a16z Crypto Founder: Why is the emergence of blockchain so important in the history of internet development?

Video Title: web3 and how we got here.
Video Author: a16z crypto founder and general partner Chris Dixon
Translation: Qianwen, ChainCatcher

In this video, a16z crypto founder and general partner Chris Dixon shares the history of the internet, the similarities and differences between protocol networks, enterprise networks, and blockchains, as well as the potential that truly open-source, decentralized networks can unleash.

Three Stages of Internet Development

There are many theories about the development of the internet, but I personally believe it can be divided into three core eras, and we are entering the third era. Each stage of development has a key network architecture. In my view, networks are the killer applications of the internet, and these networks will have various downstream impacts on the economy, user experience, and other aspects. Each stage of development has its own dominant network types and typical computing primitives. In this context, one idea is to have everyone in the world put information into a large repository for sharing, so that various pieces of information do not get mixed up. To achieve this, some simplified principles need to be followed, which leads to the three stages of the internet.

The first stage is reading. The network is like a large multiplayer social game, and websites are a fundamental invention. The underlying idea of this design is that you can put any information, such as code data, into something called a website, which can link to other websites. This is a very important encapsulation mechanism and a simplification mechanism, allowing all of humanity to have information that can link and jump to each other, making this content a digestible unit. We achieved the democratization of information access in the early days of the internet, with Google and Wikipedia being the killer applications of that era. Through the network, you could find the content you wanted, which was a magical new experience. Some say that the average user today has more information than the President of the United States did 30 years ago. However, at that time, it was still a relatively primitive (replicated) version. The network felt like you were putting magazines on a website; you could submit forms and do some one-way interactions, but it wasn't a true two-way interactive experience.

The second stage is writing. The key is how to enable true two-way interaction between the internet and users, democratizing publishing. This gave rise to blogs, where people could publish information in a simple way. Many ideas from figures like Mark Zuckerberg came from the early blog community's sharing. Blogs contained various content, such as photos and links, which extended into a whole set of interesting products.

The third stage is protocols. This was a very great design decision in the early days. People wanted to create a global multiplayer social network game, but how should it be designed? At that time, there were no tools like Google or AWS data centers, so you had to build it in a peer-to-peer manner, leading to the birth of protocols, which are a series of standards, such as HTTP for the web and SMTP for email. Protocols are like a language, like English, telling you the standards that need to be followed for certain operations. If someone establishes, opens, sends emails, or uses all other tools, then this protocol will be built into the software. When you download that software, you join email, you join the network.

Importantly, there is no central intermediary, and no one controls everything. This is a community-driven situation where people can freely discuss things. Most importantly, it doesn't cost you a dime. Of course, you need to buy a domain name, but after that, you can build a website, and these websites belong to you. Anyone with a good idea, within legal bounds, can establish a website and own the value it brings.

One very important invention is DNS. DNS is the naming system for the domain name system. For example, I own Cdixon.org, which maps to an IP address. The important thing is that I control this mapping—that is, users control the mapping, transferring all power of the network to the users. This means I can build my website, I can change my hosting provider if I don't like their high fees or low speeds, I have the right to switch, but I still retain my network, search results, and email (some might say later email was monopolized by Gmail, but at least that happened later). In short, if you don't like it, you can leave, which forces providers to follow our thinking. However, when we enter the era of enterprise networks, while you can leave, are you really leaving? For example, Facebook allows you to download a CSV file and take it away, but you can't take your followers with you. The current situation is that the network is monopolized. If you ask a blogger on the street, most people are not interested in building their own websites, which is contrary to the original intention of the internet.

Network Stagnation

The network has become stagnant. Ninety-five percent of traffic flows into the top one percent of websites, and according to MAU, eighty-five percent of mobile device usage goes to the top one percent of publishers. The economics are similar. The Nasdaq 100 index is primarily dominated by the five major tech companies. We are increasingly moving towards a monopolized world, which is far from the original vision.

For me, the internet is a tool for building networks, such as the TikTok network, Instagram network, Uber network, etc. As I described earlier, concepts like DNS and the way networks are built will ultimately have very important downstream impacts on the economy, governance, and who can build networks. For example, many of my friends were developing Twitter and Facebook applications around 2010, and all these companies were wiped out on the platform because these big companies changed the API, changed the rules, and unilaterally controlled everything. So the architecture of the network determines everything.

Here, I want to talk about the take rate. For example, on YouTube, they take a cut, but if you replace social networks with a different protocol architecture (like RSS), then a large part of these fees would flow to creators, users, and software developers, rather than intermediaries. This might be a good thing, but it could also be a bad thing. What I want to express is that architecture determines everything. For example, if you build an enterprise network, you will ultimately face economic and social issues, which are the economic and management models of today.

