a16z partner dialogues with Paradigm partner: This is the first time we have embedded the economic system into the internet

Author: chengpishu

Original Title: "Dialogue with Coinbase Founders and Investors: All Our Knowledge Comes from Existing Things, but Blockchain Allows You to Create New Things"

Compiled by: retric

Note: This is one of the highest quality dialogues on cryptocurrency content I have seen. The two main characters in the dialogue are Chris Dixon, an investor at A16Z, and Fred Ehrsam, a partner at Paradigm. The conversation took place at the end of 2017, when Fred Ehrsam was still at Coinbase.

Four years later, looking back, I feel increasingly that the cognitive explosion is profound. Recently, I printed out its content for long-term study and recommended it to many friends in traditional VC for reading.

This dialogue breaks down the Token of blockchain in the most basic way. From incentive mechanisms, developer communities, new protocols and governance models, the pros and cons of centralization and decentralization, to the crazy ICO market, the two big shots shared how they judge the reliability of an ICO project, how to evaluate and quantify the long-term value of Tokens, and why this new wave of technology needs to create a new layer of value.

TL; DR: (too long, didn't read)

1. For the first time in history, we have embedded an economic system into the internet.

2. Similar to computing power and storage, "native resources" are more suitable for direct sale through token mechanisms.

3. OpenSSL has been abandoned by us because it has no business model and cannot become a business, even though it is so important.

4. Bitcoin emerged during the 2008 financial crisis, but at that time, there was another bigger crisis, the "programmer crisis."

5. A good incentive mechanism can create a huge "collective behavior," and tokens can create a new network at an exponential speed we could not have imagined before.

6. All the companies that have truly grown in the past 15 years are those that have user data and network effects.

7. You can think of every token as its own new decentralized small bank or a small government.

8. Those volunteers and individual developers working at night and on weekends are the best prophets of the internet in the past 20 years.

9. In a chaotic structure, people will try more paths, and these paths may give birth to important innovations.

10. All our knowledge and experience come from things that already exist in the world and have been proven. But blockchain allows you to create new things.

11. Among the 50 largest websites in the world, 48 were established after the internet emerged.

12. The essence of software is to transform human ideas into code that machines can understand and execute.

Ten years ago, someone published a 9-page white paper on the internet. Ten years later, we have a $70 billion digital currency system with the largest computer cluster network in the world. Innovation, isn't that how it happens?

There is now a group of very unhappy programmers around the world who want a new way to create things. This situation coincides perfectly with the emergence of blockchain.

Chris Dixon: For many friends reading this article now, you may have heard of Bitcoin and Ethereum to varying degrees. The first question I care about is: Why are tokens so important? Why is blockchain not just a crazy speculative trick, but truly meaningful to our world? What is the value of tokens?

Fred Ehrsam: The most important reason, I think, is------for the first time in history, we have embedded an economic system into the internet.

Chris Dixon: But didn't we already have credit cards, Alipay, and such in the internet before? We can directly pay for a product on Taobao. What do you mean by "embedding the economic system into the internet"?

Fred Ehrsam: Decentralized blockchain and tokens represent a more direct economic system. Of course, you can now spend money on the internet to buy the products or services you want, but in the foundational design of the internet, we did not write this payment capability into the underlying protocol code, so you cannot directly pay for the products or services you want at the protocol level.

Taobao and Amazon are doing a layer of application services on top of the protocol layer, allowing you to buy things at the application layer. Ethereum and Filecoin are not like that. Ethereum directly sells computing resources at the protocol level, while Filecoin (IPFS) sells storage space directly. Besides computing resources and storage space, there are many native resources that can be directly sold using this underlying mechanism of tokens, and I believe these resources will perform better on decentralized platforms than on centralized ones.

Chris Dixon: I tend to look at blockchain and tokens more from a developer's perspective. We have always had two development models: one is that you work on the underlying protocol and develop it in a non-profit, public-spirited way. This was also the way we initially chose; the internet was originally designed as an open protocol by governments and academic institutions. And this way actually worked very well at first. We had HTTP, TCP/IP, HTML, SMTP (email protocol), all of which were great.

