a16z Growth Partners Dialogue a16z crypto Partners: How does AI and Crypto change the future of the Internet?

Introduction: David George, a partner at a16z Growth, talks with Chris Dixon, a16z crypto partner, to explore their vision for the new Internet, including crypto decentralized AI infrastructure; launching the network effect, and AI will become the native media form of this era, etc. This conversation also explores why the Internet’s initial business model is being solved and how the new Internet introduces a completely new business model for creators.

How technology evolves

"David George": You are now focusing on the crypto field. What do you think about the relationship between encryption technology and AI?

Chris Dixon: My macro perspective is that technology waves often appear in pairs or in three. ⼗Five years ago, mobile Internet, social networks and cloud computing were the three major trends. The mobile Internet has increased users with computing devices from hundreds of millions to hundreds of millions; social networks are the "killer applications" that attract users; and cloud computing is the infrastructure that supports all of this. These three are interdependent and indispensable. People were arguing about which one was better, but they all turned out to be important.

"David George": Yes, they are all necessary.

Chris Dixon: I think AI, encryption technology and new devices (such as robots, autonomous cars and VR) are the three most interesting trends at the moment. They also complement each other and develop together. Encryption is a new thing (this is what I have in this book) and it provides a completely new way of internet architecture to build networks. It has some unique features that can make things that were impossible in the past come true. Many people think of Bitcoin or meme coins when they mention encryption. But for me and many professionals who really understand encryption, the essence of encryption is far more than that. It has a lot of intersection with AI. One of the most basic combination methods is to use encryption architecture to build an AI system. We have invested a lot in this direction.

We have discussed a core issue within the company: Will the future of AI be controlled by a few large companies, or will it be jointly managed by a more general community? The question that this is: Is AI open source? It really shocked me that the AI ​​field has become so closed. ⼗A few years ago, all AI research was public and published in papers. But later, the industry suddenly became closed. They claim to do this for safety reasons, but I think it's for their own competitive advantage. Fortunately, there are still some open source projects, such as Llama, Flux and Mistral. But I'm a little worried that this open source model is a bit fragile because many projects don't disclose their model weights. Are these really open source? Some models are open source, but their data pipelines are not. Can it be reproducible really? They may change the model tomorrow, and you can't do anything about it. These AI models are improving every time, but if they no longer stay in the forefront, I don't know what to do.

"David George": At least for the moment, AI relies heavily on large companies.

How cryptocurrencies interact with AI

Chris Dixon: Some of the projects we invest in focus on building a decentralized Internet service architecture for the AI ​​ecosystem. For example, a project named Jensen is building a decentralized computing resource network. Its model is similar to Airbnb, allowing users to submit computing tasks and allocate them to idle computing resources around the world, thereby optimizing the supply and demand of computing power. This network is like an economic ledger, managing the supply and demand of computing resources.

Another example is Story Protocol, a new way to register intellectual property rights. Assuming you are a creator, you can register images, videos, or music onto the blockchain, which records the media and all rights. It uses existing copyright laws to clarify its copyright ownership. That way, anyone can use these contents while complying with the agreement, anyone can come and you might say, “You can use this mix, you can create derivative works, but you have to pay me 10% of the revenue.”

"David George": ...or any proportion.

Chris Dixon: On the blockchain you can set terms and create an open market. But in the current market, you can only contact the company yourself and try to negotiate. This leads people to either steal content, or use it without using it, or only large companies can reach copyright transactions. For example, OpenAI paid Shutterstock $100 million, and the blockchain created a broad master resource, and small creators could set their own terms.

One of the core advantages of encryption technology is composability. Open source software is largely successful because it allows developers to combine and superimpose innovations on existing modules. Linux is a good example, from nearly 0% of the market share in the 1990s to now occupying more than 90% of the server market, because of its composability. People contribute (even small) to the system to make it better. This is also like Wikipedia as a knowledge integration system.

Speaking of Story Protocol, it also allows creative content to be combined as freely as Lego. For example, if someone creates a role, another person writes up stories, and someone uses AI to generate animations, you can create a new superhero universe, and as long as the funds flow back, everyone will get a share of the pie.

"David George": The key to this model is that the flow of funds is transparent and fair.

Chris Dixon: This way creators can use AI tools to improve efficiency while gaining financial returns instead of being free to use. It's a great vision -it inspires people to use these new tools while providing economic models. We often think about how to find new economic models for creative workers in an AI-driven world. This is the area that excites me the most about the intersection of AI+Crypto.

