--:-- Anish Acharya: Welcome to the A16Z Consumer Pod. I'm super excited and honored to have my partner Chris Dixon here today. Chris, you're probably best known for your work in Web3 and network economies recently, but what folks may not know is that you led a lot of the most important consumer investments at Andreessen Horowitz. Prior to that, you also founded two consumer companies. I thought a fun place to start would be networks. That feels like the first place you really cut your teeth. So maybe talk about your investments in Stack Overflow, Pinterest, Instagram, and how you generally think about consumer networks.
--:-- Chris Dixon: So many of the most important internet services are networks. Going back to the early internet, email and the World Wide Web, which are still around and really important, are networks. They're networks in the sense that the service gets more valuable as more people use it. If you were the only one on email, it wouldn't be particularly valuable.
--:-- During the rise of the internet in the 90s and 2000s, that's when you had things like YouTube, Facebook, and later on Instagram and a whole bunch of other really important networks. If you're an entrepreneur or an investor during that period, they tend to be very valuable companies. They're very hard to build, and we can talk about that later. There are different tactics and strategies for doing that.
--:-- My background: I started two companies. First was a consumer security company, and the second was a consumer AI company. Then I was a personal investor. I co-founded a seed fund called Founder Collective, which was an investor in things like Uber, Venmo, and Stack Overflow, as you mentioned. I sort of revolved in a bunch of these networks as the internet evolved.
--:-- To really talk about networks, though, it's important to step back. The way I think about it, a foundational question in tech is: why do you have these companies come out of nowhere and end up being very impactful, having hundreds of millions or billions of users, being very valuable, in a way that you typically don't see in other industries? What's fundamentally different about tech?
--:-- I think the answer is that in tech, you have some very strong exponential, superlinear forces. The most famous example is Moore's Law. Moore's Law is the idea that roughly every two years or 18 months, the performance of semiconductors doubles. It's a rough approximation, but it's basically been true. You've seen this compounding improvement in processor performance. I think there's also a broader Moore's Law, which is storage, networking — all the computing resources have gotten much better, which is why you have things like mobile phones.
--:-- If you go back and look pre-iPhone, mobile phones were pretty junky and limited in capability. They didn't have touchscreens and had poor performance. Steve Jobs and Apple's great insight was that they saw this curve — they saw this exponential curve and they rode that curve. The first iPhone was also quite limited, but part of their brilliance was seeing that trajectory.
--:-- So Moore's Law is a very important exponential curve. But there are two other really important exponential curves in software. One is what I call composability. Composability is really what's accounted for the rise of open source software. Why did Linux go from a hobby project in the 90s to the dominant operating system in the world today? A lot of it is composability. Composability means the software is open source. Anyone can contribute to it. Very importantly, you can harness the collective intelligence of the internet as opposed to only relying on your employees. Anyone in the world can — the famous phrase is "all bugs are shallow with enough eyeballs." Really importantly, with open source software, it becomes like Lego bricks where anyone can take a piece and reuse it. You get this compounding exponential improvement and growth.
--:-- The third really important exponential force in tech is network effects, which is why networks are so important. They start off often quite limited. Facebook was just at Harvard, essentially a real-time yearbook for students at one school. Then it hopped by lily pads to other schools and high schools and eventually to global domination. Mark Zuckerberg and the team saw this power of network effects and rode those network effects.
--:-- Clay Christensen calls this disruptive technologies. It's a puzzle of why in tech you have these very strong incumbents who seem to miss things. You could tell the story today about Intel and NVIDIA, or even ChatGPT and Google.
--:-- Anish Acharya: That's a great example.
--:-- Chris Dixon: Neural networks 10 years ago were toys. They were cool, and a bunch of people saw the potential, but the reality is they just didn't work that well. I remember there was a chatbot VC thing in, I want to say, 2016 or something.
--:-- Anish Acharya: Yeah, of course. Chatbots had a moment back then.
--:-- Chris Dixon: They had a moment. But the reality is they just weren't that good. They just couldn't do the job. But of course, they got much better. The genius of OpenAI and other pioneers in the space was to make that bet.
--:-- Anish Acharya: That's right.
--:-- Chris Dixon: And then Google today is in an awkward position because they have this huge incumbent business that depends on sponsored links, and they're trying to layer in AI. In some ways, it didn't come out of nowhere, but it grew faster than even some of the optimists predicted. It improved faster.
