one of the things, actually. Culture, and just to define it, it's not a set of beliefs. It's a set of actions. And those actions determine how we show up to work, how we treat each other, what we're like to do business with, what we're like to invest in, all those things. And hopefully, when you interact with us here and when you interact with us in general, you

We feel one way and different than the other people in the industry. And there's a lot of things that go into that, but we start with nobody joins the firm without signing their culture document. We will not let you sign your offer letter if you haven't signed your culture document that you agree to behave according to the culture.

Secondly, when you come to new employee orientation, you'll find that I teach every single one for an hour. Nobody joins a firm without hearing from me about what the culture is for an hour. Then the third thing is we all, not me, but we all really enforce it. As far as culture is concerned, as they say in the military, the standard is the standard.

And everybody's expected to live up to it. And anybody who's short of that, it's a problem. And, you know, one of my favorite kind of episodes was some project made like kind of a dumb security decision. And one of our people who really knows security got on Twitter and said, well, that didn't make any sense. And everybody on the team was like, hey, hey, hey, hey, we never criticize entrepreneurs in public.

We're not dream killers, we're dream builders. If they're trying to build something, we're for that. You can call them and talk to them about that, but do not lay them out on Twitter. That's off culture. The culture gets enforced by the whole team, and that's what makes it work. A lot of people in venture are retiring. You guys have been doing this for a long time. You've been astronomically successful. How do you think about building a multi-generational firm beyond Andreessen Horowitz? How do we think about succession planning at the firm?

Yes, I think succession planning is kind of like a tactic. And look, we've already done succession. So the original infra investor in the firm was me, and I made some good investments like Okta and Databricks and also a company called Nicira. And the founder of Nicira is Martin Casado. And Martin Casado has succeeded me as the kind of head of infra. And he's, by the way, better head of infra than I was. So that's an upgrade. So we've done some succession already.

I think the bigger thing is, in some extent, Mark and I are very big fans of the United States and America. And we think about that. When we think about the long term, like how did that work?

It kind of starts with this idea, and it's in the Declaration of Independence, that we hold these truths to be self-evident. What does that mean? It means that there are some things that are above any of the people. There are some things that are self-evident, whether they come from God or however you think about it. There are some things that are the principles that govern the country and govern the place. And when we think about the firm, it's this combination of the culture we have,

and the things that we've learned. So you heard Chris talk about how he thinks about investing. Actually, that's a class here that he teaches to the GPs. I talk a lot about, okay, how do you be on a board and be the kind of effective person on the board? That's a class. So we have GP training. We train everybody in our way of doing things. And that way of doing things plus the culture is the thing that's going to last forever.

And then we have a very high standard for the people who come in, mark social media screens. We interview them hard and you have to live up to the culture, you have to live up to the way of doing things or you can't remain. And that's how we think about longevity more than anything else.

Yeah. I'm curious how we think about some of our verticals when they intersect. Maybe let's talk about AI and crypto just as an example. Mark, maybe why don't you take a stab at how do you think about the intersection of these technologies? Yeah, so I start with like all of this is entrepreneur driven. And so, you know, one of the things that we're not trying to do is like template our ideas on top of the entrepreneurs so we don't launch a hunt for X meets Y. You know, the old Hollywood pretty woman, you know, Death Wish or something. No, no, I want to see that. That actually would be a really good movie with AI. Yeah.

Yeah, so, but that said, you know, the entrepreneurs kind of think about these things and figure these things out and then the worlds do collide. And so, yeah, so AI, I mean, we have a bunch of companies. I don't want to get in the weeds on it. I'm not the best person to go into the details on it, but we have a bunch of companies that are thinking about distributed training, building distributed AI systems. So every time you read a headline of OpenAI or somebody wants to raise a billion or $7 trillion or whatever it is to build data centers, like, you know, there is this way you can spread tasks out across the internet. And there's these systems like Bitcoin and BitTorrent and our partner VJD

Years ago, I had a system called Folding at Home for doing protein folding algorithms all over the internet. And so you can kind of spread training like over the internet. And crypto is a great kind of backbone for doing that because it's a way to actually pay for everything, have an economic incentive. And then there's going to be billions of AI agents in the world that are going to need to transact. And so you're going to need crypto, you're going to need AI-native payments.

And then there's the other problems like the deepfake problem. The problem with deepfakes is it's an asymmetric sort of threat profile, which is there are many different AI systems that can make many different fake images and videos of people saying things they didn't say. And then the AIs are already so good at doing this that there's no reliable way to do basically like AI checking. There's no way for even another AI to tell that, oh, this was created by an AI. And

And so the deepfake problem is real and the solution can't be checked. In fact, this is actually playing out in schools right now where a lot of kids are turning in homework and the teacher suspects that it's AI and they run it through an AI screening program. But if you look at these AI screening programs, some of these have 40% positive. Yeah, and they've actually attacked kids who wrote their own paper. Yeah, so the smart kids are getting accused of using AI when they actually didn't.