Blockchain is a new type of virtual computer built on a computer network, and its main feature is that it disrupts traditional computing architecture. If Google's servers were placed on the network, they would ultimately control the network because they control the hardware, and they can modify the software; they have been doing this. Imagine if Google announced the launch of Google Coin and promised to issue only a certain number of Google Coins; no one would believe it because they would backtrack at some point, and there would be changes in management, such as the founder leaving; in short, things would change.

What blockchain does is say: we give power to the software. For example, if you write a smart contract on Ethereum, you can control everything. The most important thing is: you can build networks on the blockchain, gaining the social benefits of protocol networks while also obtaining the competitive advantages of enterprise networks.

The benefit of protocol networks is that there are no cuts; all funds flow to the edge of control. This allows for various interesting innovations to occur. However, they have a significant weakness: funding is not easy, primarily obtained through government funding and other means.

In fact, the RSS protocol once rose, but later lost to YouTube, partly due to their different strategies. YouTube initially had no traffic, and to attract traffic, they adopted a free hosting strategy. For example, I have traffic on my own website, but YouTube can host my videos for free, so I uploaded my videos and let them host it, then they embedded a widget on my website. That's how YouTube started. However, RSS had a different strategy; you could buy a domain name, but you had to pay for hosting. Hosting videos and other costs are very expensive, so YouTube subsidized this, which is an architectural advantage of enterprise networks and the reason why RSS lost to software like Facebook and Twitter. Initially, companies like YouTube and Twitter supported RSS functionality and API usage, but later they stopped these services.

So I think this is a cycle; at first, traffic is minimal, and companies try everything to get creators on board, offering various subsidies. But ultimately, as companies scale, they create huge network effects, and the incentive mechanisms change, with profitability becoming more important. So this is just a core logic. For these enterprise networks, the larger they become, the less motivation they have to interoperate with other networks, support other networks, and allow creators to enter. I call this the "extract circle." Initially, you try to attract all different people to join your network, but at some point, you start to extract from them. Of course, there are many efforts to try to alleviate this situation, such as Google promising they won't do this, but the reality is that after the founders leave, the company's strategy will shift to profit-first because if you don't do this, other competitors will. Thus, we find ourselves in a world that is almost monopolized, and innovation is suppressed.

Blockchain Should Be an Organically Growing City

Blockchain still has a core, which is the blockchain network. Taking Uniswap as an example, the smart contract code is the core that runs, and the core idea includes code immutability. Uniswap is a great example; even the community cannot change the code. If they release a new version, it is a new set of code, immutable code. This is the "do no evil" principle. They cannot change the rules because it is stipulated in the code, so as a developer or user, you can decide for yourself because there is competition and choice.

The core of blockchain is much smaller than that of enterprise networks. It is more like building a city rather than a theme park. Enterprise networks are like a theme park because every feature of Facebook and Twitter is built by their own employees. No matter what kind of network you choose, it is created by the company's employees, essentially stifling all innovation.

In the late 90s, spam was rampant. How was this problem ultimately solved? Although it hasn't been completely resolved, the situation has improved at least. Ultimately, the market intervened; at that time, there were about 50 venture-backed companies solving spam issues, and eventually, Gmail acquired the startup Postini. Similarly, you can leverage the power of the market to create blockchain. If you want to build a blockchain network here, you would want to establish the core of the network. Just like building a city, you need a mayor, a city hall, a fire department, and police, all of which involve basic rules and governance.

You can build such a network to create incentive mechanisms for people, leaving ample design space for mechanism design, while also accumulating funds to provide funding for developers and achieve all the things that RSS couldn't do. This is what I call "the best of both worlds," where you will get this open ecosystem, these cities, these digital cities. On this basis, you can build future applications and networks, as well as all the wonderful things around protocol networks, while also obtaining funding and functionality. Moreover, from a technical perspective, some new naming systems like ENS can also solve the domain name issues I mentioned earlier, modernizing the user experience.

Ideally, in today's era, if you cannot achieve functionality comparable to enterprise networks, you are likely to fail. User demands are high; they expect a good user experience. So essentially, this architecture is logically centralized but organizationally decentralized. For example, ENS has a central database, but it is not controlled by individuals; it is owned by the community.

Federated Networks

Federated networks are still protocol networks; their logic acknowledges the problems that come with individual control, so the solution is to break this massive dictatorship into many smaller dictatorships. Each server is a small mini-dictatorship, and as long as the federated sub-networks are small enough, people can have autonomy, and they can check each other, which is a way to keep power decentralized.