--:-- The big takeaway here is that whether you're an investor or entrepreneur, the most important thing to start with is to look for these forces, to look for these exponential forces. One of the lessons I learned in my career was that you can do all sorts of tactical product things, but these forces are going to overwhelm you, for better or worse. The first thing to understand is the landscape of these forces, how they're moving, and how you can hopefully be on the right side of them.
--:-- Anish Acharya: Chris, how intentional do you think you have to be as a founder about building? You're building a tool — do you have to be thinking about the network a priori, or can the network emerge? Because in AI so far, we've seen a lot of tools and not a lot of networks. In hindsight, everybody was designing a network from day zero. What's your instinct?
--:-- Chris Dixon: That's a great question. I wrote a blog post years ago called "Come for the Tools, Stay for the Network." The idea was a tactical pattern I observed among entrepreneurs. I cited Instagram as an example. Young people remember this, but Instagram's network was not a big part of the product initially. It had a button where you could share an Instagram, but why would you do that? No one was on it.
--:-- Two things happened. One, they had these cool filters, which at the time you had to pay for in other services, and they gave them away for free. Just effects or lenses or whatever you want to call them. Secondly, they piggybacked off other networks. You would share to Twitter. Then a year or two later, Twitter blocked them.
--:-- You see that today with Substack. Substack starts off piggybacking on the email network, on Twitter. I'm not a firm's investor, I'm not personally involved, but my sense is they're now getting traction with their own network. You go to the Substack app.
--:-- I think it's a similar tactic. You can see some of this "come for the tool, stay for the network" with modern productivity tools like Figma and Notion. They're useful single player — you can go to Notion and it's a really nice way to edit a document, or Figma to do design. But there are also social features that become essential.
--:-- These things are all degrees. Google Docs — I love it, I use the social features. But is it really a network? I could probably switch and just share links with somebody else. The social features layer on. Some products like Instagram, it becomes essential. You simply can't leave Instagram if you have a following and you want to keep that following. It varies by use case.
--:-- By the way, I think you see some of this now in Stripe doing the Link product, which is a payment app. I think Shopify and the Shop product. There's a really nice user experience — I don't have to type in my credit card again. Now there's a network. Shopify originally was just a tool for a merchant to get online.
--:-- Anish Acharya: That's right.
--:-- Chris Dixon: So I think it's a really powerful tactic. Network effects cut both ways. They're great when you have them, but they're really hard at the beginning. No one wants to be on a dating site with two people. So how do you make these things useful from day one? But the problem with single player is it's just hard to defend.
--:-- I think you're seeing this in AI now today. You know much better, but you're seeing a lot of really cool tools because it's an amazing technology. But then it's like, okay, you can change your face app or whatever, but how does it move beyond faddishness? How does it move to something that really engages people over a long period of time? Often the answer in consumer products is networks. So you have to layer in a network. The challenge is you don't want to just layer it in for the sake of it. You need it to actually be useful. So, I'd love to hear from you. What are you seeing in that area?
--:-- Anish Acharya: It's actually interesting because it feels like the big networks have become hyper-sensitized to this idea of new networks emerging that were bootstrapped on their networks. Twitter of 10 years ago would have been a lot more asleep at the wheel to the threat of a Substack. They were pretty aware of this potentially happening. Facebook has de-platformed a ton of companies that they thought were going to do this, including Instagram and others.
--:-- So one, I think the networks are more sensitive. And on the tool side, because the tools have been specializing in their own directions, and part of it is product features, but part of it even for some of the multimodal tools is aesthetics. Midjourney just has a different aesthetic than Ideogram. So they can both coexist and they're not directly competing. Even though the tools are seemingly substitutes, so far we haven't seen that trade-off and they're all working. Maybe that's just where we are in the product cycle.
--:-- But I do think it's a topic for a lot of AI founders: there's not an obvious network to build around a lot of these tools. How much of that should be pre-designed versus let's just keep pushing the edge and the network will emerge? It will show up in two ways. One would be in the usage — you might see some of these tools not get used as much. But the other is in pricing. Even if you carve out a niche, how much more are people willing to pay for that niche versus competitors?
--:-- Anish Acharya: Prices have been going up, interestingly. Google's top SKU is $250 a month. Grok is $300 a month. I don't think we've ever seen a time where consumers were paying those kinds of prices. One of our extreme views here is that the future of consumer disposable income will be like food, rent, software. Software is going to subsume a lot of the other areas of discretionary spend today.
