Who Will Own the Internet? a16z’s Chris Dixon on AI and Crypto

0:00What will be the economic models for creative people in an AI world? Don't stop the inevitable, which is the technology progressing. Lean into it and rethink those models. If that to me is the most exciting area for this intersection. In the last few years, AI has been the talk of the town. Founders have tiveted, incumbents have plot capital into new projects. VCs have upended their investing VCs. All of this is part of the race to capitalize on what seems like be biggest platform shift in decades, and equally a new generation of the internet. This generation is not only an opportunity to rethink the past, but with parallel technology tracks from new hardware to crypto intersecting, we can build things we never could before.

0:44So, what will the economic model of this wave be when so much is being upented? You go to their websites, they give you an answer. And so what happens to the billion other websites if they aren't getting traffic. It's a question, right? When will we move past the skeomorphic phase of this generation to building net new behaviors? And could crypto be the counterbalance to the centralized gravity of AI, targeting more data, more compute, and more complex models? Where we're headed is a world where do you have five big systems? Let's call it three to five big AI systems. Joining us to discuss all this and more are A16Z Growth, General Partner David George, and A16Z Crypto Founding Partner, Chris Dixon.

1:23In the last year, of course, Chris wrote his book Read Right Own, building the next era of the internet, all about how blockchains might finally bring us back to the early promise of the internet, a decentralized democratic network of innovation, connection, and freedom. So without further ado, let's dive in. By the way, if you did like this episode, it comes straight from our AI Revolution series. And if you missed any of the previous episodes of that series, with guests like AMD CEO Lisa Sue and Thropic co -founder Dario Amade in the founders behind companies like Databricks, Waymo, Figma, and more, head on over to a16z .com slash AI revolution.

2:02As 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. Please note that A16Z and its affiliates may also maintain investments in the company's discussed in this podcast. For more details including a link to our investments, please see a16z .com slash Disclosures.

2:32First, thanks for being here. Yeah, thanks for having me. I always love hanging out with you. Obviously, you spend most of your time in crypto today. How do you generally see crypto and AI interacting? Yeah, I mean, so I think first of all, my kind of met of you is that The technology waves tend to come in pairs or triples 15 years ago, it was mobile social cloud. And I'm always giving this speech to entrepreneurs. They tend to reinforce each other. And so mobile was what took computing from hundreds of millions to billions of people. Social was the killer app that hooked them. And cloud was the infrastructure that made it possible.

3:04Right. And so you couldn't really have all three of them. And I remember back then, people having debates, which were better, it turned out they were all better. And they were all required. They're all required. And so I think of that with AI crypto and maybe new devices is the other kind of probably robotics and self -targing cars and VR and things I think of those as the three interesting things going on And I think they all kind of complement each other and work together It's a new way to architect internet services a new way to build networks that has a bunch of different properties Which I argue are beneficial for a bunch of reasons and can do a set of things you couldn't do before Right And so I think a lot of you think of it as Bitcoin or meme coins or something and so that's fundamentally not what it is to me Or I think to the kind of smart people working in the space There's many different ways in which it intersects with AI.

3:44So the first way, which is something we've invested a bunch in, is just using this new architecture to build AI systems. And so, for example, one of the core questions I think, we've just talked a lot about this from about the future of AI, is to what example AI be controlled by a small set of companies or controlled by a broad community? The obvious first question there is, is it open source? Yes. It's negatively shocked me how close source the world has become. Ten years ago, everything was open and put in papers, and then it all shut down and was closed. And they said this was for safety reasons.

4:14I think it just happened to be very good for their... I just think it's a beneficial business. I don't believe it's safe to think. But thankfully there's these ones like Lama and Flux and Miss Rall and things who are open source. I worry that's a little fragile because first of all, I don't know. A lot of them don't put their weights open. Is it really open? Some of it's open. Like the data pipelines not open. Is it really reproducible? They could switch it tomorrow. These models get better every month. and if they don't start doing the new frontier, I don't know. So it's like, it's very heavily dependent on one large company.

