Written by: Techub News Compilation

Introduction

With the rapid development of AI agents, the possibility of them interacting autonomously, holding assets, and even participating in governance is moving from science fiction to reality. This brings unprecedented opportunities and challenges to Ethereum, which serves as the "global settlement layer." In the latest episode of the WuBlockchain podcast, Ethereum co-founder Vitalik Buterin (V God) was interviewed with a focus on the long-term fundamental trends at the intersection of AI and cryptocurrency technology. In this forward-looking dialogue, Vitalik Buterin (V God) did not dwell on technical details but instead, from a philosophical and systems design perspective, outlined the new roles Ethereum needs to play and the infrastructure it must fulfill under an AI-dominated interaction paradigm.

Summary

• AI is the new user interface (UI), not the users themselves. It will fundamentally change the way people interact with blockchains, giving rise to complex workflows that interact with multiple protocols simultaneously.
• Blockchain is a natural platform for decentralized AI to achieve trusted collaboration, serving as an "economic layer" that coordinates different parties' controlled AIs through economic incentives and rules.
• Identity models need to be deconstructed, utilizing technologies like zero-knowledge proofs (ZK) to prove specific attributes (such as reputation or source of funds) on demand, without fully exposing identity.
• The development of layer two networks (L2) needs to go beyond simple EVM scaling and should provide specific functions (like privacy, high-speed settlement) based on application needs, forming a modular architecture complementary to L1.
• Excellent AI product experiences must prioritize user interests, place a high emphasis on privacy and security, and support seamless, secure interactions with users' custom environments and other internet protocols.

Is the “operating system” metaphor of Ethereum outdated in the AI era?

The interview began with a fundamental question: If blockchain is viewed as an on-chain operating system, what core foundational components must it provide in the agent era? What does Ethereum currently lack? Vitalik Buterin (V God) first dissected the two existing core functions of Ethereum: public bulletin board and on-chain computation. The former allows anyone to publish data that can be interpreted by applications in multiple ways, while the latter supports complex financial and application logics like DeFi.

He believes that the arrival of AI will not overturn these foundational use cases, but it will completely change the interaction paradigm. Before AI, users typically interacted with a specific application through a specific interface. In contrast, AI, especially the AI agents that will become commonplace in the coming years, will act as a "super intermediary." Users can interact through an AI assistant with various "skills," which can call and coordinate multiple distinct tools and applications simultaneously to accomplish complex tasks.

"Intuitively, this will greatly increase the number of workflows interacting with both Ethereum and other entities simultaneously," Vitalik Buterin (V God) explained. Therefore, simply comparing blockchain to an "operating system" may no longer be accurate. The future "operating system" may become smaller and simpler, with a series of tools and skills around it that are assisted by AI, each fulfilling its own role.

In this context, the core value of blockchain becomes increasingly prominent: it is the most natural way to achieve long-term, trustless collaboration among multiple parties. When the discussion turned to the difference between decentralized AI and centralized AI, Vitalik Buterin (V God) pointed out another key role of blockchain—economic layer. "If you have more decentralized AIs, it means that various AI entities (agents, programs, etc.) built and controlled by different people need to interact with each other." He stated, and this interaction is either based on centralized control or on economic incentives, rules, rewards, and punishments. "If we can establish a good economic system, then more decentralized interactions between AIs become possible."

Reconstructing Users, Identity, and Interaction Paradigms

When agents can autonomously trade, hold assets, or even participate in voting, how do we redefine "users"? From the perspective of Ethereum governance, how do we view these non-human participants? Vitalik Buterin (V God) is very clear on his position: humans are always users, while AI is the new user interface (UI). He likened the evolution of internet use: in the past, people obtained information through Google searches, visiting Wikipedia pages, or Stack Overflow; now, you can directly ask an AI assistant, which performs a series of actions to provide the answer. Blockchain interactions will also undergo similar transformations.

This transformation directly impacts the characteristics that blockchains need to have, with latency being the top priority. Low latency is usually important when humans interact directly. However, in the AI era, the situation becomes more complex. On one hand, if AI thinks a thousand times faster than humans, reducing latency to 1 millisecond can bring tremendous benefits; on the other hand, there are many scenarios where latency is not critical. "When you pose a complex question to Claude, GPT, or DeepSeek, it thinks, searches the web, gets replies… basically, after you ask the question, you could grab a coffee and come back to see the answer. So, even if the latency is a minute, that's actually not a problem."