"David George": In the past, social platforms have obtained 100% of advertising revenue, while creators can only rely on traffic monetization. And what we want to see is a new system in which creators can freely price and trade. This can drive more innovation.

"David George": Because economic incentives are consistent.

Chris Dixon: Based on this, we are seeing more of this 'crowdsourcing' way to do AI. Judging from the data level, AI needs more data. The breakthrough point of encryption technology is that it can design new incentive systems. The key is how we use these systems to collect more AI training data? Data can be used as input to AI, for model evaluation, or other purposes. This is similar to what Scale AI does, but the difference is that we want to do it in a decentralized way, rather than being controlled by a centralized company.

One of the projects we invested in is WorldCoin, a project co-created by Sam Altman. Its core idea is that in a world where AI can fake human identities and content, we need a way to prove that a person is real, and the best way is to use blockchain to complete identity verification using encryption technology. WorldCoin has designed an incentive mechanism that allows users to register and obtain identity authentication, such as a spherical scanner (orb) to scan the iris, but this practice has caused some controversy. Now they offer other ways, such as authenticating with a passport. Once you complete the authentication, you can obtain an encrypted credential on the blockchain, which can be used in various services.

A simple application scenario is verification (CAPTCHA). Current verification codes have become so complex that even humans themselves may not be able to easily pass. Compared to these cumbersome anti-fraud systems, we can use encryption verification methods. The user can receive an encryption code that proves that he is a human, and then add an additional verification level to this. This is another interesting intersection.

There are still many opportunities for decentralized AI at the infrastructure level, such as dismantling the centralized AI system to make it decentralized at both the code and service levels. There are also some new possibilities, such as machine-to-Machine Payments. etc.

I think the most exciting part is exploring new business models in the AI ​​era, especially those for creators.

Break the economic contract of the Internet

"David George": You pointed me out "Hey, we're probably breaking the Internet contract" after the ChatGPT moment, which I think is a very interesting question.

Chris Dixon: There is a chapter in the book about this, which is close to the end. I call it the new contract. If you consider incentive systems, one of the main reasons for the success of the Internet is that it has a very smart incentive system. How do you get 5 billion people to join a system without central authority? This is because of the incentive mechanism of the Internet.

ChatGPT shows signs that the Internet economic contract may be broken. Over the past 20 years, the Internet has formed an implicit economic contract: Search engines and social platforms have access to content, and in return, creators can gain traffic. For example, travel websites, score websites, illustrations, etc. will allow Google to crawl content in exchange for search traffic. This model supports the development of the Internet. But now AI directly generates content, users don’t even have to click on links, and Google doesn’t have to distribute traffic to websites anymore. In this way, the creators' income source is cut off and the Internet's original economic model is also solved.

In the past, Google would also divide some of the traffic. For example, when users search for questions, Google would display a summary, but would still guide users to visit the website for more information. But later, Google began to "intercept", such as StackOverflow content, which directly displayed the answers in the search results instead of allowing users to access the original website. This has led to a decline in traffic on many websites and the monetization capability has been affected. Google is also doing similar things in industries such as travel and catering (such as yelp), and will even give priority to displaying its own content rather than the content of the sole creator. Although these problems have existed for a long time, the AI ​​era has made this problem even worse.

But if AI can directly generate illustrations, scores, and travel suggestions, users will no longer need to visit those content websites. This may be a better experience for users, but it is a devastating blow for content creators. In the future, we may only have AI giants left, and the original unique websites and creators will lose their living space.

This is the question we need to think about: Will the Internet in the AI ​​era still support innovation and entrepreneurship? If we don't solve this problem, the Internet could become the TV industry in the 1970s, with only one giant controlling all content. This is not the Internet future we want.

So how should a new website rise? How can new things be created? We haven't really thought about this issue clearly yet.

I don't think I have the only answer, and the solution to this problem does not necessarily have to rely on encryption technology. But we need to realize that this is destroying the original incentive mechanism of the Internet. Secondly, we need to think: Is this a good thing? I don't think so. We need to find the right solution – should we create new incentives?

This is why I have been focusing on investing and thinking about new incentive systems, such as Story Protocol. We need to explore new ways to superimpose new economic structures on existing systems to ensure that the Internet can continue to innovate and develop.