4:42Yeah. So one of the things we invested in is a stack of internet services that are built for the AI stack, but open services at different layers. As an example, there's a project called Jensen, which is building, think of it as crowdsource compute layer. So you as a startup can submit a job that goes beyond the compute you control, and it goes out to a network, kind of Airbnb style, people that have access compute, and the network manages that supply and demand. And that's the economic ledger. Yeah, that's one example. Another one is one called Story Protocol, which is a new way to think about registering intellectual property.

5:17And so you could create image or video or piece of music and then you register it on a blockchain, which keeps a record of the piece of media and the rights around. It uses existing copyright law. So actually, the blockchain record mirrors a legal agreement that's been crafted to work internationally. And then anyone can come along, and as long as they abide by your terms that you said, you might say something like, you can use this, you can remix it, you can create derivative works, but any revenue you make, you have to pay me 10%. That's true, and you set the terms. But that creates this sort of open marketplace where right now you have to call up some company and try to do a BDDL and this and that, and so you end up having this kind of thing where people either basically steal it or don't do it, or they're scaled enough to make a deal or something like you have open AI going to Shutterstock and they paid them $100 million.

6:06But this is really just for the very high -end companies. This is creating a broad democratic kind of resource where anyone can, small creator can set the terms. And then ideally what you create, and this is a recurring theme in the blockchain world, is you have this kind of what we call composability. I think the kind of core force behind the success of open source software. I mean, people forget this, but open source software, certainly the most successful open computing movement in the last 80 years. but Linux went from 0 % market share in the 90s to probably, I don't know what 90 plus percent market share today.

6:36And a lot of that's because of what we call Composibility, which is basically all these different people coming along and contributing little pieces to the system and the system collectively getting much better. In the same way that Wikipedia is a collective knowledge system. And so something like story protocol, you get the same kind of Lego break effect with media. So someone comes along and they create a character, someone else creates another character, someone else remixes them, someone else. And then you can use whatever AI tool, you can create generative AI, and you can create your story.

7:00I created a new superher universe where I use these other LEGO bricks, and as long as the money kind of waterfalls back, that's all okay. I think it's a really great vision that both allows for people to embrace these new tools, but also provides an economic model for creative people. I think that's a recurring theme in our investing, is like what will be the economic models for creative people in an AI world? Don't stop the inevitable, which is the technology progressing. Lean into it and rethink those models. Like that to me is the most exciting area for this intersection. You go from social networking companies which keep 100 % of revenue for themselves when creators create stuff effectively to something where hopefully the creator can capture an upfront amount that they set.

7:44And then ideally the composability allows for actually more creativity built on top. That's right. Because of the economic sentiment. Yeah. We're seeing people do interesting stuff with kind of crowdsourced model evaluation. just think of it as all the data side of things. Like you need more data, and we have this crypto as a breakthrough in new ways to design incentive systems. And so you combine that, and you say, well, how can you use new incentive systems to get more data for these AI systems, right? Data can either be an input or it can be a model evaluation or whatever it might be. So it's kind of what these companies like SCAL AI do, but in a crowdsourced way, instead of a centralized way.

8:18The project is co -founded by Sam Alman that we're investing in called World Coin, where the thesis is that in a world where AI can replicate humans and content, we need a way to prove your human. And the best way to prove your human is cryptographically using a blockchain. And so the idea is they have an incentive system to be able to sign up and originally it was this, or the scan drive -alls, some people, it was controversial. They now have systems where you can identify yourself in other ways including your passport and other things. But the idea is you prove who you are, you get cryptographic proof on a blockchain.

8:49And then you can use that for a bunch of different services. to think of very simple examples, think of captures. Today you have to go and play these puzzles, which I think have gotten so complicated. Not yet. Not yet. I proof anymore. I don't say I proof anymore. And they may be human proof. I have trouble with a lot of them, but replace those with a set of systems like that and other kinds of clunky fraud systems have an actual cryptographic thing. So I have a code, essentially, this is how cryptography works. And that code proves that I'm a human. And then you can layer onto that. Other kinds of things you prove on top.

