ETHTaipei Interview with Vitalik: Ethereum Faces a Critical Moment in the Next Five Years, Most Often Eats Dark Chocolate

Compiled by: GaryMa, Wu Says Blockchain

During the recent ETHTaipei 2024 conference, Vitalik participated in interviews with several media outlets, covering topics such as lifestyle and diet, the Dencun upgrade, Ethereum-related roadmaps (upgrade expectations, centralization issues, modular restaking, etc.), security and privacy (quantum resistance, ZK privacy security), decentralized social media, AI, industry advice, and more.

Q: As a nomad, what fast food do you often make for yourself lately?

A: Dark chocolate would be one of the main components.

Q: When you travel and live around, which experimental community activity has inspired you the most?

A: It depends on what you mean by experimental community activity, but broadly speaking, last year's Zuzalu might be the biggest thing in that category (Zuzalu is a pop-up city experiment initiated by Vitalik in Montenegro from March to May 2023). There was also an event in Vietnam, which was the first event I attended that was entirely outdoors. It reminded me of the atmosphere of West Lake in Hangzhou. It was also the safest pandemic conference I have attended, which was cool.

Q: What recent work are you most interested in? It doesn't have to be Ethereum or crypto-related.

A: This year, Ethereum-related affairs have been quite low; even if there are any, they are in the community aspect. Various independent groups are forming their own projects, and I am following up and providing help as much as I can. I'm excited to visit some of these project ecosystems soon. Recently, I have been trying out some of the latest AI tools, running models locally, and applying them to various tasks. Deepening my understanding in this area and exploring new possibilities is very important. Learning is crucial, and I believe exploring various topics is valuable. So, these are the main highlights.

Q: What contribution does the Dencun upgrade make to the Ethereum ecosystem?

A: The goal of the Dencun upgrade is to significantly improve scalability and reduce transaction fees, especially for L2 and Rollups. It achieves this by creating a separate database section (called blobs) within each block. This data is inaccessible to the EVM, which is important because it means that clients validating Ethereum blocks do not need immediate access to this data. However, the Ethereum blockchain guarantees its availability. This feature is particularly beneficial for Rollups and any L2 projects relying on secure data availability, ensuring that nodes can synchronize when the current node disappears or challenge incorrect answers.

In the past week, we have observed a significant decrease in layer two fees, sometimes even by 50%. However, it is important to note that as more users adopt blobs, fees may increase again. Nevertheless, this still represents a significant improvement in scalability, and we expect the number of blobs supported by the Ethereum chain to continue to increase in the coming years.

Q: Has the performance after the Dencun upgrade exceeded expectations?

A: I think it depends on your understanding of "exceeded expectations." From a technical perspective, the upgrade has progressed smoothly. The network participation rate has only dropped from 99% to 95%, which is better than any of our previous hard forks.

Interestingly, the usage is currently very low. The goal is about 3 blobs per block, but the average usage is only one blob per block. This means blobs are currently very cheap. If you want to publish a blob, you basically only need to pay the Ethereum transaction fee. The high Ethereum transaction fees may be one reason for this, but if the price of blobs is close to zero, they can be used for various purposes, such as backing up encrypted hard drive copies. There are infinite uses for guaranteeing databases. I believe usage will eventually increase, but for the current Rollups, having a very low price is beneficial. I look forward to seeing usage increase in the coming months.

Q: What do you think will be the most transformative impact on society in the next five years?

A: I believe the next five years will be crucial for Ethereum, as many applications that were once theoretical or small-scale are now ready to be applied in the real world. The impact of blockchain technology has begun to permeate the broader world, often unnoticed. For example, Reddit's upcoming IPO provides active community members with the opportunity to participate alongside institutional investors, thanks to ideas from the crypto space.

In terms of practical use, stablecoins have had the most significant impact, facilitating savings, trading, and remittances. In the next five years, the user experience and fee structure of Ethereum will improve, and the development of projects like L2 and Base will pave the way. As Ethereum becomes more user-friendly, it is expected to lead in making stablecoins accessible and decentralized.

In addition to the financial sector, non-financial applications are also experiencing growth, particularly in decentralized social media alternatives like Farcaster and Lens. Decentralization offers unique advantages, such as allowing anyone to develop new clients or access and write the same content without starting from scratch. This vibrant ecosystem is expected to expand further.

Ethereum-based identity solutions are also rapidly growing, including advancements in personality proof protocols and social graph-based systems. These solutions aim to address the challenges of verifying human users on online platforms, mitigating the risk that centralized solutions may exclude certain groups. Ethereum has the potential to play a leading role in developing decentralized alternatives to this issue.

Overall, these developments highlight the importance of Ethereum's evolution and its continued growth potential across various fields.

Q: How do you view the current challenges in PoS, and how will SSF (single slot formality) and other upgrades address these issues?