From mobile Internet, social networking and cloud computing to

encryption, AI and hardware

"David George": One thing you talk about is three technology products that appear simultaneously - Generative AI, cryptocurrencies and new hardware platforms. How do you view the combination of these three?

Chris Dixon: The analogy is of course mobile, social and cloud computing. In the last wave, they promoted each other and jointly promoted the development of the Internet. We have seen some such combinations today.

Now, we are in another wave of technology, with core technologies this time being AI, encryption and new hardware such as robots, autonomous vehicles and VR. These technologies are not unique to each other, but complement each other and form a new ecosystem together. New hardware devices, such as AR and VR glasses, rely on AI to provide a better interactive experience, such as smart assistants like the movie "She". Autonomous driving car, Tesla's robotics, and various humanoid robot projects are also deploying AI technology in the physical environment to apply it to the real world. Encryption technology provides a new way to enable decentralized networks to support these AI applications. So one area that I am interested in is DPIN - decentralized physical infrastructure. The most prominent example is Helium, a community-owned, crowdsourcing telecommunications network project that is competing with traditional operators like Verizon and AT&T. Helium has designed a set of incentive mechanisms where anyone can build a node at home to support the network. These nodes function similar to wireless signal transmitters, and currently, tens of thousands of people have installed these nodes across the United States.

Now, Helium also launches a web service, and it’s much cheaper than Verizon – only $20 per week, while Verizon costs $70. This is mainly because Helium's network is built by the community and does not require tens of billions of dollars to build infrastructure like traditional telecom companies.

How to use encryption technology to start network effects

Chris Dixon: Encryption technology has great advantages in solving the "cold start" problem.

Many network effect projects faced a challenge in the early stages: how to attract enough users to make the network work?

For example, Helium is jointly built and operated by the community. But assuming there are only 10 nodes, it obviously won't work. The construction of network effect is a problem of chickens. If a new social network has only 10 people, it is not very attractive to new users. But if it already has 1 million users, the value of new users joining will increase significantly.

What is unique about crypto technology is that it can inspire early users through the token economy, thereby promoting the formation of network effects. Helium is just one example. Other fields such as weather data, autonomous driving data, electric vehicle charging stations, decentralized maps, and even scientific research can be used to build networks in similar ways.

Is AI frosting or sugar?

"David George": Marc gave me a metaphor that I like very much: Is AI "frosting" or "sugar"? If AI is just "frosting", existing industry giants will win because they can simply add an AI chatbot to their existing products and continue to dominate the market with their existing distribution channels, sales capabilities and customer relationships. But if AI is "sugar", that is, it's the core ingredient, then you can't just "add it in" it, but you need to build the entire product from scratch. In this way, the AI ​​field is more likely to be dominated by emerging companies.

At present, we have not seen a clear answer. The more a product follows the traditional model (such as just using AI to enhance the original business), the more it will benefit industry giants, not startups.

Chris Dixon: You can look at this problem from the perspective of Clayton Christensen. He proposed the concepts of "disruptive innovation" and "sustainable innovation". Many people misunderstand the meaning of "disruptive innovation", which is just "new technology", but rather that this innovation does not conform to the business model of existing companies. That’s exactly why even the largest businesses struggle to deal with true disruptive innovation because their core customers don’t need it.

This is in line with Marc's concept of "frosting vs. sugar" - if AI is just the "frosting" of existing products, then industry giants will naturally dominate; but if AI completely changes its business model, the situation will be completely different.

For example, today's database market is basically dominated by traditional relational databases (SQL), and AI may bring about completely different computing architectures and even completely subvert the concept of databases. If AI is only used to optimize SQL databases, it is just "icing" and has no threat to existing enterprises. But if AI completely changes the way data is stored and retrieved, making traditional databases meaningless, it is "sugar", which will disrupt the entire industry.

"David George": We have not seen such cases yet. I only see the impact on price (like cheaper AI services), but that doesn't bring about industry disruption.

Chris Dixon: Yes, that's the issue at the first level. I usually use a framework to analyze the implementation process of these emerging technologies, but before talking about this, we can first talk about consumer-grade AI. At present, I don’t think there is a product that has truly network effect in the consumer AI field. While AI chatbots like Claude and ChatGPT have been successful, they have not formed a powerful network effect. Users can change AI tools at any time, with little switching costs, which makes them easily fall into price competition.

"David George": We once believed that the data network effect would become a moat for AI products.