But if you look at the development over the next 20 years, you will find that something seems off. For example, the current problem with OpenSSL------OpenSSL provides security guarantees for internet users accessing a website. The entire world, all internet users, rely on OpenSSL to encrypt sensitive data such as bank passwords and user passwords entered online to prevent theft. However, OpenSSL later encountered a fatal bug because it was maintained only by a group of spontaneous, part-time programmers. The protocol itself has no commercial value; OpenSSL has been abandoned by us because it has no business model and cannot become a business. So very few people actually work on it, even though it is so important.

The second development model is corporate-based, where we have giant companies like Facebook doing advertising and Amazon doing e-commerce. These companies have developed very rapidly, creating a lot of revenue and profits, which are then used to support their continued growth. Meanwhile, all the smart people, the best engineers, all the capital, productivity, and everyone's energy are being siphoned off by these highly centralized systems; on the other hand, many individual developers or small startups are gradually being squeezed out of existence by these centralized systems and platforms. They suffer from the "suppression" and "pursuit" of the giants.

Either you are directly blocked by the platform, which is very common on WeChat, Twitter, and Facebook, or if you are lucky enough not to be blocked, you still have to pay a 30% toll to these platforms (Apple's App Store).

So, there is now a group of very unhappy programmers around the world who want a new way to create things. And this situation coincides perfectly with the design of blockchain, the incentive design of tokens, and the design of the Bitcoin economic system.

Fred Ehrsam: If I were an entrepreneur, would I be willing to build my own service or product on a centralized platform? In fact, as you go along, you will find that working on these centralized giant platforms can easily lead to a ceiling over your head and a guillotine around your neck. But in a decentralized system, this problem does not exist because no one can shut down your service or restrict the development of your product.

Chris Dixon: Many people say that Bitcoin emerged during the 2008 financial crisis. But actually, I believe there was another larger non-financial crisis at that time, which is the "programmer crisis." In many developer communities, there is really a "cleansing" of developers happening. Moreover, the new generation of programmers, these young technical personnel, will not suffer the same losses as the previous generation. They will not continue to be "persecuted" by these giants and large platforms like their predecessors.

Fred Ehrsam: Yes. A person does not fall into the same pit twice.

All the companies that have truly grown in the past 15 years are those that have user data and network effects. But now, every token is a newborn bank or government.

Chris Dixon: I think most platforms------when I say platforms, I mean communities or networks that can gather developers and users, such as Windows, iOS, iPhone, Android, Twitter, Facebook------all these platforms are deeply entangled in struggles among themselves. For example, Microsoft and Netscape have fought, and every time Apple holds a launch event, there is a list of companies they have taken down. There is a long history here. But in this blockchain movement, it establishes a new way to build new networks, where all participants can receive fair rewards. They do not need to kill each other.

Fred Ehrsam: That's right, this is the truly key point of the entire blockchain movement, which is "incentives." What kind of participants in the network you incentivize, how you incentivize the entire network, involves how the incentive mechanism promotes a sufficiently large group to engage in beneficial behaviors, which in turn can promote the development of the network.

The most typical example is Bitcoin. Satoshi Nakamoto wrote a 9-page white paper in 2008, proposing a new incentive mechanism, and ten years later, we have a $70 billion digital currency system with a rich ecosystem, many companies, investment institutions, and users participating, and this system has the largest computer cluster in the world, with computing power unmatched by any other country, institution, or organization.

The example of Bitcoin shows how powerful an effective incentive mechanism can be. Moreover, if you have good engineering implementation capabilities, you can create a huge "collective behavior" through such an incentive mechanism, a collective behavior that was previously unimaginable and could not be initiated using traditional methods.

Chris Dixon: Incentives here are not just about giving these developers some returns, like in classical entrepreneurship, giving entrepreneurs returns and equity financing. The incentive of tokens also extends to roles like miners. Miners are service providers in the network, and incentives are given not only to developers but also to these service providers and users in the network. The incentive mechanism of tokens now seems to be able to create a new network at an exponential speed we could not have imagined before. Ethereum is an example.

