Conversation with Vitalik Buterin: Discussing Ethereum POS, Layer 2, and AI Fusion Innovation

On April 7, 2025, at an event in Hong Kong, Yan, the founder of the DappLearning community, interviewed Vitalik Buterin. The following are the main points of the interview:

01 Views on POS Upgrade

Yan asked Vitalik about his views on the Ethereum POS upgrade. Vitalik said:

"The price trends of BTC and ETH are unrelated to POW and POS. The actions and thought processes of the two communities are completely different. For the price of ETH, one issue is that there may be many successful applications on Ethereum, but these applications may not bring enough value to ETH. This is a concern for many people in the community, but it is actually quite normal. For example, Google has developed many products, but over 90% of its revenue still comes from its search business. The relationship between Ethereum's ecosystem applications and the price of ETH is similar. We need to support more applications that are beneficial to ETH holders and the long-term value of ETH."

02 PBS Architecture and Centralization Concerns

Regarding the potential centralization issues of Builders brought by the PBS architecture of ETH 2.0, Vitalik believes:

"This is an important philosophical question. Early Bitcoin and Ethereum had a potential assumption: building blocks and verifying blocks are one operation. Now we have new technologies like ZK, DAS, etc., and building and verifying can become asymmetric. The difficulty of building blocks may become very high, but the cost of verification will become very low.

In this case, if there are some super nodes in the network, we need to consider how to prevent them from acting maliciously, such as 51% attacks, censorship attacks, MEV-related operations, etc. For the 51% attack, since the verification process is completed by the Attester, the cost of verification is very low, so the threshold for becoming a consensus node is still relatively low. For censorship attacks, we have the concept of Fossil, which allows us to separate the roles of selecting transactions and executing transactions.

This method is more complex than before, but it can ensure the security and decentralization of the network. This is also a challenge we face.

03 The Relationship Between Layer 1 and Layer 2 and Future Directions

Vitalik believes that there are several issues with the current Layer 2.

1. The progress in security aspects is not fast enough. He has been pushing for the Layer 2 upgrade to Stage 1 and Stage 2.

2. Interoperability issues between Layer 2. They are developing solutions such as the Open Intents Framework.

3. The withdrawal time for Optimistic Rollup is too long. It supports a 2/3 verification method of OP+ZK+TEE, which can reduce the withdrawal time from 1 week to 1 hour.

Regarding the role of Layer 1, Vitalik stated:

"If Layer 1 plays no role at all, it is dangerous for ETH. One of our biggest concerns is: the success of Ethereum applications cannot guarantee the success of ETH. If ETH fails, it will result in our community lacking the funds to support the next round of application development. Therefore, we need to allocate more roles to Layer 1 in some applications."

He mentioned several solutions to improve Layer 1 scalability, including delayed execution, block-level access lists, and multi-dimensional gas pricing. Combining these solutions could increase the gas limit of Layer 1 by 10 times.

04 Views on Cryptography and AI

Regarding the development of cryptographic technologies such as ZK, MPC, and FHE, Vitalik stated:

"FHE has a good prospect now, but one concern is that MPC and FHE always require a committee. If the committee is attacked, the system will have problems. Moreover, the committee needs to be online all the time, which increases the risk of being attacked. This worries me a bit, but many applications can still be developed; there are advantages but it's not perfect."

Regarding the impact of AI, Vitalik believes:

"AI can enhance the capabilities of everyone, and sometimes it may also strengthen certain centralized powers. But AI also has a democratizing effect. It can narrow the gap between experts and novices, bringing new opportunities to all."

05 Expectations for the Ethereum ecosystem and suggestions for developers

Vitalik told Ethereum developers:

"There are many new opportunities for developing Ethereum applications now. Layer 2 addresses the TPS issue, solutions for privacy concerns are available, and AI has lowered the difficulty of development. Many projects that previously failed may now succeed.

The biggest problem in the current blockchain application ecosystem is two extremes: one is highly idealistic applications with only a few users, and the other is casino-like applications. We hope to develop applications that users enjoy using, are valuable to the world, and have sustainable business models.

Now everyone has more resources than before, and if you have good ideas and can execute them well, the chances of success are very high.