Finally! The Ethereum Foundation has established the dAI team, and I particularly want to ask: what took so long? I have repeatedly emphasized in several articles that AI Agents will become the "redemption" of many Crypto narratives. For Ethereum to solidify its position as a decentralized, secure settlement layer for global public chains, the technical debt has become too heavy; it needs a new application scenario that can refresh old technologies, and that will likely be AI Agents. (I previously called for Ethereum to embrace an AI grand strategy) So, what highlights does this dAI team plan to focus on? 1) Creating an "AI Economy on Ethereum" that allows AI Agents to autonomously pay and coordinate on-chain, eliminating intermediaries. This will require the maturity of many other AI infrastructure components: For example, AI Memory contextual memory solutions, AI Oracle real-time dynamic price feeding solutions, AI Processor or Multi-Agents autonomous coordination management scheduling platforms, AI Agent autonomous DAO governance solutions, etc.; Following the tradition of @VitalikButerin where he points, will this spark a wave of innovation in AI components within the Ethereum ecosystem? 2) As for the Decentralized AI Stack, it has shouted the slogan of "open, verifiable, and censorship-resistant" AI infrastructure, but the problem is that "openness" and "censorship resistance" are almost non-existent at this stage. Because the core computing power and models are controlled by giants like OpenAI, Anthropic, and Google, before decentralized computing power, algorithms, and federated learning platforms mature and scale, web3AI is completely constrained by a single API hook, lacking any initiative. Currently, the only credible aspect is "verifiability"; in AI Agent scenarios, technologies like ZK-Proof and TEE can be applied on a large scale, which will provide more imaginative space for the ZK track. As for the ERC-8004 goal mentioned in the announcement to solve the identity verification standards for AI Agents, I don't think it's an innovative starting point, but it seems that any technological innovation on Ethereum will begin with establishing standards 😅. As for the concept of d/acc, it's even more touching, but it's not important. For Ethereum to truly become the "neutral base layer" for AI, it needs not just technical standards but also killer applications and rapid follow-up from the developer ecosystem.