Fred Ehrsam: Exactly. That's right. The reason I first joined the blockchain industry, which I believe is also the most important reason, is that you incentivize all early potential users and service providers willing to join this network with tokens, even potential "protocols on protocols."

If you look back at all the companies that have truly grown in the past 15 years, most of them are companies with user data and network effects. "User data + network effects" is monopolistic; it leads to the solidification of power. All the internet we are currently in is a solid block. And the promise brought by tokens is that for the first time in history, when establishing network effects, you can overcome the "which came first, the chicken or the egg" problem.

Chris Dixon: "Which came first, the chicken or the egg," if we expand on this, the good side is that once you grow the network, you become very powerful, but the downside is that you only become very powerful once you grow large.

A dating website with only one user is probably the worst website in the world, but if you have a million users, that dating website will be much better. But how do you grow your users from 1 to 1 million?

From my experience in investing and entrepreneurship, 99% of networks die before they reach 1 million users, while companies like eBay always seem to have some particularly strange methods to achieve growth, ultimately surviving and becoming larger. There are probably about 50 such large network platforms that have truly survived and eventually grown into big trees, but there are countless thousands of networks that may also be quite useful and could help the world become better, but these networks have died out before they could gather many users.

The biggest problem is actually the expansion problem, right? Tokens are actually a universal solution to the expansion problem. When a new network platform does not have enough network value, you use some financial value to incentivize early seed users to enter your network earlier, and then this financial value will gradually narrow as your network value grows.

Fred Ehrsam: This is a bit like joining a startup like Tencent very early on. You take on a higher risk but also have a higher potential return. The only difference is that the people who can participate here are not just company employees but also users on the network or other potential service providers or application developers.

You can imagine a coordinate graph where the horizontal axis is the number of users and the vertical axis is network value. If you are the first person to use this network, the network value is certainly zero, but as more and more people join the network, the network value will suddenly rise sharply. It is very likely that this network will ultimately fail, but once it succeeds, you will reap tremendous value, and later entrants will be more willing to enter this network, accelerating its growth.

Chris Dixon: Moreover, interestingly, not only can these early developers or users enjoy the dividends of the network, but they are also the early regulators and maintainers of this network. Sometimes I feel quite helpless when I see some articles. Because people keep saying that Facebook manipulating elections is terrible, while on the other hand, they say that blockchain is the worst thing happening in Silicon Valley. They do not even realize that the entire decentralized movement and the significance of cryptocurrencies we are discussing now lie in leveling the inequalities within networks like Facebook. This is the core point of the blockchain wave.

Fred Ehrsam: This is true in economics, politics, and many other areas. I think there is an interesting perspective on blockchain------we rarely have the opportunity in the real world to try different methods in the economic and political fields to explore the best organizational structures and achieve better outcomes. New regimes or new banking institutions rarely emerge in the world every day, right? This is generally a low-probability event. But you can think of every token as its own new decentralized small bank or a small government. There are probably thousands or tens of thousands of tokens globally, so essentially, we have created a new platform where you can try various commercialized management systems or economic systems. This is a new path of innovation.

Chris Dixon: Is Tezos such an attempt?

Fred Ehrsam: Yes. Many people may not yet understand Tezos. Tezos is a new blockchain platform very similar to Ethereum, where you can run smart contracts, but the only difference is that Tezos maintains a self-amending ledger. In other words, the future direction of Tezos's development, how its protocol is designed and formulated, is unified and advanced under community consensus. What Tezos becomes is entirely decided by all members in the community. Users of Tezos and members holding tokens can vote on how to modify the Tezos protocol. It essentially provides a mechanism for self-evolution that can adapt to different developments over time. Developers on Tezos will contribute to the Tezos protocol in the future and will gain significant rewards for doing so.

Chris Dixon: Speaking of this, it is actually a pity for Ethereum. Ethereum still has many issues that need to be resolved, such as scalability, but many of its core open developers also hold Ethereum tokens, yet many of these developers are still engaged in other different projects.