Chris Dixon: Indeed, the data network effect is a theoretical concept, but it is often not that powerful in practice. Many people believe that the more data AI trains, the better the model will be and users will rely more on it, thus forming barriers. But the reality is that the data generated by individual users actually contributes very little to AI training. In other words, the use of data by a single user will not significantly improve the capabilities of AI, so it is difficult to form a strong network effect. This leads to AI companies facing a major risk: market competition will intensify and price wars are inevitable. Although AI products like ChatGPT currently have strong brand awareness, the question is how to avoid entering pure price competition?

If the cost of switching between different AI tools is low, the ultimate market competition is likely to turn into a "price war" where all companies are forced to lower prices to attract users. In this way, these AI companies will not be "ruling" companies.

"David George": So do startups still have opportunities?

Chris Dixon: If AI is just used to improve existing products, it will be difficult for startups to compete with big companies. But if AI is the core architecture to create a brand new business model, it will be different. At present, many AI consumer applications we see, such as face swaps, image enhancement, etc., although they exploded in a short period of time, they were quickly copied by TikTok or Instagram, and eventually the startup lost its competitive advantage. If an AI product has no network effect, it will be difficult for it to maintain competitiveness for a long time once its functions can be copied. That's why, if you want to build a truly successful AI startup, you must find an entry point that can form a network effect, rather than just providing a feature.

Come for tools, stay for the network

Chris Dixon: A classic user increase strategy is: "First comes from the tool, then stays from the network." That is to say, many users initially use a product because of a tool, but the reason for staying behind is the network effect. For example, early Photoshop users might just want an image editing tool, but later, they found that the Photoshop ecosystem was very powerful, so they became a long-term user. The rise of social networks is similar, with many users initially joining because of a feature (such as friend address book), but ultimately staying because of a social relationship chain. AI can also adopt similar strategies, such as the AI's image generation tool, which can serve as an entry point, but the ultimate formation should be a complete AI creative community, not just a tool software.

Imitation and Native Technology

Chris Dixon: Before we get into the deeper discussion, it is important to discuss how the main technologies can be launched in phases. The development of new technologies usually goes through two stages:

•Imitation stage: New technologies imitate old technologies so that they can be more acceptable to users.

•Native stage: New technologies create completely different new experiences.

There is a third phase later: more general changes brought by new technologies. For example, after the invention of the car, we built other infrastructure such as expressways, suburbs and cattle.

For example, the early web pages were like an electronic magazine, and all the content was static and there was no big difference. This imitation phase may exist in years or even years, such as Mosaic in 1993 to YouTube and Facebook around 2005.

But with the development of the Internet, we are beginning to see native Internet products, such as social media, search engines and online video platforms, which do not have offline corresponding business models.

AI is still in the sham stage, and the AI ​​applications we see are mainly replacing human resources, such as AI customer service, AI writing assistant, etc. But the real AI life will appear on AI-native products, such as the game world generated by AI, the interactive content generated by AI, etc. It's like when photography comes to life, cultural critics worry about its impact on art. Walter Benjamin's famous article "Works of Art in the Age of Mechanical Replication" asked what happens to an artist when anyone can take a picture.

Today, similar problems exist in generative AI. If AI could create the entire movie, what would happen to traditional film production?

"David George": We've seen this in the image.

AI as the foundation of creativity

Chris Dixon: Yes, this trend has started with images and the video may keep up soon. When photography first appeared in the past, people were worried that it would replace painting, but in the end, photography and painting each developed a unique artistic style. Art turns to abstraction, away from photography. On the other hand, photography technology has prompted the rise of films. People realize that while machines can replace photography, they can also create a new art form that has never existed before.

The same is true for generative AI. The negative view is that AI will replace human creation, but in fact, AI may spawn a completely new art form, providing a new canvas for human creativity, which may be a virtual world, a game or a new type of movie. In addition to the creative industry, it can be applied to other fields such as consumption and social networks.

When you create a new website, more general changes will follow. Social networking is a good example. It rose in the 2000s and reached a peak by the 2008 and 2012 Obama elections. News articles at the time also pointed out that social media has since changed from a secondary position to a major position. Then we start to see unexpected social changes. These changes are likely to unfold in the next 20 to 30 years.

Balancing supply and demand in AI

"David George": The technical stages you mentioned are very interesting. The development of the Internet took a lot of time, one of which is the need to build a huge network. This involves supply and demand issues - the development of the Internet requires the laying of wireless infrastructure such as fiber optic and cables. What AI needs is computing resources, such as large-scale GPU clusters. But the main limiting factor for AI from the "imitation stage" to the "innovation stage", may not be technical capabilities, but human creativity and ideas.