9:15So I think there's a bunch, so this infrastructure layer of like take AI systems that exist today in a centralized way and decentralize them, both in terms of code and services. There's new things you couldn't do before, like machine -to -mean -machine payments, and then there's these sort of really far off things that I find the most exciting, which are like what are new business models in this world. One of the things that you pointed out to me right after the chat, GPD Moment, is, you're like, hey, we have the potential for a sort of a break in the pack to the internet. Oh, yeah, yeah. Which I think is a super fascinating.

9:41Yeah, yeah. There's another chapter on this in the book toward the end. I call it a new covenant. So you think about the incentive system. One of the main reasons the internet succeeded is that it had a very clever incentive system. How do you get five billion people to opt into the system without having a central authority tell them to? This is because of the incentives of the internet. And specifically, there's been what's emerged over the last 20 years is, I call it an economic covenant between the platforms, specifically social networks and search engines, and all the people that create websites that essentially those link to.

10:13And so if you're a travel website or a recipe website or an artist who has illustrations, there's an implicit covenant you have, let's say, with Google, right? Which is you say to Google, it's okay if you crawl my content and you index me and you show snippets in your search engine, if you send me traffic back. This is how the internet is evolved, right? And why do you want traffic back because you have some business money? Maybe it's a free site, maybe it's an ad -based site, maybe it's subscription -based site, but whatever it is, somehow you have a way to make money on traffic. There's some understanding, right?

10:44Any case - It's really beneficial. It's really beneficial. And occasionally that has been breached. So there was a thing Google does called One Boxing, which is they would take your content and just put it like I was on the board of Stack Overflow for a long time and they would do this. Where they would take, you type in a thing for Stack Overflow instead of clicking on it, they would just show you the answer and remove the click. They'd done that with Wikipedia. They did it with lyric sites. Yeah, but they did with Yelp. They did it in a chat. And people get very upset. Or they with Yelp, they promote their own content on top.

11:07And so there were issues, but it worked, right? Now, the question in an AI world, is if you have these chat bots, if you go and you say, I want an illustration and it just generates an illustration or you say I want a recipe and it gives you a recipe This may be a better user experience by the way. I'm not against it I think it's probably better in the end for the users of the internet But the problem is it breaks the covenant right they took this data These systems were trained on data That was put on the internet under the prior covenant under the premise that they're gonna get traffic back That's right and they can monetize that's right and that was the premise and that was the promise right now And now you have a new system which may not send a traffic.

11:41In fact, it probably won't. If these things can just tell you the answer, why would you click through? And so that's probably where we're headed. It is a world where you have five big systems. Let's call it three to five big AI systems. You go to their websites, they give you an answer. And so what happens to the billion other websites if they aren't getting traffic is a question, right? And I'm surprised slash disappointed that I don't see anyone. I feel like I'm the only person I've been seeing in the business. Yeah. I feel like I'm streaming the abyss. I'm a little bit surprised that the AI people who just, it's fine, they took all the data and they'll be copyright lawsuits and I'm not going to apply them.

12:13Yeah, they've done some data. But aren't we a little bit, even forgetting about the societal questions and all the small businesses that will be, don't we worry about the internet? Because I worry about just the internet. Well, if you have internet, if you have five companies and it becomes broadcast TV in 1970s, there's four channels. Is that the world we want to live in? Is that a world that's pro -start -up, pro -innovation? for creativity, like a long tailed website. It's like that next generation, long tailed website. Yeah, how do you break out? How do you create new things? So I just worry without thinking it through.

12:46And so to me, look, I'm not saying that I have the only answer to it, or you have to be a crypto answer. I realize some people that's controversial, but I think step one is we should say, okay, wait, this breaks all the incentives of the internet. And step two is, you know, is that a good thing? I don't think so. And then so what is the right answer? And should we create new incentives? And this is why a lot of what I've been trying to invest in and think about has been, okay, like the example I gave a story protocol is let's think about new incentive systems to layer on top. One of the things you've talked about is just this trifecta of technology products that have come along at the same time.