--:-- Chris Dixon: It's also possible — I've always suspected in tech, in Silicon Valley, we underestimate the power of brands and consumer inertia. I think you're seeing that today with ChatGPT. It's such a household name, almost overnight, that even though it doesn't have in this technical sense maybe network effects, memory and things, but it's more stickiness. The brand effects are so powerful. You become known. Cursor is known as the best vibe coding platform or whatever.
--:-- Anish Acharya: That's right. I was going to ask you about that, Chris. Network effect is a gold standard for defensibility. You've talked a little bit about how brand is underappreciated. You just mentioned it. Do you think being a high NPS, high DAU product is enough of a moat? Or do you think we really have to push for building around these compounding forces?
--:-- Chris Dixon: It's a really interesting question. One argument would be that maybe a lot of the network effect has been externalized to the internet. We were having this discussion at one of our partner offsites. The idea being, you're Cursor, and suddenly it becomes popular. Or Midjourney. Then you get all these Midjourney influencers, YouTube videos, websites, how-to guides. You still, in some sense, have a network effect, but it's not a network effect that's in the product itself. It's externalized to the internet.
--:-- Maybe that's a difference now. The era I was discussing was the era when the internet was being built. In some ways, the internet is built now. I'm sure it'll improve and change, but it's built. It's 5 billion users. Maybe the rules are different now. Maybe that effect of getting all of those adjacent networks around you gives you, in a sense, a network effect. You try to show up top in search. ChatGPT recommends you. The algorithms feature you.
--:-- There's a soft sense of a brand. People have heard of you. But it's also this whole giant system with all of these different interconnecting networks that might strongly favor those products. Then it becomes a timing thing. Getting in early, being the first to own the meme, the category, and get that effect going. Then maintaining it through product velocity and high quality and everything else, which is non-trivial. It's very hard to do, particularly in AI — you tell me — to always stay on the cutting edge. It's expensive, a lot of capital.
--:-- That's another thing, by the way, the capital effects in AI. You do well, you raise the most money. I assume the people raising a billion have already proven a bunch of things. At some point, the capital becomes a moat.
--:-- Anish Acharya: A hundred percent. It's very interesting because there's this barbelling that's happening even in software where the bigs are getting bigger, but we're also seeing single-person, 100 million run rate companies coming, or maybe it's already here. But certainly the bigs are getting bigger and capital is a part of that. Maybe the market's just so big that the answer is both, all of the above. It may just be, as you said, it becomes like food and rent. Software is moving beyond the quote-unquote software budget. It really hasn't been zero sum so far. It's been shocking — prices are going up and everything feels like it's working. So maybe we'll look back and say that was a sign, but so far so good.
--:-- Chris, since you mentioned Vibe Coding, I thought it would be fun to talk about movements. It feels like you've been early to a bunch of movements, products like Coinbase and MakerBot. Those felt like niche communities on the internet when you started paying attention to them. How do you think about investing in movements? How do you think about building around them when there are questions around: is this a toy? Is this structural? Is it durable or ephemeral? Maybe talk a bit about that.
--:-- Chris Dixon: It's a little bit to the point we were talking about, about networks becoming externalized. I used to spend a lot of time, 10 to 15 years ago, on subreddits and niche communities. Partly because I'm interested in that stuff, and partly because I think they're very powerful. If you look at Wikipedia, Stack Overflow, a lot of these interesting community sites, they're often 20,000 people. They aren't the millions you might think. There are millions doing a little bit here and there.
--:-- If you look at open source software and crypto projects, a lot of things that have been popular movements that grew were really led by a relatively small — on the internet scale — relatively small group of hardcore enthusiasts who are really smart, often technical.
--:-- There's a famous old quote, I think by William Gibson: "The future is already here. It's just not evenly distributed." I've always believed that. If you go back historically, that's the case. Talking about neural networks, that's been going on since 1943. There have been communities of people, including a lot of the people that lead the labs today, who 15 years ago were seen as niche. Neural networks weren't the dominant approach.
--:-- With that thesis, you want to find the next big thing. One way to do it is to look around and see where these hyper-enthusiastic, sometimes cultish communities are. They have their own language, their own norms, a sense of insider versus outsider. I got into that a while ago, and that's how I got into Bitcoin originally. I just followed those people. I found it was one of those things where it sounded kind of silly at first, and then as you learn more about it, it seemed a lot more interesting. That's always an interesting feature. Some things you learn more about and they aren't that interesting. Some things are kind of silly — the conspiracy theories that the earth is flat, or the moon landing conspiracies. You dig into that stuff and it's just crazy. Whereas you dig into this stuff and you don't have to agree with everything, but there are smart people. It's very interesting.
