Vitalik Chinese AMA: Ethereum as the World Computer

Author: Jiawei @IOSG

Recently, Ethereum founder Vitalik Buterin and FSL CRO Mable Jiang conducted a flash interview AMA in Tako App. This is also the first Chinese AMA conducted by Vitalik in recent years. It aims to let more people in the community understand the current status of Ethereum from the perspectives of technology, ecology, and future direction. We have sorted out the core views given by Vitalik on some specific issues that everyone is concerned about, hoping to give readers some inspiration.

Summary

1. Vitalik's definition of "world computer" is: "a place where the world's applications can interoperate with each other", rather than "a computer that can support every application in the world at the same time". Ethereum is both a digital asset with long-term value storage capabilities (such as Bitcoin) and an underlying platform that supports global application interoperability. The two are not contradictory. The "world computer" is realized through the collaboration of the highly decentralized and trusted L1 (underlying chain) and the L2 (scaling layer) with scalability and special features. It is not feasible to rely entirely on L1 or L2.

   The core mission of L1 is to ensure security and network resilience-even if the foundation no longer exists, global developers can still maintain the ecosystem based on open clients and protocols. L2 carries the mission of performance expansion and scenario innovation, such as high throughput, privacy or AI model collaboration.

   The relationship between the two is like the foundation and architecture of a city: L1 ensures geological stability, and L2 allows free design of the form. However, the premise of this architecture is that L1 must maintain sufficient throughput so that any user or application can interact directly with the main network when necessary, while ensuring interoperability to prevent L2 from becoming a closed island.

   In summary, Ethereum does not need to make an either-or choice between "digital gold" and "world computer" in its finale. The compatibility of the two is precisely its unique value: Ethereum needs to be a decentralized, censorship-resistant, long-term and reliable value storage carrier like Bitcoin, and it must also serve as the underlying infrastructure to support the interoperability of global applications.

   Compared with the proposal of "L1 as the base layer to fully support L2" a few years ago, Vitalik proposed "hybrid L1 + L2" in this AMA, that is, L1 also needs to play a certain role, increase blobs to give L2 more space, and promote cross-L2 interoperability. According to actual needs, the market will figure out which transactions should be in L1 and which transactions should be in L2.

2. In the past three years, the development of the Rollup route has shown a certain polarization. On the one hand, L2 represented by OP Stack achieves transaction confirmation and MEV control in seconds through centralized sequencer; on the other hand, the architecture that relies too much on the operation of a single team has also raised questions about censorship resistance and long-term credibility.

   Vitalik believes that at the technical level, censorship resistance needs to be guaranteed by forcing transactions to be on-chain. At the ecological level, Vitalik encourages the exploration of "dedicated L2". For example, Aztec achieves privacy transactions through fully homomorphic encryption, Gensyn focuses on distributed AI training, and MegaETH uses centralized sequencer to achieve a high-performance scenario of 100,000 TPS - these experiments are essentially answering "what degree of decentralization is suitable for what application".

3. Vitalik has contacted and talked face-to-face with many community members, and believes that in the future, the Ethereum Foundation needs to focus on supporting public goods, formulating technical standards, and maintaining ecological balance, rather than pursuing corporate governance. Corporate governance means the pursuit of profit. Ethereum is a decentralized ecosystem, not a company. If Ethereum becomes a company, it will lose most of its long-term significance. Vitalik does not approve of the culture of short-term speculation (such as the over-hype of memecoin) and excessive catering to political needs. He advocates creating long-term value and attracting long-term developers. Of course, Vitalik is not completely against memecoin. As early as March last year, he proposed the practical uses of memecoin in the article "What else could memecoins be?" In Vitalik's eyes, a good application should meet the real needs of users, have the ability to achieve sustainable profits, and have social value. The problem with many projects is that they fabricate non-existent or difficult-to-verify needs, and do not need the anti-censorship, verifiable, and collaborative characteristics of blockchain. Some successful products often start with the "minimum necessary scenario." A healthy profit model must be rooted in real value exchange, such as Uniswap fees on the financial side, and annual fees and rare domain name auctions on the non-financial side of ENS domain names.

Core viewpoints on some specific issues

I think these two ways of thinking are compatible with each other.