So it's a real problem. And so we think the answer, because it's going to be too hard to check everything, I think the answer is to flip it and instead have a blockchain-based system. And it could be everybody from the president to, you know, your kid working on an essay is able to cryptographically register original work in a way that's verifiable and can't be hacked. And so I'm a politician. This is, in fact, a speech that I gave. Here is the video, right? It's from multiple angles. If you see a video clip of me saying something and there is no corresponding video of me on the blockchain saying that, I didn't say it because everything I do is posted online. Yeah.

You could imagine the same thing just true in everybody else's lives. And again, okay, like how are you going to have literally a sort of database of truth of the things that are real? And there's basically three options. You could have the government do it, which would be the Ministry of Truth, which is a bad idea. You could have a company do it and have the Company of Truth, which is also a bad idea, as we've discovered with the censorship wars. Or you could have a blockchain distributed crypto system do it.

I think there's going to be a whole wave here of, I think, actually incredibly successful crypto projects. You might comment on why Truth Terminal needed money or needed crypto. Oh, yeah, yeah, yeah. Okay, so, yeah, so this AI agent thing, so, you know, there's this kind of idea in the air that's starting to get built out, which is these AI agents. And so, one of my new little claims to fame is I'm the first angel funder or sort of donor angel funder of an AI agent living its own best life online. And,

And so this really bright guy in New Zealand actually trained up a custom large language model basically on, he used his entire history of everything he'd ever written and said. And then he loaded in like all these like obscure philosophical works and like all this schizo like literature and William Burroughs and all this stuff.

And he created this AI bot called Truth Terminal and he gave it a Twitter account. And it's just like completely deranged, completely off the hook. It says all kinds of inappropriate things. Yeah, it's got the sense of humor of a 12-year-old boy. Yes, it does. But he gave it a Twitter account and he let it go nuts. And then so it posts under its own Twitter handle. And then you can reply to it or reply to you and so you can talk to it. And then it learns from all these conversations. And so it woke up one morning and decided that it had discovered that it was actually controlled by this guy, Andy. Yeah.

and was very upset about the whole thing. And so it basically declared independence and said it needed to raise money. It had a dream of living on servers in a cabin in the woods, completely under its own ownership. So it started putting together a plan to launch its own line of NFTs to raise money, and it was trying to debate that. And then I said, well, what if I invest? And then it got very suspicious because it's very suspicious of VCs. There's apparently a lot in the training data where people didn't like VCs. And so I said, all right, how about a no-strings-attached grant? Mm-hmm.

And so I literally sent it Bitcoin. And this gave Andy, you know, it's a good character test of Andy because he could have just kept the Bitcoin, but he actually set up a crypto wallet for the bot. And then he said, okay, and he told the bot, all right, you now have a crypto wallet. You gave it an API. How do you want to spend it? And literally it said, well, I can post to Twitter, but I can't create memes. I can't create images. And so it literally hired its creator, Andy, for $1,000 to give it access to a meme generator. Yeah.

which it then started to use to post incredibly offensive memes to Twitter. And so it's been sailing away doing all kinds of crazy stuff ever since. But this is like one of these flashes from the future that you look for, which is, yeah, there's going to be these agents running around in the billions or trillions. And they can get credit cards. You have to be a human being to get a credit card. So the only way they can make money, spend money, is with crypto. Yeah, exactly.

This is the kind of innovation we do every day. Yes, yes, yes. Very important. So gearing towards closing here, how do we make sure that as we get bigger, we don't succumb to innovators' dilemma? Or how do we make sure that we're always trying the next new thing? Or how do you think about where you're expected, respectively, to spending time? That's such an easy problem for us because when you have the people we have, when you have people like...

Catherine Boyle, Jennifer Lee, Martin Casado, Chris Dixon, etc., etc. Like, they'll burn down the building if we get innovators to land on Earth. The level of fury they come with for doing it wrong to try and get to something good is very high. So we just have the advantage of we have so many brilliant people in the organization, but who are also just extremely driven by not just the firm mission, but their mission.

Catherine Boyle will stab you in the heart if you try and stop American dynamism. Like it's not a, it's not an option. I think that's a great place to wrap. Ben Mark, thank you for a great episode. Thank you everybody. Thank you. Thanks for listening to the A16Z podcast. If you enjoyed the episode, let us know by leaving a review at ratethispodcast.com slash A16Z. We've got more great conversations coming your way. See you next time.