--:-- For me, it was 3D printing. That led to my investment in Oculus and Coinbase, both from that thesis. Seeing the developers and the Kickstarter community enthusiasm around when Palmer Luckey was first creating VR. I also got into nootropics, which led to an investment in things like Soylent. When I joined the firm in 2013, drones — we did a few investments around that. Just looking at these interesting hobby communities.
--:-- There are a bunch of reasons why it's an interesting way to look at it. One is those are the people that create these things. If you have 20,000 interesting technologists, they often build things. They're going to build some interesting products. It's also a great marketing engine. They're out there. They often have outsized influence on the internet. They have followings. They help get the energy going, build things, and market them.
--:-- It's not foolproof. It's hard because a lot of these things just end up being niche or don't have — going back to the exponential forces — nootropics is still a thing that's around, but I don't think it's created a big tech company as far as I know. I think it's partly because it's got linear forces, not exponential forces behind it. There's not some engine exponentially driving it to have better and better products.
--:-- Anish Acharya: Maybe actually though, if you look at a company like Function Health, Function Health is sort of the catalyst for this huge movement, consumer movement around health and quantified self. Nootropics was a bit of the predecessor to that. In a sense, there is this slow exponential and then very rapid uptake. Timing is a really interesting question here because with these movements, you don't know if they're going to play out over 100 years or 100 days sometimes.
--:-- Chris Dixon: That's right. I don't know that. You know much more about Function Health, but that's interesting. You're right. 3D printing is a good example where it's still around. It didn't get as big as people had hoped. I had an investment in MakerBot back then, which was a leader and got acquired. It's still a hobbyist thing. It's interesting. I think the limiting thing is it's in the physical world. There isn't a Moore's Law driving it. That said, I expect that over 50 years or something, it will become a more important thing. You're right. It could just be a timing thing.
--:-- Anish Acharya: The Vibe Coding thing, to come back to that, feels like this irreversible consumer phenomenon where everybody is maybe not quite programming, but creating software in a way that they weren't 10 years ago. How do you think of that as a decentralizing force? You talked about the economics of software versus the means of production. The means of production are getting decentralized through these new tools like Replit and Cursor. Is that sufficient to lead to a renaissance of the open web? What are the second-order implications of everybody programming?
--:-- Chris Dixon: It's a great question. The internet has become increasingly consolidated. If you look at metrics like the amount of revenue generated, the traffic — it's more and more. It's like 95% plus of both of those metrics are now in five to ten companies' hands. You can make an argument either way. With AI, we're already seeing this in the data. A lot of AI obviates the need to click through and go to a website. I think there was just a report out that a bunch of travel sites and others were seeing some alarming drops in SEO.
--:-- It's a mixed thing. On the one hand, I'm a user of ChatGPT, and it's amazing to just get an answer and not have to go searching and go through all these websites. It's a vicious cycle where websites lose traffic, they get more desperate, they put up pop-up ads and other things, and it becomes an even worse experience. This has been going on for like 10 years, this kind of negative flywheel.
--:-- On the one hand, it's great for consumers. Vibe Coding — we were investors in Stack Overflow, which got acquired, but I think their traffic has dropped a lot because of Vibe Coding. Some of the training data probably came from Stack Overflow and GitHub and places, but then it becomes better. I've used Cursor to do some fun projects. It's an unbelievable tool. I think it's clearly good for the world. It is bad for those websites.
--:-- I hope what we're seeing is a renaissance of paid software businesses that don't need to dominate the internet and be Facebook, but can get to hundreds of millions in revenue. I think we're seeing this.
--:-- Anish Acharya: Right.
--:-- Chris Dixon: From an entrepreneur's perspective, it's a very exciting time. I think we can see a lot of great products. It's a great time for consumers. Maybe that will change over time. Maybe they'll need to layer in ads and the incentives will shift and do things that are more adversarial towards consumers. I think right now I like the AI products in that they feel very aligned with users. They're really just genuinely trying to get great products and charge for them.