If you need to distinguish which blockchains are "truly decentralized", you can use a relatively simple test: If his foundation disappears, can the chain survive? I feel that only Bitcoin and Ethereum can answer clearly: of course. Most of Ethereum's development is outside the foundation, and the client team has an independent business model. Now many researchers are not in the foundation, and almost all activities except devcon are independent.

It is difficult to reach this stage. Ethereum was not like this 5 years ago.

It is a big mistake to give up these advantages in pursuit of TPS, because there will always be new chains that suddenly have higher TPS than you. But decentralization and resilience are precious, and few blockchains have them.

These characteristics are conducive to making a digital currency with long-term value, and also conducive to having a good world computer. But the world computer also needs to solve the problem of expansion. "World computer" does not mean "a computer that can support every application in the world at the same time", but "a place where the world's applications can operate with each other". High-performance computing can be placed in L2, which is fine. But this role still requires L1 to have enough scale. The specific details can be seen in an article I wrote recently.

ETH is a digital asset suitable for use between the world's applications (including finance, and also other things, such as ENS, etc.). ETH doesn’t need every transaction to be on L1, but it needs to have enough throughput to allow anyone who wants to use L1 to use it at least occasionally.

So the two directions are also compatible here: the features that help Ethereum become a better world computer are also the features that make ETH a better digital currency.

So far, our expansion method can be roughly understood as hybrid L1 + L2, but I don’t think anyone has clearly defined which transactions should be on L1 and which transactions should be on L2.

The answer of "everything is on L2" is difficult to accept because:

• This makes it easy to lose ETH’s position as a medium of exchange, store of value, etc.

• If you are worried about L2 stealing L1 users and not giving L1 anything in return, this problem will be more serious in a situation where "L1 does almost nothing".

• Cross-L2 operations still require L1. If an L2 has a problem, users still need to have a way to move to another L2. So there are some use cases where it is difficult to avoid L1. I wrote an article on this topic here.

The answer of "everything is placed in L1" is also difficult to accept because:

• If L1 supports many transactions, it is easy to become centralized, even if technologies such as ZK-EVM are used.

• The world’s demand for on-chain transactions is infinite. No matter how high the TPS of L1 is, you can always find an application that requires 10 times more TPS (for example, artificial intelligence, micropayments, micro-prediction markets, etc.).

• L2 not only does capacity expansion, but also provides faster confirmation speeds through preconfirmations and avoids MEV problems through sequencers.

So we need hybrid L1 + L2. I think the role of L2 will continue to change, for example, it seems that evm-equivalent L2 is enough now, it is possible that we will see more privacy-focused L2 (aztec, intmax, etc.), and there may be more application-specific L2 (if an application wants to control its own MEV situation, this is beneficial, etc.)

So in the short term, I think we should continue to improve the capabilities of L1 at the same time, increase blobs to give L2 more space, and promote cross-L2 interoperability, and then the market will decide which expansion method is suitable for which application.

Centralized sequencer actually has many advantages:

• Centralized sequencer can ensure that it will not steal users' money by frontrunning, etc.

• Instant preconfirmations can easily turn a traditional application into a blockchain application, because the server directly becomes a sequencer.

• The decentralized characteristics of blockchain can be used to avoid the risks of centralized sequencers: the forced inclusion mechanism prevents the sequencer from censoring users, and the optimistic or zk proof mechanism prevents the sequencer from changing or violating the rules of the application (for example, suddenly inflate a token or nft collection).

However, centralized sequencers still have risks, so we cannot rely entirely on centralized sequencers to solve the problem. It is also important to have the ability to trade based on rollup or directly on L1. So I support having two parts of the ecosystem to promote these two methods at the same time, and then we can see which method is more suitable for which application.

It is of course critical to maintain the ability of ordinary users to issue censorship resistant transactions.

There is no such thing as ETH 3.0 now. Some people would say that Justin Drake's 5-year plan is it, but that plan is only for the consensus layer, not the execution layer, so it is only part of the future of the Ethereum blockchain.

The relationship and balance between L1 and L2 is an execution layer issue. Here is another roadmap: strengthen the capabilities of L1 (increase gaslimit, add stateless verification (such as Verkle) and other functions, etc.), improve interoperability across L2, improve blobs, etc.

I also think that the question of whether L2 pays enough transaction fees to L1 should not be viewed too much from a short-term perspective. For example: before 4844, everyone's complaints were the opposite: Is L1 sucking the blood of L2?