A: I believe the main challenges currently facing proof of stake primarily involve various centralization risks. One significant concern is regarding MEV (miner extractable value) and the risks associated with proof of stake and the functionality of becoming a validator itself. In terms of MEV, we are witnessing increasing centralization and censorship risks, with relayers becoming another centralized participant. Technologies like execution tickets and inclusion lists aim to mitigate these risks, striving to maintain decentralization and fairness in block creation while assigning specific centralized functions to builders.

Another challenge lies within proof of stake itself. According to recent surveys, the main reasons for not participating in proof of stake are holding less than 32 ETH and the perceived difficulty of running a node. However, a technical roadmap has been developed to address these issues. For example, vertical trees have made significant progress in reducing the storage requirements for running nodes from multiple terabytes to more manageable levels, potentially allowing nodes to run in RAM. Additionally, with technologies like ZK-SNARKs, the computational requirements for running nodes will be further reduced, enabling a broader user base to run nodes easily.

The issue of having less than 32 ETH for staking is more complex. Initially, the choice of 32 ETH was to balance the need for a significant amount of ETH for validation and to avoid too many validators seeking balance, which could lead to processing difficulties. New approaches to proof of stake include relaxing the requirement for each validator to participate in every round of consensus, allowing for single slot finality and the ability to validate with less than 32 ETH. Rainbow staking and BLS signatures are among the proposals being researched.

Q: We have seen many projects proposing "modular blockchain" solutions, and we have also seen the idea that Ethereum L1 could be responsible for shared ordering. How do you view this broader issue: what functions should L1 centralize, and which should be left to individual L2s?

A: Modularity suggests that the functionalities of various independent chains will be reduced in the future, with different components managed by different parts, while shared ordering becomes increasingly important, as advocated by one of our researchers, Justin Drake.

Currently, Ethereum L1 is responsible for shared security and settlement, ensuring that each L2 can interact without relying on centralized participants or validator sets. Ethereum provides data availability for Rollups but does not include volumes, and the ordering of transactions is determined on a per-Rollup basis.

Opinions on shared ordering are divided. Some advocate for its benefits, while others believe it is overestimated. They argue that the benefits of shared liquidity are somewhat limited, and cross-L2 MEV is not as significant as people imagine, as it can be decomposed into MEV between different L2s.

Expanding Ethereum's ability to directly support more data is crucial. While ideally, everything should be a Rollup, practical limitations require exploring optimistic data constructions for off-chain data, with high-security features retained on-chain through Rollup.

Account abstraction also presents another challenge. Determining where to store account state and effectively managing updates across multiple locations requires innovative solutions. One approach is to adopt a minimal key vault Rollup, where state is stored in a neutral Rollup above Ethereum and is accessible to other L2s.

These ideas are still in the early stages and are under ongoing discussion within the Ethereum community. In practice, key functions that need to be handled by L1 should remain there, while providing flexibility for different L2s to handle various functionalities is also very valuable.

Q: If you have a system that theoretically gives you some level of ZK privacy, how do you, as a user, know that you actually have that level of privacy?

A: I think this essentially continues the issue that Ethereum has already faced: ensuring the security of assets in smart contracts. If you are entrusting assets to a smart contract rather than an individual, how can you ensure there are no backdoors allowing unauthorized access?

Current solutions include reading contracts on Etherscan, where people can publish source code. While this is useful for knowledgeable users, it is unrealistic for ordinary users who cannot audit complex code themselves. Wallets have started to address this issue by providing more warnings, such as alerts when interacting with less familiar applications.

One improvement I envision is a versioned DApp user interface hosted on IPFS, where each update requires a blockchain transaction for authorization. This eliminates the risk of unauthorized updates through server breaches. Then, wallets can display information about the website's recent updates and approval status.

We also need better mechanisms to aggregate opinions from high-quality researchers and auditors. Wallets could play a key role in assisting users by aggregating this information. The same tools used for Ethereum should be applied to ZK technology, including publishing source code and verifying its authenticity.

For example, tools like Etherscan could publish ZK source code and verify it on-chain. This approach ensures transparency and security for both on-chain and off-chain ZK technologies, such as Zoopass.

These efforts will extend the same principles used in Ethereum contracts to the realm of ZK technology, ensuring a secure and transparent ecosystem for users.

Q: How do you think Ethereum will address the urgent issue of quantum threats, and what impact will this have on the broader cryptocurrency ecosystem?

A: An important point is that, from a technical perspective, we have quantum-resistant algorithms for every vulnerability against quantum computers. Quantum computers can break existing elliptic curve signatures, but we have various solutions based on hash functions, lattices, and homomorphic foundations. For instance, lattice-based and homomorphic solutions can address the impact of quantum attacks on elliptic curve encryption and implicit addresses. Additionally, recent technologies like Starks have made breakthroughs in reducing size, providing quantum resistance due to their reliance on hash functions.

Fully homomorphic encryption is fortunately inherently quantum-resistant because quantum computers do not affect lattice-based cryptography. While this theoretically solves the problem, there are still logistical challenges between theory and practice. However, emergency recovery mechanisms can protect most funds, although efforts are underway to ensure complete quantum resistance for users and protocols.