--:-- Anish Acharya: Exactly. We sort of call it the emergence of narrow startups where they charge high prices and deliver exceptional value. Maybe a controversial statement right now is that there are no marketing problems, only product problems, because the technology allows you to be so ambitious on behalf of your customer. The costs actually ironically lead to better business models because consumer founders need to think about monetizing early. Otherwise, they're just going to go to business. It does feel like there's a renaissance in paid software that's happening that makes it a more fun time to build than five years ago. Do you think that over time that will shift potentially because people will realize that the low-hanging fruit is picked, the higher paying consumers, and to get the rest, you need to layer in different business models, ad-based business models and so forth?
--:-- Anish Acharya: I don't know. It feels like there's so many more consumer needs that are addressable, and they're addressable in such a significant way by the technology that you can actually specialize and go very, very deep. There's AI therapy generally, then there's AI therapy for people that have ADHD. And then there's people who have ADHD that are in a certain life stage that perhaps want to interact in a certain way. You can just go extraordinarily deep. So I don't know if it leads to consolidation over time or if you can continue to specialize and for a small number of people be their primary.
--:-- Anish Acharya: That actually might lead to a good topic around the idea mazes. Chris, you've talked a bunch about platform shifts. You've invested around platform shifts. You predicted them. One of the interesting things about this platform shift is that the properties of the platform are sort of emergent. They're not explicitly defined by Apple as iOS was. They're things that founders, and even the people training the models, are discovering. Does that sort of change your mental model around platform shift? And maybe how similar or dissimilar is that to Web3?
--:-- Chris Dixon: Yeah, the idea maze concept originally came from our friend Balaji Srinivasan. I wrote about it a while ago. The way I think about the idea is that there was this old debate: with startups, are the ideas more important or the execution? With the idea maze, I think it says they're both important in the sense that it matters which maze you enter. I'm entering the AI maze for healthcare, or I'm entering the AI maze for image generation. You go in with an initial product idea, and clearly that matters. But it also matters that it's a maze, meaning it's dynamic. The world will shift. You can't predict it.
--:-- Chris Dixon: The canonical example in my mind is Netflix. Netflix started off mailing CDs. The hypothesis was that movies will become—the internet has changed the way people consume movies. People will subscribe to them. But today we need to send them by mail. Then over time they pivoted to digital distribution. Then they started getting pushback from the content providers and they pivoted to original content. They really did two almost complete company pivots, but their core maze was right. The core maze was that the internet will lead to subscription movies in some broad sense. But they were extremely agile with respect to the implementation.
--:-- Chris Dixon: To me, that's the idea maze concept. You're entering a maze. As an investor and as a founder, you need to think: am I a person who wants to be in this maze for 10 years? Am I willing to be agile and often persevere through difficult periods? It's often emotionally challenging, not just intellectually challenging. That's the life of a startup.
--:-- Chris Dixon: When you think about AI, we have a very clear megatrend of AI being intelligence. It's a very broad and important technology. Everyone knows that. Secondly, you have these scaling laws which seem to be quite powerful. The models are getting much better. An important distinction is there's specific scaling things like LLM pre-training, which people may have debates about—at what point do you have diminishing returns? Maybe we're hitting that, I don't know. I defer to the experts. But then there's the meta process. The meta process is AI overall. There's people working on reinforcement learning and 100 different techniques. The meta process is really an economic phenomenon: there are all these smart people, there's business models behind it, there's funding. There's not just one process; there are many processes being explored.
--:-- Chris Dixon: It reminds me of Moore's Law. From the outside, Moore's Law naively seems like semiconductors magically get better every two years. If you read books about it, from their perspective, they run some fabrication technique, hit a wall, freak out, and then some brilliant person from another lab comes up with a new fabrication technique. Each process would run and have diminishing returns and asymptote at some point. But the meta process, the bigger industry flywheel, did not. It led to this smooth growth. My sense is AI is in that semiconductor-like place where you have this meta process that's very likely to continue scaling exponentially for a very long time.
--:-- Chris Dixon: That creates a huge opportunity for entrepreneurs. It's also a challenge. The opportunity is you can build things where the capabilities will grow. There'll be all these new opportunities. The challenge is: are the incumbent models going to be sort of God models that subsume your use cases? How do you play that? What you're seeing is people say, "I'm going to go so deep on a domain that that will be my edge. I know everything about this specific domain. No matter what the incumbent models do, I'll always be able to have an edge in my product, or I'll have such good brand recognition or strong user base or reference selling." That's the threat.