In the 2000s, all these services like Twitter and Tumblr promoted themselves as RSS servers, but the final situation was that one or a few servers either didn't work or the servers did work but clustered around a single server. Especially in a federated architecture, there is no typical nation, so there is no central place to hold the namespace and everything else. Thus, these servers intersect and cannot operate "gracefully," ultimately requiring a sharding system. If the entire system can operate in the end, then some servers will become powerful, and we will experience the history of monopolization again. I believe you need a typical central core to have a system that matches the future and modern enterprise networks.

What I want to express is that cities are better than theme parks. Importantly, the interaction between private ownership and public ownership in a city is very important. For example, if I open a small restaurant, I need to rely on the traffic on the street and public facilities; this is an interaction. If only a central authority makes decisions, they are unlikely to understand the specific needs of people. Therefore, in a well-designed city, there will be diversity among public facilities, parks, and infrastructure, giving a sense of an open, shared city while also promoting entrepreneurship. But at the same time, you can leverage the creativity of the free market; the network is a typical example, reflected in a network where you can achieve truly interesting things, like Google and Wikipedia. There are all kinds of things in a great city, where various organic beautiful things interact, which is far superior to a carefully planned and managed place.

Why is Blockchain Important?

Now let's discuss why people would want to see a world where blockchain networks are widespread. Because this new network architecture combines the essence of previous architectures. First is community ownership, which is similar to protocol networks. Here, you have community ownership; the concept of read-write ownership is a form of ownership where you can truly own a part of the network.

Tokens are one of the key concepts of this era, encapsulating ownership. I believe tokens can be anything; they can be things in the physical world, digital world, code, games, art, or anything. The impression that tokens leave on the public often resembles currency characteristics, which is a very narrow view. To me, tokens are like a website where you can put anything inside. In a well-designed blockchain network, you can put voting power, the ability to own part of the network, and other things in to share economic benefits.

Lower take rates. Money ultimately flows somewhere; do you want it to flow to network operators or to network participants? This is not just an altruistic principle; it is also a competitive advantage. You can seek out creators, musicians, and all these people, offering them lower take rates, which is one of the powerful factors that will drive this movement to success. In fact, for all participants, this is a better economic proposition.

Incentivizing innovation. Everything you build belongs to you. The rules are written into the code, and others cannot change them, while also having the fundraising advantages of enterprise networks. I like to illustrate this with an example of an SSL vulnerability known as "Heartbleed" a few years ago (SSL can be understood as a standard technology used to protect internet connections). This was a very dangerous vulnerability that could threaten all encryption technologies on the internet. Half of the developers were studying SSL, but there was almost no funding, and this issue was only partially resolved at that time.

Some might say, look at Linux; it is well-funded. But if you look closely, the biggest contributor to Linux is Intel, and Intel has ample strategic reasons to support Linux because their main competitor is Windows, so they must analyze all political alliances. However, networks do not naturally have their own allies, so networks need to have their own native form of funding to have a chance of success, and that is what blockchain can provide.

Tokens

Tokens are a product of traditional video games, which pioneered virtual economies. For example, "League of Legends" and "Fortnite" are mainstream video games. In the past, people would buy physical games, but now almost all modern game companies do not charge based on the game itself but make money through virtual economies.

In blockchain network design, if you want to design a virtual economy, you can think of tokens as flowing water, with places where tokens flow out (drains) and places where tokens flow in (faucets). For example, staking and airdrops are a form of faucet. Access fees and transaction fees are the drains. You want to achieve a balance between the two, and there needs to be a balance between all supply and demand, but you also want to incentivize good behavior through the faucet, hoping that blockchain can make up for the shortcomings of mechanisms like RSS.

Incentive mechanisms are very important here; you can use them to provide funding for developers. For example, the YouTube hosting example I mentioned can be understood as a faucet. Ideally, the drain should align with the utility of the downstream applications. For instance, Ethereum's drain is great; Ethereum is a globally shared supercomputer in the cloud, owned by no one, and anyone can access it. It is like an old-fashioned mainframe computer, its resources are limited, and it charges you for using the computer, billing you in Ether. The convenience of this is that as people build more applications for Ether, the demand for Ether should rise, and Ethereum will become a more popular network.

So, it is closely related to the practicality of the economy. If someone creates a popular game based on Ethereum, it will increase the demand for Ether; this is economics. This is the same as what economists have been discussing about internet economics for hundreds of years. This is also economics concerning supply and demand. If you build a popular network and link the network token to the utility of the network, you will generate basic sustainable demand rather than speculative demand.

