Two years later, Mr. V tweeted again. I followed the research paper two years ago and said that even the time was the same. February 10 Two years ago, V God had already subtly expressed that he was not very optimistic about the various Crypto Help AI that were popular at the time. The three popular strategies in the industry at that time were the assetization of computing power, data, and models. My research report two years ago mainly talked about some phenomena and doubts observed by these three horses in the primary market. From the perspective of V God, he still prefers AI Help Crypto The several examples he gave at that time were AI as a participant in the game AI as a game interface AI as a Game Rule AI as a Game Goal In the past two years, we have actually made many attempts on Crypto Help AI, but the results have been minimal. Many tracks and projects are just about issuing a coin without a real commercial PMF, which I call the "tokenization illusion" 1. Capitalization of computing power - Most cannot provide commercial grade SLAs, are unstable, and frequently experience disconnections. Can only handle simple small and medium-sized model inference tasks, mostly serving edge markets, with revenue not tied to tokens 2. Data assetization - There is high friction, low willingness, and high uncertainty on the supply side (retail investors). On the demand side (enterprise), what is needed is a structured, context dependent, trusted, and legally responsible professional data provider, which is difficult for DAO based Web3 project parties to provide 3. Model assetization - Models are inherently non scarce, replicable, modifiable, and rapidly depreciating process assets, rather than terminal assets. Hugging Face itself is a collaborative and propagation platform, more like GitHub for ML than App Store for models. Therefore, the so-called "decentralized Hugging Face" tokenizing models is mostly a failure In addition, we have also tried various "verifiable reasoning" in the past two years, which is a typical story of using a hammer to find a nail. From ZKML to OPML to Gaming Theory, and even EigenLayer has transformed his Restaking narrative into Verified AI based. But it's basically similar to what happened on the Restaking track - few AVSs are willing to pay continuously for additional verifiable security Similarly, verifiable reasoning is mostly about verifying 'something that no one really needs to be verified', and the threat model on the demand side is extremely vague - who are we really guarding against? AI output errors (model capability issues) far outweigh AI output being maliciously tampered with (adversarial issues). As we have seen in various security incidents on OpenClaw and Moltbook recently, the real problem comes from The strategy design is wrong Too much permission given I haven't figured out the boundary clearly Unexpected interaction in tool combination ... There are almost no imaginary nails such as "models being tampered with" or "reasoning processes being maliciously rewritten" I sent this picture last year. I wonder if Lao Tie remembers it The several ideas provided by V God this time are obviously more mature than two years ago, also because we are concerned about privacy, X402，ERC8004， Predict the progress made in various directions such as the market You can see that he has divided the four quadrants this time, with half belonging to AI Help Crypto and the other half belonging to Crypto Help AI, instead of clearly leaning towards the former two years ago Top left and bottom left - Utilizing Ethereum's decentralization and transparency to address trust and economic collaboration issues in AI 1. Enabling trustless and private AI interaction (infrastructure+survival): Using technologies such as ZK and FHE to ensure the privacy and verifiability of AI interaction (I don't know if the verifiability inference I mentioned earlier counts). 2. Ethereum as an Economic Layer for AI (Infrastructure+Prosperity): Enable AI agents to conduct economic payments, recruit other robots, pay deposits, or establish a reputation system through Ethereum, thereby building a decentralized AI architecture rather than being limited to a single giant platform. Top right and bottom right - Utilizing AI's intelligent capabilities to optimize the user experience, efficiency, and governance of the encryption ecosystem: 3. Cypherpunk mountain man vision with local LLMs (Impact+Survival): AI serves as the "shield" and interface for users. For example, local LLM (Large Language Model) can automatically audit smart contracts, verify transactions, reduce reliance on centralized front-end pages, and safeguard individual digital sovereignty. 4. Make much better markets and governance a reality (impact+prosperity): AI deeply participates in Prediction Markets and DAO governance. AI can act as an efficient participant, amplifying human judgment through large-scale processing of information, solving various market and governance problems such as insufficient human attention, high decision-making costs, information overload, and apathy in voting Previously, we were crazy about wanting Crypto Help AI, while V God was on the other side. Now we finally meet in the middle, but it seems to have nothing to do with various XX tokenization or AI Layer1. I hope to look back at today's post in two years, and there will be some new directions and surprises