Right now, the blob fee for the last 30 days is 500 ETH.

If the blob target is raised from 3 to 128, according to our plan, if the blob gasprice is the same, it will burn 21333 ETH per month, 256000 per year.

So the narrative here is easy to change quickly. Now we need to strengthen L1, so that what should happen in L1 can happen in L1, increase blobs, and then maintain the adaptability of our community.

I think the blockchain community, and the whole world, is in a more dangerous state now. There are many things that have no long-term value, or even malicious, happening, and these sentiments and the people behind them have received a lot of attention.

But we can't just shout against these things and not mention a better alternative. So our goal should be to do this alternative well and demonstrate that a stable and brighter future is possible.

Here I am talking about both the blockchain circle (if memecoin, which fell 97% in one day, is not our future, then what is?), and a macro-social aspect: many people now think that democracy is impossible and can only rely on the leadership of strong men to do things. But at DevCon, a political scientist told me that one of the reasons he respects Ethereum very much is that we are a truly open and decentralized ecosystem, and we have succeeded at this scale so far, which gives him hope. So if we can succeed in this way, the positive impact on the world may be very large, and it will give many people a bright example of success that they can follow.

But "decentralization" does not mean "doing nothing." The philosophy of subtraction of the Ethereum Foundation does not mean "reducing the foundation to 0", but a way to maintain ecological balance. If there is an imbalance in the ecology in one place (for example, a part of the ecology is too centralized, or there is an important public good that others do not do), we can help counterbalance. After solving this problem, the foundation can withdraw from that area. If an imbalance occurs in a new place, we can move resources there, and so on.

In Chinese culture, the way we pursue may be most similar to the ideas of the Tao Te Ching, but taking this path requires wisdom and the ability of the foundation to be improved in some areas. It is not a matter of "success by doing nothing". So in the short term, we need to put more effort into making some important pivots.

Here we actually need to find a way to solve three problems at the same time:

• Attract more developers

• Encourage developers to develop more open source, secure applications that comply with public standards and have long-term value, etc.

• In the process of solving (2), avoid the ecosystem from becoming a closed circle (the phenomenon of "we are aligned because we are good friends of developers")

So I recently said that Ethereum alignment should be a technical game, not a social game.

I want to emphasize this issue because I think that in the area of ​​decentralization, the most pressing centralization issue is often not the L1 issue, but the L2 or wallet or application issue. So the entire ecosystem needs to work together to expand and attract new developers and make progress in these decentralized and trustless aspects.

There are several ways we can help achieve this:

• Education, so that developers can more easily understand why blockchain is, what should be on the chain, what should not be on the chain, what to care about in the field of blockchain, etc.

• If some blockchain-specific technical parts are too difficult for application developers, the foundation can do it itself, so that developers can more easily integrate. For example, zk's programming language, and a16z's helios, etc.

• Give developers clear standards. For example, if you are making an Ethereum client, there are many tests. You can run the tests yourself to see if your client can pass. If you are making L2, there are frameworks such as l2beat stage 1, stage 2, etc. This should also be available for zk applications, wallets, etc.

I am very interested in many non financial zk use cases, such as:

• anti-sybil verification. Many services require you to log in with kyc not because they really want to know who you are, they just want to know that you are not a robot, or if you are banned, you can't reopen your account 100,000 times. To implement this use case, you only need zk proof of personhood, or proof of reputation. In fact, sometimes proof of tokens is also enough, such as anonworld.

• Use cryptography to develop privacy-preserving AI applications. Zk may not be the most suitable technology here, but FHE may be. FHE has made a lot of progress recently. If we can further reduce the overhead of FHE, there may be opportunities.

• Wrap any web2 account with zk-snark and use it in web3. zkemail, anon aadhaar, zkpassport, zktls, etc. are good examples.

I think this technology has many opportunities to solve many security, governance and other issues in social and other fields by protecting personal freedom and privacy.

Here we need to make an important correction: d/acc is not de-acceleration, but decentralized defensive acceleration.

This is important, because there are indeed people in this world who support deceleration, degrowth, etc., but I think this direction is wrong. In a peaceful world, it will delay the improvement of important medical and infrastructure and cause more people to get hurt. In the more dangerous world now, if we don't accelerate, we will be eaten by those who are willing to accelerate.