Today's token prices contain speculative behavior, just like speculative behavior exists in every tech market. However, speculation is unsustainable because the focus is on creating fundamental value, building new networks that have the advantages of protocol networks, and can better leverage the advantages of enterprise networks, creating new virtual economies. Using tokens and incentive mechanisms to achieve adoption, promote behavioral adjustments, and realize all these wonderful outcomes I mentioned, such as returns.

The above image shows the take rates. I speculate that as market competition emerges, there will be downward pressure, meaning that over time, many of these take rates will decline. For example, OpenSea does not own NFTs; they cannot force you to stay there. Networks must have the right to compete; competition lowers prices, take rates, or network prices. So I believe this interest rate may decline in the future.

Composability

Composability is the truly exciting part of software. Core created by network designers, then all these things are built around the core by the ecosystem. This is a very exciting thing. If you trace back the history of technology, you will find that the history of composability has a long-standing tradition, specifically focusing more on networks and open-source software. I remember in the 90s, Linux was very niche, and open-source was a particularly radical idea that came from Richard Stallman in the 80s, who believed that no one should own software and opposed intellectual property.

In the 90s, there was a famous article called "The Cathedral and the Bazaar," which presented two concepts: one is the cathedral, like Microsoft, which was the Google of that era, dominating the market, while the bazaar was a bustling crowd where everyone was developing open-source software. Now, the vast majority of operating systems in the world are Linux. Your Android phone is Linux. All data centers and all IoT devices are Linux, so this system has completely won. It turns out that letting a group of people package on the internet is a better way than doing software in a "cathedral"; you could say the bazaar is a better way, as evidenced by networks and Wikipedia.

Because of the composability of the network, a global knowledge base is being established, empowering everyone. Personally, the most frustrating part of enterprise network architecture is that all innovation is confined to the "clergy" (like product managers), and innovation is largely stifled.

Our Current Development Stage

As shown in the image, our team has made a rough classification of applications and infrastructure. Our theory is that when a situation occurs, such as a price increase, people around will also get excited, for example, saying, "I want to read the Ethereum white paper, I want to learn about it," and then they really become fascinated with it. Some people are speculators, but some genuinely develop an interest, so they start companies, contribute to open source, and so on. Typically, seeds take time to fully bloom, and then you will see a new round of exciting things. So we did the data analysis project shown in the next image.

As shown, it is not a very stable development like the development of internet software; it is somewhat chaotic. This reminds me of artificial intelligence. AI has also gone through very chaotic development. Since Turing published his paper 70 years ago, AI has experienced "summers," "winters," and so on. Therefore, the larger the technology, the more likely it is to experience such chaotic cycles multiple times, as this involves market sentiment, funding, and other factors. If you know the S-curve of technology adoption, we are still at the bottom.

Golden Age

I am a devoted fan of history; I have read every book on computing history, technology, and history. I try to read these materials because I am a fan of stories. I used to be an entrepreneur, and I love those stories in the room, just thinking about three people sitting together, drinking beer, and discussing an idea is truly incredible. I have been an early investor in many companies that became big, and in this world, being able to create software and be part of these early movements is truly amazing.

When you read history, you might think that doing personal computers in the 80s was certainly obvious, just like "history doesn't repeat itself, but it rhymes." There will always be a set of different things emerging, such as cryptography, which of course faces a lot of criticism. Bill Gates' vision at that time was also revolutionary, aiming for every household to have a computer, and everyone would own a part of the network, which was a great vision. In fact, the entire history of computing in the 1930s is very interesting, such as the stories of logicians. For example, the story of Fairchild Semiconductor and the stories of Robert Noyce and others. I encourage you to read them.

The story of building personal computers at that time is somewhat similar to our current hackathons, where someone says, "Let's make PCs; that would be cool." The price of an Apple computer was five thousand dollars, but there was very little you could do with it, so it was better to build personal computers. At that time, no one realized it would change the world.

I feel excited at this moment; why? Actually, you should also feel excited because these are very special moments when you can witness a new computing movement, AI, VR, and a series of technological changes.

This transformation really affects the future of the internet, and we must make a decision about what kind of internet we want. Is it an internet monopolized by a few companies? Or do we want an open, decentralized, innovative internet? Currently, I don't seem to see good regulations emerging that can achieve an open internet. Perhaps the form of protocol networks will make a comeback after 30 years. For me, I think this is the most credible way so far to bring the internet back to its roots.

Of course, this is also why you are here. I believe this is a great entrepreneurial opportunity where you can truly make a difference.