This means that blockchain and its interface layers need to adapt to this new normal of polarized latency demands. Taking wallets as an example, Vitalik Buterin (V God) believes that core functionalities of the SDK (Software Development Kit) or API layer (such as sending tokens, querying balances, executing privacy functions) will remain unchanged, but the entire user experience surrounding them will undergo significant changes. "We may no longer need a traditional 'wallet' app in the conventional sense; you just need to interact with the SDK." In the future, we will see software packages that have undergone rigorous security audits and even formal verification, designed as "skill files," ensuring that they remain secure regardless of how they are invoked, while AI operates around these secure building blocks.

Regarding on-chain identity, Vitalik Buterin (V God) advocates for deconstructing identity as much as possible, considering what claims about oneself truly need to be proven in specific interactions. "For most interactions, revealing who you are completely doesn't make sense. But in most cases, revealing certain aspects of your identity is meaningful." He mentioned using zero-knowledge proofs to demonstrate various attributes like reputation and source of funds, and called for establishing a framework that facilitates such proofs while maximizing user information protection in meeting application needs.

Public Goods, L2, and Costs: Adapting New Rules for the Agent Economy

If agents become the main participants in on-chain economic activities, how should the financing mechanisms for public goods evolve at the protocol level to address these non-human capital flows? Vitalik Buterin (V God) acknowledged this is a tricky question, as public goods financing is fundamentally a governance issue, which requires clarity on who the stakeholders are. He believes that behind each agent is still a human—those who have reasons to run these agents. AI and zero-knowledge proofs (ZK) technology open up more possibilities for how this governance operates, but at the same time, AI disrupts many old mechanisms. "Many things that used to require great effort to undermine can now be automatically destroyed." This implies that governance mechanisms need to be continuously iterated and updated.

When the discussion shifted to the low latency and low cost needed for high-speed interactions between agents, the conversation naturally turned to layer two networks (L2). Vitalik Buterin (V God) affirmed the importance of L2 but criticized the industry's overly uniform development approach: "What we've been doing for too long essentially boils down to: take the EVM, make a copy, scale it up by 1000 times, make it more centralized, and then publish the hash to Ethereum, and that's it. The Ethereum community has realized that this actually doesn’t make much sense."

He proposed a more imaginative approach to L2 building: starting from application needs to think about what components L1 fails to provide, then let L2 fill those gaps. In an ideal world, different parts of applications would be distributed between L1 and various L2s. For instance, parts that are updated infrequently, like user accounts, can be placed on the more costly L1; while individual transactions, order matching, and other high-frequency operations could be placed on L2. Furthermore, L2 could specifically provide privacy functions. "Privacy solutions like Tornado Cash, Railgun, or Privacy Pools are, in my view, already L2—they weren't designed for scalability but for privacy, yet they are L2." Vitalik Buterin (V God) predicts that we will see more L2s developing towards various functionalities and directions in the future.

Building an Excellent AI Product Experience Centered on Users

As new aspects like agent wallets, agent app stores, and agent payment SDKs continually emerge, what constitutes an excellent AI product experience? Vitalik Buterin (V God) believes it encompasses multiple dimensions: it must be intuitive and user-friendly; it should operate well as part of a larger ecosystem, rather than attempting to take over the user's entire digital life; and it must take privacy and security extremely seriously.

He particularly pointed out the current open-source AI ecosystem's shortcomings in security and expressed the hope for building AI that genuinely aligns with user interests: "We need companies and teams to build AI that is not dependent on any specific company or application. Their goal is to serve users as well as possible." More intelligent beings that represent and help users will help ensure fair treatment and reduce incidents of user exploitation by applications.

At the same time, excellent AI products should be able to interact seamlessly with the users' existing, potentially highly customized settings, and securely connect with other tools that users want to utilize. Vitalik Buterin (V God) also revealed that the Ethereum core team and the related Kohaku team are exploring agent wallets, focusing on ensuring privacy, security, and supporting local AI. He emphasized that if the agent experience always relies on calling a third-party server, then it lacks both decentralization and privacy.

Finally, he envisioned future protocol-level native agent standards, highlighting ZK payments and the development of ZK APIs. Their goal is to keep each API request (whether invoking a large language model, conducting a web search, or making RPC calls) private, and requests unrelated to each other. "Especially for AI, if you use a persistent anonymous identity, you still leak enough information to lose privacy. What we need to do is isolate each request from one another." He admitted this is challenging, especially to avoid making on-chain transactions for every request (too costly), but they are exploring methods like combining secure collateral to prevent abuses on both user and application ends without compromising privacy.

The entire conversation painted a grand picture of the deep integration between AI and blockchain. In Vitalik Buterin's (V God) view, this transformation is not just a technological upgrade but a reconfiguration of interaction philosophy, economic models, and governance mechanisms. If Ethereum and its ecosystem want to continue to play a key infrastructural role in this wave, they must transcend existing paradigms and build a foundation of trust and collaboration in a more modular, privacy-first, and user-sovereign manner for the upcoming agent era.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。