--:-- Chris Dixon: If you go back to the semiconductor analogy, the canonical case study in Clay Christensen's Innovator's Dilemma is the PC industry, the hard drive makers. It was a very fruit fly Darwinian struggle where you had thousands of companies and very short life cycles for a lot of them. But also a lot of very successful companies. It may be a very brutal process for entrepreneurs in the sense of a lot of competition, a lot of other smart people, a very dynamic idea maze, but also massive opportunity.
--:-- Anish Acharya: How do you think about native versus skeuomorphic technologies in that context? Everything is changing, especially when you're building for a consumer. Does the consumer change their preferences when they get this magical new technology? Or does the emergence of the native technologies also depend on consumer preferences changing and being informed by these external forces about things like AI?
--:-- Chris Dixon: Great question. Let me define the term. Skeuomorphic is a term Steve Jobs used to use with respect to design. He liked design like the original bookshelf app on the iPhone that had grainy stuff on the background. Design that harkens back to a different form factor. There is a common pattern in technology and media: when you have a new platform or media form develop, people start off imitating prior media forms. Early films were shot sort of like plays with a camera and a better distribution model. Then people invented a native grammar of film—close-ups, establishing shots, all those kinds of things.
--:-- Chris Dixon: Early internet in the 90s looked like you take a commerce catalog and put it online, or a brochure and put it online. It took 10 to 15 years before you had things like YouTube and modern social networking—things that really couldn't have existed prior to the internet, like user-generated content where anyone can upload a video. Some of it is the technology. YouTube couldn't have existed until you had really wide broadband penetration. Some of it is the underlying technology takes a while to get there. YouTube also started off as just funny viral videos. A lot of it was copyright violations. It took a while to develop native YouTubers, content creators. That's often a generational thing. It literally is a new generation sometimes—people that don't look at the technology as a threat but as an opportunity.
--:-- Chris Dixon: Part of it is entrepreneurs just have to figure out the idea maze. They have to figure out what people want. At YouTube's time, there were a lot of debates. People said, "Tastes aren't going to change. Why would people want to watch four people joke around?" There were analogs—is that like talk radio? But they just really didn't understand. I don't think humans fundamentally changed in a deeper sense, but it was understanding the capabilities of technology, the cultural shifts around it, the network effect around it.
--:-- Chris Dixon: With AI, a really interesting question—and I'm sure you've thought much more deeply about it than I have—is that most likely we're in a skeuomorphic phase right now. What is the native phase going to look like? I like the native phase better because it's kind of crazier and more interesting. If you look at image generation, they're kind of just basically taking what illustrators do. But one cool thing with photography is that when it first came along, it seemed like a threat to representative painting. You saw art move to more abstract art to get away from that. There was a lot of hand-wringing about whether this would cheapen the art form. But an interesting thing happened: a new art form emerged, which is film. Photographs were the skeuomorphic app of cameras, but film was a native one. You had a new art form.
--:-- Chris Dixon: I wonder about that with AI. Right now you have image generation, which is taking what a human might do and automating it, and movie generation and other videos we see online. But is there a new medium that hasn't emerged yet? Maybe it's virtual worlds or something. There are probably a bunch of hypotheses. But my experience has been that it's often surprising and hard to predict. That's where a lot of the cool, creative, interesting stuff comes in. It may take another generation or five to 10 years for a new set of AI-native kids to grow up.
--:-- Anish Acharya: That's right. It's actually really interesting because in a sense, we're in the command line era of AI. There are some things you can articulate well with words. But if I describe to you what kind of music you like, it's hard to say. Most people lack the language to articulate the art that they love. Even the idea of prompt to media feels skeuomorphic. There's got to be a more native way to explore it. I don't know what that looks like yet, but I'd be surprised if it's prompts in the long term.
--:-- Chris Dixon: People are now calling it context engineering, not prompt engineering, which I think is a nice rephrasing. That is kind of what you're doing. You're taking the fact that all of this stuff I do in the real world that ChatGPT isn't able to see, and I'm trying to summarize all that knowledge that's hidden to it—the context—and put it in there. It does feel like something that should be automated. If you have these intelligent machines, I think that's what people are doing with these potentially new ambient devices people are creating.
--:-- Anish Acharya: Even in the media case, my Spotify library is probably much more useful for generating music that I like versus my articulation of it.