13:19So, generative AI, crypto, and new hardware platforms. So, how do you think about the three of those coming together? So, yeah, and the analogy, of course, is like mobile social cloud, the last way where they all ended up reinforcing each other. So, you're already seeing some of this. You have these new devices, the AR and VR glasses and things, which use a lot of AI and sort of per -style kind of stuff. There's a whole area of cryptomic side about called Deepin, which is decentralized physical infrastructure. Most prominent example is a project called Helium. And Helium is a community -owned crowdsourced telecom network that tries to compete with Verizon and AT &T.

13:51And so basically what they did is they created an incentive system where anyone can put a Helium node up in their house and that adds a little bit to the network. It's a wireless transmitter. They got hundreds of thousands of people in the country to put these networks up. And now they offer a cellular service that's, I think, significantly cheaper than something you get from Verizon. It's like 20 bucks a month, sort of 70 bucks a month. And it's cheaper because much of the time, it's using this homegrown network that you don't have to spend tens of billions of dollars to build that. But what's interesting about it is that crypto is very good at creating incentive systems.

14:21And traditionally, the hardest part of a network is the bootstrap phase. Once a network has critical mass, it's clearly valuable. Once I can sign up for the cellular network and use it anywhere in the country, clearly I'll pay for that. When you start it off, when there's only 10 houses with the cellular access, it's not something you want to use. Think of a dating site. If there's 10 people on dating site, you don't want to use it. If there's millions, you do want to use it. This is a classic problem with building networks is how do you get over this early phase when the network affects a week?

14:48Yeah. And so crypto is the perfect complement to that. Crypto is a great way to provide incentives in the early areas of building a network. And it turns out a lot of interesting networks in the world are physical networks. So there's people doing this for climate weather modeling. There's people doing it for mapping self -driving data and mapping cars. People doing it for electric car charging, for cellular networking. We just did one that's around energy, metric monitoring. And there's people doing decentralized science, which is you mix it in with more scientific applications. So one sort of simple heuristic is anywhere where you want to build a network and is a challenge to build the early phases of the network.

15:23Crypto can be a really useful way to help bootstrap that. Oh, interesting. Right. And so that's one of my favorite areas where the physical world and robotics intersecting with data collection and all these other themes at Intersect with AI are relevant. Mark actually gave me this framework which I like a lot which is the AI for Austin or Sugar. You know if the AI is for Austin is a core ingredient. If it's for Austin all the incumbents are going to win because you just slap a chap on your existing product and you've got distribution. You have that like selling reference power, incumbent relationships.

15:53if it's more fundamental of an ingredient, like you can't actually just slap AI into the product you have to build it from scratch. And that favors the newcomers. It's just very TV. We haven't seen anything that tells us what the answer is. The more sealier thing, the more skew more thick, you know, it is, which is a really cycle thing. The more probably favors the incumbents. Another way maybe to frame Mark's thinking is a Clay Christiansen view, is it disruptive or sustaining? And specifically, I think what people misunderstand about Christiansen view, right? Disruptive doesn't just mean new.

16:22It means misaligned with the income and business model. That's sort of the interesting part of his book, right? Is that even when the smart income and sees it coming, it's very, very hard for them to react to it, because it's not what their best customers are asking for. Yeah, exactly. And so that's where I think somewhat overlapped with Mark's frosting, icing thing. Well, it could be that the business model is a fundamentally shifted business model. Yeah, so you come in and you're like, instead of databases, it's some radical new architecture that's database -free. I don't know what. It's something that cannibalizes the income and business model and therefore makes it organizationally and economically harder for the incumbents to layer it on.

16:54Yeah, right. We haven't seen it yet. We've seen people talk about outcome -based pricing. But let's talk quickly about consumer. So in consumer right now, at least, I don't think you see a lot of network effect businesses, right? So like, as successful as the cloud and chat GPs are, I don't think they have a network effect. They're switching costs are relative. Maybe they learn your history. But the question is, how do they avoid in the steady state having just like a model and price competition to the price at the bottom, right? Obviously, they're important big businesses, but will they be dominant?