To achieve this, several key steps are necessary. First, implementing account abstraction will allow users to choose quantum-resistant signature algorithms. Additionally, the Ethereum consensus layer needs quantum resistance, although engineering challenges have arisen due to the efficiency of current signature schemes (like BLS). This highlights the importance of exploring alternatives, such as 8192-bit signatures per slot, to accommodate less efficient quantum-resistant algorithms while researchers work on optimizing and benchmarking post-quantum alternatives.

Q: In your view, what are the benefits of combining AI with cryptocurrency? How might this reshape the industry? You discussed the role of AI in debugging code in February; could you provide more thoughts on how AI can be used in crypto and Ethereum?

A: AI and cryptocurrency are two very important technological trends of our time. There is a saying that AI tends to centralize, while cryptocurrency is striving for decentralization. There is a certain complementary relationship between the two. But the question is whether we can transition from this concept to practical application examples in a way that is meaningful and productive for both.

One of them is that AI can participate in predicting markets on Ethereum and other types of markets, essentially making markets more micro-focused and creating APIs that can play a role in them. Another is integrating AI as part of wallets to help users understand the crypto environment they are interacting with. The third is leveraging cryptography, including ZK, SNARK, MPC, etc., to embed AI, creating secure, robust, and privacy-preserving AI models that become central participants in on-chain applications, whether DApps or other forms like Oracles. Then the fourth is that if successful, it can be applied to AI in other fields.

Among these applications, I think the first two are the most obvious short-term applications, while the latter two are more speculative. I absolutely do not want to give the impression that the application of AI and cryptocurrency will become the next major narrative driving the industry forward. However, I do believe these intersections are worth exploring in depth.

Additionally, another application is that AI can perform rolling debugging of code, which is one of the biggest challenges currently facing the crypto space: code errors and vulnerabilities. A promising possibility is that AI could greatly simplify the process of using formal verification tools to prove that a codebase meets specific properties.

Q: What are your thoughts on the wave of restaking in Ethereum?

A: Restaking is indeed an interesting idea, but the potential risks are that demand could be taken over by centralized participants, along with various systemic risks and impacts on the Ethereum validator set. I know there are some projects researching different approaches, and so far, I am just observing this area, looking forward to future developments.

Q: You seem to be more active on Farcaster than on Twitter. Can decentralized social applications like Farcaster really compete with Web2 social applications? Or can they disrupt social models?

A: Yes, I think one interesting aspect of the social space is that it has network effects but also anti-network effects. Twitter is a platform that gathers all kinds of crowds, but it is also a gathering place for those truly annoying people. I find that Farcaster has also accumulated enough users and is interesting; I can get higher quality interactions on this platform, although it still needs improvement.

I believe that while Twitter may differentiate between high-quality content and crowds versus low-quality content and crowds, many criticize these mechanisms, including moderation and blue verification, arguing that they may become tools of centralization or privilege, serving specific viewpoints and groups. The crypto space, along with many non-financial applications like Ethereum, fundamentally addresses this issue and attempts to solve such trust problems. Decentralized social media can truly resolve this issue.

Another important feature of Farcaster is that it is not a server but a shared platform. Therefore, in principle, you can use your own client, and your client can read or write the same content that all users on the network can see. I think this can definitely become a place where people start trying to create their own clients and add interesting new features. For example, you can imagine people posting their community notes on Farcaster or publishing various AI-based workshops or even predictions. On Farcaster, you can envision people creating their own mechanisms to identify high-reputation participants and low-reputation participants, etc. Different groups can adopt different approaches. All these things can be accomplished by different clients.

Another thing that excites me is that I find the Farcaster application particularly practical, smooth, simple enough to use, and is being used by some non-crypto people. I think this is something many applications have not achieved, and it sets a template for other similar applications.

Q: Do you have any specific advice for non-native English speakers participating in the public goods ecosystem?

A: First, if you have a certain level of English skills, translating some content is valuable, as well as aggregating some information or actively participating in the co-construction of some protocols (communities), as many protocols are trying to welcome more participants from around the world or to become voters or badge holders.

Q: Why are you so interested in longevity? If you were immortal, what would you do?

A: I just love life; life is beautiful. If I were immortal, what would I do? Probably the same things I do now, just for a longer time.

Q: What has surprised you the most in the development and evolution of cryptocurrency? What do you think the cryptocurrency industry will look like in five years?

A: In 2020, NFTs really surprised me, as did the resurgence of meme coins. On the positive side, the rapid advancement of ZK technology is particularly surprising, as such progress is rare in software development. Additionally, the speed of changes in L1 protocols, such as the merge and Cancun, is impressive.

Q: Can you give some advice to new blockchain developers?

A: I think the most important point is to motivate yourself to participate and stay engaged, to truly start joining and becoming part of the community. Finding a clear direction and then creating something that is genuinely "usable" is much more important than just thinking without doing. You can also try to join this field through writing, which is a way to push yourself.