Fred Ehrsam: The current market value of all Ethereum tokens is $25 billion. If you can successfully scale it, such as implementing sharding, the overall market value of the Ethereum network can easily increase by 10-20%. That’s nearly a $2.5 billion increase. This is a very significant change.

But the question is, how do you convert the value brought by this change into incentives that truly motivate the core developers of the community to get this done?

If you can put out half of the future value increase, which is $1.25 billion, to incentivize developers to scale and shard, then everyone benefits.

The interesting thing about Tezos is that it uses token inflation to complete this incentive. If you submit a new piece of code to Tezos's code pool, this piece of code will be tied to a bill. If Tezos ultimately accepts your submitted code, it proves that your contribution is beneficial, then Tezos will pay you more tokens. And this portion of newly generated tokens through inflation is ultimately valuable because the overall network has increased in value after accepting the new code changes.

This is somewhat like in a startup where a new employee joins, and then every old employee shares a bit of their equity with this new employee. The only difference is that after this operation, Tezos's entire pie becomes larger, so the equity each employee holds ultimately appreciates. Everyone benefits.

Those volunteers and individual developers working at night and on weekends are the best prophets of the past 20 years. So if I had to bet, I would definitely bet on this group of people.

Chris Dixon: Next, let's talk about the proof-of-stake mechanism. Ethereum is transitioning from PoW to PoS, and it seems that everyone suddenly realizes that PoS has new design possibilities. For example, under the PoW proof-of-work mechanism, you cannot punish bad behavior in the community, but now PoS can provide this capability.

Fred Ehrsam: Exactly. It provides a universal method to punish behaviors that are detrimental to community development. This is very important for the development of the entire community.

Chris Dixon: Email is actually an example. Email lacks this punishment mechanism. Because there is no punishment mechanism, the spam problem requires a lot of manpower, material resources, and funds to solve every year. Otherwise, this problem will be transferred to user experience costs, such as requiring you to fill in a verification code every time you send an email. Ultimately, the user experience becomes very poor.

Fred Ehrsam: Yes.

Chris Dixon: Besides the consensus mechanism, another important point is centralization and decentralization.

Fred Ehrsam: Yes. The advantage of centralization is that it gives you more control over the entire process, the development speed will be faster, and generally speaking, centralization will also excel in user experience or performance efficiency. But there is a downside to centralized mechanisms: you can only try one path.

For example, there is only one Facebook in the world, the largest social database globally. As a company, Facebook can only plan to go in one direction; they can use A/B testing methods to decide how to modify their products next, but overall, you cannot run five different versions of Facebook simultaneously. The same goes for politics, such as the establishment of the U.S. government.

But it has been proven that if you can try many different experiments simultaneously, you often get many interesting results that you could not have imagined before. Now it is difficult to do more on Twitter's API because they can block you at any time, but everyone is still using the same SMTP protocol. If you use the same standard, the network effect becomes very strong.

Chris Dixon: You just mentioned that the efficiency of centralization is higher than that of decentralization. I actually wonder if that is really the case. Is the usability, efficiency, and user experience of decentralization really worse than that of centralization? Because you need to look at this issue dynamically and extend the time frame to 10 or 20 years later; will it still be the same?

It's like there are two startups, one of which produces a terrible product on the first day, with weak efficiency, user experience, etc. But it has the power of developers. If over time, these developers continuously improve it, it may become better and better.

The benefit of SMTP is that you can construct any business model you want on email. You can build a business on email. Now there are hundreds of startups creating new services on email, email inboxes, anti-spam services, Microsoft's email service, Google's Gmail, etc. However, Gmail now seems a bit too powerful, powerful enough to have somewhat centralized the SMTP protocol, but that’s another topic.

But overall, the biggest difference with SMTP is that you can access this protocol equally. Developers and users can freely and equally use this standard. This keeps the SMTP platform vibrant. This is different from platforms like Facebook; you don’t have to worry about whether your friends are also on Facebook, right? This is a network effect issue. If that were the case, SMTP would have been dead long ago.