17:21And then what's the opportunity for new startups? If you're doing venture investing and AI consumer, like you see a lot of these things that make your face prettier, like these kind of fun apps, and they zoom up in the app chart, and then TikTok copies it and so forth, right? Because it's just not, because again, the network of networks. No, no, that's not true. And there's just technique, kind of strategy. I like to talk about, called Come For The Tools, Stay For The Network. And the idea is maybe you can use that, make my face prettier, and then use that as a hook to get people into your new network, like your social network, possibly, although it just feels very, very hard today, given the scale and power of these incumbents.

17:55And that, by the way, will intersect back to crypto, because what crypto is, and what I argue in my book, is that crypto is a new way to build networks. And so, you know, you sort of have the chocolate and peanut butter. You have AI with all these really interesting use cases, and then you have this new technique for building networks. AI, the interesting use cases, but no network effect. And then you have this new thing that's like all network effects are their interesting ways to combine them. But before I get to that, I think it's important to talk about how big technologies roll out in multiple stages.

18:18So there's a distinction, it's not my distinction, but I've talked about a lot. It's sort of one way to think about technologies that they can do one of two things. They can do old things better or they can do new things you couldn't do before. We call the first one SQ Morphic. This is a Steve Jobs term, which sort of refers to products and designs that kind of harken back to a prior era to make them more understandable. And then there's what we call native apps, which are things which are the kind of new things that couldn't be done before. And then there's There's actually a third stage, I think, which is second order effects, which is you created the car, and now you have the highway system, and now you're able to create suburbs and trucking infrastructure, right?

18:55Those are second order downstream effects. There's a famous line that good science fiction writers predict the car, great science fiction writers predict the traffic jam, right? So like, it's like that idea. So it's like that. Like, what are the second order? Like, yeah. Bitcoin is something that couldn't have existed before social networking, right? So 30 years ago, you say, some day people are going to have their own media, and you're remove these gatekeepers, who would have thought, then you're gonna create these digital currents. There would have been no way to create the community. Yeah, yeah, it would have been a nerd times bullet, saying it's stupid, and then it has to end of it, right?

19:24And then it just nowhere to get together and talk about it and create, or I mean, to really kind of religious movements, you know, most token communities, and they need places to congregate and discuss it. And now they have that. And so there's all these kind of second order, meaning we're seeing effects in politics and all these other things. There's the whole arguably our society and world is changing as a second order effect of social networking. So one way to think about, AI. So the first stage is the Schymorphic phase, which is this is the stuff you see talked about all the time in the business and startup community of like your customer service spots, right?

19:52You take a job that's currently done by a person sitting in a call center and you replace that with an AI voice and chatbot, right? In the simplest case, it's a one -to -one exchange. It's cheaper and is more systematic and it will displace jobs. Hopefully it will also create equal or more jobs and better jobs. But that's sort of an obvious for stage and look and this is I think one of the reasons people get so excited about the opportunity for AI is you can just see that happening in I don't know tens of millions of jobs I guess like the whole laptop middle kind of section of the economy and you can see many of those jobs everyone including us who spend their days typing emails are hard to listen to.

20:31It's like we can speculate on it but we're part of that group too. So that's phase one right? Just give more. But that's phase one can last 20 years so just to be clear. Yeah. The next phase is the native phase, and to me, that's what gets me more excited. And by the way, let me give you a little analogy to the internet. So the Schumorific phase with the 90s. And so basically, if you look at 90s internet, people were taking offline things like magazines and catalogs and putting them online. So you would go buy things, you know, and it was much easier. You could type in a website and go buy this rare book on Amazon, and it was much easier and it was convenient.

21:00But it was fundamentally something you could have done before. It just would have been clumsy and getting some weird magazine, some catalog or something. But it wasn't until the 2000s that people did things like social networking. And these things were just brand new things. There's no analog in the offline world to a lot of these new behaviors that people created. I talked a lot in detail about this in the book if people aren't should. So anyways, so you saw the internet play out that way. 93 was mosaic and 2000, I would say five -ish was sort of YouTube and before, I think, with Facebook or whatever it was.