--:-- Chris Dixon: That's right. I think we're in a different era now. I've come to believe that we're sort of in a different epoch. The internet is built. Some of these things like network effects, maybe they're less important now because it's in the network itself, it's externalized. Some of these dogmas that people like I have as entrepreneurs or investors have believed for 20 years are changing. In that sense, experience can be a hindrance.
--:-- Anish Acharya: You'd mentioned in another podcast that if you had one issue to get passionate about in the world of AI, it was open source and open source AI. Do you want to speak to that for a moment?
--:-- Chris Dixon: We were talking earlier about the democratization of the web and how consolidated the internet is. I would argue that open source software has been an incredibly important force for democratizing technology. The reason you can get an Android phone for $10 and get on the internet so cheaply is that basically all the software is free. Imagine if operating system providers used to charge $100, and you'd be paying that on clients and on the back end. Instead, you're not. Most internet users—the vast majority of the bits being hit—are open source. It also is what makes startups exist. We can fund startups and they can spend hundreds of thousands of dollars, or even less sometimes, and be up and running with really competitive, great software. That's because of open source.
--:-- Chris Dixon: On a policy side as a firm, we've been big advocates for making sure open source is around and competitive. That means not banning it. There are bills out there, particularly at the state level, that want to put in de facto bans. For example, California had a bill that would have created unlimited downstream liability for software developers, which would have effectively killed open source. That's step number one. Step number two is whether the incentives are there to create open source.
--:-- Chris Dixon: I watched an interview with Satya from Microsoft recently. It was a really good interview. He argued that open source will always exist because enterprise customers always demand at least one open source alternative. They'll end up funding it. That's why you always see this proprietary open source combo. But then you have Facebook doing it with Llama. I don't know if they'll continue to do that. There are some startups doing it. China has been very into open source. Maybe that's a national strategy. Maybe that changes things. Maybe you do it at first to create attention and marketing, and then you change it.
--:-- Chris Dixon: The thing with AI that's different than operating systems is that with operating systems and databases, you just needed a bunch of coders sitting around. With AI, you need massive capital expenditure to train the models. So I just don't know. I think it's an unknown question long term. Are there good steady state funding models for open source? I think a possible outcome, which is a pretty good outcome, is that open source is just always a little bit behind, the way OpenAI is now releasing older models. I think that's probably a fine outcome for startups to exist, for consumers. We want consumers to get inexpensive healthcare advice. The next best model in five years will probably be good enough. For most startups, it will probably be good enough. For the super high-end stuff, people pay for it. Maybe that's a good equilibrium state. Maybe that's where we're headed. I hope so. It would be a bad outcome if you had four companies that had vastly better closed-source technology and could effectively charge rent to consumers and startups.
--:-- Anish Acharya: I agree. I think a lot about the early ethos of Android, which felt like it matched Google's open-web mindset. Then when it became clear that iOS was beating their pants off by being a closed ecosystem, Android became very closed and started to mimic the closed iOS strategies. We'll see what happens with Meta and Llama if they replicate that. That's a worrying dynamic. The more optimistic case is that we haven't yet seen the same sort of app platform feedback loop and the lock-in that you get from the foundation models. There is a case for them to continue to release the next best model and for the models to be somewhat substitutes for each other so far.
--:-- Chris Dixon: The Android case is a good cautionary tale. In some technical sense, some of the code is open source, but it de facto isn't. All the services, everything else, permissions. They made lots of overtures that way. That would be the worry. But it does feel a lot better than it did three years ago with the China open source stuff. The policy stuff is better. We're seeing OpenAI doing older models. It seems like we're in a better spot for open source. Some of that is the scaremongering that a chatbot is going to murder everyone. Literally zero people have died from ChatGPT so far. I think people are chilling out on that. We're in a much better spot. I'm cautiously optimistic on open source.
--:-- Anish Acharya: Two years ago, the conversation was a lot about if OpenAI is the only game in town over time, they take all the economics of the complements. It doesn't feel like that's happened. To your point in the amazing book about ballooning, it feels like that's why there's a lot of interest in acquiring IDEs. They understand that if the foundation models start to become more interchangeable, they're going to have to move upstream and own user-facing economics.
--:-- Anish Acharya: Amazing. Chris, thank you so much. It's great to hear you talk about consumer and AI and all the implications. We're super thankful to have you at the firm.
--:-- Chris Dixon: Well, thank you. This was fun.
--:-- [Outro music begins]
--:-- Host: 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.
--:-- [Outro music fades]
--:-- Narrator: 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. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com forward slash disclosures. Thank you.