Fred Ehrsam: Then I want to ask you a question: if you had to bet on a centralized company to create a social product, like developing a Facebook, versus a decentralized organization building an open protocol for a social product, which one would you bet would ultimately be bigger?

Chris Dixon: Are both starting from zero? Because Facebook has already grown too large; this competition is unfair. Facebook has a lot of funding, high-end engineers, and talent.

Fred Ehrsam: Let's assume that this PK is happening when WEB 2.0 has just emerged.

Chris Dixon: This is a very interesting question. There is indeed a lot of risk.

On one hand, you have a team working in the same office, you have funding, and the team is focused on the same goal, with KPIs, clear work metrics, better computers, servers, better office facilities, etc.; on the other hand, you have a bunch of volunteers willing to participate in development during their evenings and weekends.

In fact, it is a competition between amateur volunteers and individual developers working in their spare time versus the 9-to-5 office workers, right?

The volunteers during the weekend will definitely outnumber the employees of the same company by a lot. Linux, Wikipedia, etc., these open-source projects probably have millions of volunteer developers, right? And these people are often very smart.

Those volunteers working at night and on weekends are the best prophets of the past 20 years. So if I had to bet, I would definitely bet on this group of people. This is my choice and my investment strategy. I previously wrote a blog saying that the toys smart people tinker with during the evenings and weekends will be things that ordinary office workers take 10 years to accomplish. I think this is also the biggest inspiration the internet has given us in the past 20 years.

Fred Ehrsam: Yes. You can actually see the "motivation" of these two types of people working.

Chris Dixon: It is motivation, but it is also… how should I put it, time is the best judge. In a company, no matter how ambitious you are as a leader or how far-sighted you are, essentially, you are still "short-sighted." Zuckerberg, Larry Page, they are all great entrepreneurs with long-term vision and a big enough perspective, but in the end, they can only make short-term plans of 2-5 years, right? Those who work at night or on weekends, those in the lab, they generally consider issues with a time span of over 10 years.

Fred Ehrsam: Exactly. To innovate, it often involves issues of brand communication or whether it can bring cash returns to the company in the short term, which large companies generally cannot avoid, so they choose not to do it.

In this paradigm shift of decentralization, the interesting thing is that compared to traditional companies that can only try one path, you can try many things simultaneously. This means you can copy all of Facebook's source code and all of its user databases, and then simultaneously try five different versions of Facebook. Some of these versions may fail, while others may run very well, and then you pick the best-performing version and continue to split it into another five versions. This is somewhat like natural evolution.

The decentralized token mechanism may seem chaotic, just like free markets are always more chaotic than planned markets. But the key point is that you bet that in this chaotic structure, people will try more paths, and these paths may give birth to important innovations. Moreover, in such a system, you often learn faster because the feedback cycles brought by this mechanism are very short and direct. Which paths are worth trying can be seen quickly.

Chris Dixon: The reason I felt more confident about the entire industry last year was like this. Compared to the winter phase of 2013 and 2014, of course, many people will say it is because the price of Bitcoin has risen. In fact, it has nothing to do with the price of coins. The most important reason is what you just mentioned, that the proportion of this "attempt" path is increasing.

Fred Ehrsam: I believe that as time progresses, we will become better at development, whether in centralized companies or in more chaotic and loose decentralized organizational structures. For both, development capabilities will improve. What you just mentioned actually focuses on the most critical aspect of the entire industry, which is the activities of developers------what they are doing, what they are trying.

In 2014, we at Coinbase would pay attention to how many people were following Bitcoin's GitHub repository, how many developers referenced Bitcoin's source code, and then find some reliable entrepreneurs or entrepreneurial teams from these people. Now, we look at how many people are following Ethereum on GitHub, how many people are referencing Ethereum. New things have actually emerged. So suddenly, the entire industry has become lively.