Decentralized and defensive technology need to compete with other technologies. If the sword advances rapidly, but the shield does not, the world will become more and more dangerous. If centralized technology advances rapidly, but decentralized technology does not advance, the world will become more and more centralized. So we need to counterbalance these trends. Blockchain is part of this story, but only part of it. There is also decentralization outside of blockchain (e.g., p2p networks), software and hardware security (digital "shields"), and many things in the biological field, etc. The latter is the most important now. Technological breakthroughs have already been made: zero-knowledge proofs, consensus algorithms, virtual machines, etc. User adoption has already occurred, but the adoption that has received the most attention is the memecoin that fell 97% in one day (I am not against all memecoins here, I was an early buyer of DOGE anyway, but the current one is a completely different type). I think the applications we need should pass 3 tests: Can you imagine yourself or someone you know actually willing to use it? That is, the difference between something that is interesting in theory (decentralization uber! Great!) and something that can actually be used.

It is difficult to pass all three at the same time. Maybe only payment and store of value applications, which may predict the market, have passed now. We need to add 10 more successful examples.

Perhaps the most disappointing time for me recently was when someone said that Ethereum is not good and intolerant because we don't respect the "casinos" on the blockchain enough. Other chains are happy to accept any application, so they are better. If the blockchain community has this kind of moral reversal, I have no interest in participating in the blockchain. But I found an interesting point: on the Internet, many people will say those things, but when I chat with the community in person, everyone's values ​​are still the same as before, so I feel that I have a responsibility to this community and cannot abandon them. We Ethereum need to work together to create the world we want to see. This will require some changes, such as the foundation may not be too neutral at the application layer and need to specifically support some things, but this project is worth doing.

This is also a problem I am very concerned about. Personally, I have been working hard to move most of my conversations from Telegram to Signal in the past two years. But Signal is also imperfect. Although it is confidential, it is still centralized, has no interoperability, requires a mobile phone number to log in, and the server sees a lot of your metadata, etc.

But it is difficult to make a higher quality messenger. I try status every year. They try to be completely decentralized. They do a good job, but they still have some reliability issues. In fact, there are various small teams making their own messengers, but they are not united, so it is easy for each one to be not good enough.

I recently started using fileverse to make my various documents. I found that the user experience of this is good enough. Now many people in the foundation use it. If there is a decentralized, encrypted messenger that can achieve this quality, I will definitely work hard to help the community move to this messenger.

I think Ethereum is a decentralized ecosystem, not a company. If Ethereum becomes a company, we will lose most of the meaning of Ethereum's existence. Being a company is the role of a company. In fact, there are many big companies in the Ethereum ecosystem: Consensys, various client teams (Nethermind, Nimbus, etc.), Coinbase, L2 teams (including Aztec and Intmax, whose privacy technology is very interesting and underestimated by many people)

The best way is to find ways to give these companies more opportunities to realize their advantages, and the foundation plays a coordinating role.

There are many different people with different stories.

For example, many people in the blockchain circle would say 10 years ago that the goal of blockchain is to be a global neutral system, protect individual freedom, and counterbalance government hegemony. Now, if a president issues a memecoin, they would say, wow, this is real world adoption, so good, but why is it happening on other chains? If we can be more friendly to those politicians, it will happen on our chain next time! I personally think that such people have gone astray. Of course, they will say that I am too purely idealistic, unrealistic, etc. Each side has its own story.

Some people will also say that the Ethereum ecosystem is too controlled by OGs and there is not enough space for new people to come in. But this criticism is in another direction, and there are different groups making these arguments.

I think there is only one suitable way for us to get out of these predicaments: we need to have some updated stories to explain why Ethereum is, what the ETH coin does, what L1 and L2 do, etc.? This is no longer the era of infra, but the era of applications, so these stories cannot be abstract "freedom, openness, anti-censorship, solarpunk public goods, etc.", and some clear application layer answers are needed. In the near future, I plan to support more: info finance (this is also the direction of AI + crypto), privacy protection, high-quality public goods financing methods, and continue to do a good job in the world's open financial platform, which of course must include real world assets. There are many things here that are valuable to many users at the same time and are in line with the values ​​we have always had. We need to support this direction again, and this will also give new people more opportunities to come in.