Chris Dixon: I think so, too. The bottleneck in the development of AI is likely not about technology, but about the speed of human adaptation and the impact of policies and regulations, which are closely related.

"David George": In other words, the problems in AI development include both the supply side (computing power) and the demand side (user acceptance). But the key is still the demand side?

Chris Dixon: Yes, the challenge on the supply side is to develop a powerful AI model and have enough computing power support. But the real challenge is how to get users to embrace AI and incorporate it into daily life.

We now see that many entrepreneurs are exploring how to use AI to solve practical problems. But unlike 20 years ago, the entrepreneurship ecosystem has become much more mature now. ⼗⼏Before the new year, most smart people would not choose to start a business, but instead worked in a big company. But now, the entrepreneurial ecosystem is more complete, and financing, talent and market are more mature than before.

But there is another big problem with AI, which is how people work and how the industry adapts to AI.

How AI changes the industry

"David George": For example, how quickly will Hollywood adopt AI?

Chris Dixon: That's exactly what I'm thinking about. When I was writing, I wanted to use AI to generate my own audiobooks, but both publishers and Audible explicitly prohibited the use of AI. Part of this is that the industry's unions are boycotting AI, but there are deeper reasons.

"David George": So, the ability of AI to generate content exists, but the industry is not ready to accept it. We can see that many potential applications of AI face regulatory barriers. For example, in the medical industry, the technical capabilities of AI diagnosis are already strong enough, but regulations still limit their wide application.

Chris Dixon: In the next five years, a US judge may rule whether AI training data is reasonable or Congress may introduce laws to regulate AI training data. At present, the legitimacy of AI training data remains controversial. AI companies believe that AI training data is "learning" of information, not "replication". But copyright holders believe that AI uses their content without permission, which constitutes infringement.

"David George": This is a debate in almost all AI-related industries.

Chris Dixon: Yes, it may be necessary to ultimately require laws to judgment the rationality of AI training, otherwise the issue will remain unresolved.

"David George": When will AI be truly implemented in regulated industries, such as healthcare, finance, etc.

Chris Dixon: At present, these industries are under extremely strict regulation, and it may take a lot of time for AI to enter these areas. But in some areas, such as autonomous driving, we have seen significant progress.

"David George": Waymo is an example. The data shows that it is already 7 to 10 times more safe than human driving and has millions of real data support.

Chris Dixon: Maybe this is the model of AI’s wide application—first to make breakthroughs in a specific field (such as autonomous driving) and prove that it performs better than humans, and then push it to other industries.

What is the ideal future of the Internet?

"David George": What do you think the ideal Internet should look like?

Chris Dixon: We are at a junction. The initial vision of the Internet was a decentralized network that communities could jointly own and manage, and the economic benefits of the network should also flow more to users than to a few large companies. But now, the flow of funds on the Internet has changed, and more and more returns are concentrated in the hands of a few technology giants.

"David George": Yes, social platforms' advertising revenue has reached tens of billions of dollars, but creators can only get a small share.

Chris Dixon: Currently, the top five Internet companies in the world's market value may have occupied more than 50% of the entire industry's market share. The Internet has become a closed ecosystem dominated by a few companies.

"David George": So now technology companies have mastered users and have begun to find ways to let users spend more time on their platform.

Chris Dixon: Yes, they have climbed to the top of the Internet and kicked off the "ladder" to prevent new competitors from entering. This is also why we are so concerned about the construction of blockchain and decentralized networks. If the Internet in the future is completely controlled by a few companies, the space for innovation will be greatly compressed. Relying on centralized platforms to build business is like building on quicksand, which may collapse at any time. True innovation should be built on an open ecosystem, rather than being controlled by a few companies.

"David George": So, the focus of our attention is how to make small technology companies survive and grow in this ecosystem. I'm still optimistic about the future. Through your efforts and the industry-wide drive, decentralized technologies and open source AI are being accepted by more and more people. Today’s discussion is great, thanks for your participation.

Chris Dixon: Thank you for your invitation.

*All content on the Coinspire platform is for reference only and does not constitute an offer or advice for any investment strategy. Any personal decisions made based on the content of this article are the sole responsibility of the investor. Coinspire is not responsible for any gains or losses arising from this. Investment is risky, so be cautious in making decisions!