Chris Dixon: Ethereum has brought us a lot of inspiration: First, you can gather a large network in a very short time, even though there were already super apps and super networks; Second, Ethereum is Turing complete, with Solidity built in as a development language like JavaScript, allowing you to develop various applications on it. This is the potential that Bitcoin initially showed but has never realized; Third, you can not only create various applications on it but also create new token networks and establish a new layer of protocols. So, Ethereum may have these three aspects of innovation happening simultaneously.

Fred Ehrsam: Yes. This is also why I believe blockchain is a very special industry. It is so close to finance, but it is not just finance. If you are a person from the financial industry, coming from Wall Street, your understanding of blockchain in financial and economic aspects will gradually enhance your understanding of blockchain in applications, innovations, and systems. In other words, you must first understand its significance as money, but that is only one side. If you only understand blockchain from this level, you will simply regard it as------"money," and you will not see its intrinsic value, why this thing could become a global innovation competition platform, you will not know.

Chris Dixon: I think what they overlook is that this is a complete re-architecture and adjustment of the internet, which will allow everyone involved in the internet to have the opportunity to own a part of the network, and then the network will ultimately have value, and that value will be realized into benefits and returns. But to obtain these benefits and returns, you actually need to first understand what we are creating. To understand what we are creating requires you to comprehend how the internet operates from both technical and cultural perspectives.

All our knowledge and experience come from things that already exist in the world and have been proven. But blockchain allows you to create new things.

Chris Dixon: You create a network and then use the tokens in the network to participate in the development of the network. This involves another concept called ICO. Although I don't particularly like the term ICO, but since everyone calls it that now, let's talk about what ICO is all about.

Fred Ehrsam: Well… it seems that overnight, everyone rushed out to issue coins, thinking there was money falling from the sky to pick up. Issuing a coin could make a fortune. This is actually quite similar to startups. A very small number of startups can truly raise funds and create great products, but most startups are garbage; they may never be able to raise money in their lifetime.

Chris Dixon: Actually, it’s the scammers who entered the scene.

Fred Ehrsam: Yes, most ICOs now are bad, but there are a small number of ICOs that look good. It really comes down to whether a coin issuance project is reliable. Unreliable projects may have several signs: for example, tokens based on rent-seeking models; or a token is used to eliminate middlemen, which is unnecessary; or tokens are used to raise funds for a centralized system, which is also unnecessary.

Chris Dixon: If you derive a token from a company's profit distribution, the credit and security of such a company are very high. In other words, you actually do not need to use tokens; you can completely solve the problem through traditional means.

Fred Ehrsam: Yes. Other signs of unreliability include that the white paper you released is just a marketing manual, completely unrelated to the technical specifications of the protocol layer.

Chris Dixon: So conversely, a reliable project must at least include a white paper with technical details, some running code, and a project team that can truly write high-quality code, with the team itself having a good software engineering background.

Fred Ehrsam: Exactly.

Chris Dixon: I must emphasize that many tokens we mention here do not recommend anyone to invest in. Because the risks involved are very high. If you really want to invest, you must understand a lot of information behind the tokens, conduct thorough research, or be mentally prepared to lose all your money. Because it is venture capital.

Fred Ehrsam: Yes. I think the industry is still in a very early stage. We are still laying the foundation and have not yet been able to truly start building houses. If you look at Ethereum's current processing capacity, it cannot handle more than 20-50 transactions per second, which means that to create a real Facebook on-chain, there is still a distance of about 2000 times. So, I think providing infrastructure services now is more valuable than directly creating applications. I believe that the success probability of infrastructure and system projects is much higher than that of applications.

Chris Dixon: If we compare it to the internet, we are still building servers and protocol layer infrastructure. This is a process that must be experienced. It took 20 years for the internet to give birth to a Facebook; you must have AWS, you must have other intermediate layer services to have a super application come out.

Fred Ehrsam: In this technological wave transition, one common mistake we often make is that we try to directly transplant existing solutions and ideas from the world into the new platform. Because all our knowledge and experience come from things that already exist in the world and have been proven. But the emergence of new technological waves is interesting because it allows you to do new things, not old things. So I personally think the most important thing is