Some people always think that my criticism of web3 AI means that I am biased towards web2 AI. Actually, that is not the case. Any content I write about web2 DeepSeek, Manus, MCP, etc. does not mean that I am "partial". From the first day of waving the flag for AI+Crypto, I have been thinking about how to innovate web3 AI, how to evolve the narrative, where the development problems are, and how to adjust. Now and in the future. The reason is very simple:
1) My main focus is on the AI + Crypto field. I know a little about the innovative development of web2, but not enough to see the whole picture, so I can't make an absolute evaluation. Like most people, I only know that after a series of innovations in web2 AI, web3 AI collapsed. After careful study, I found that web2 AI is indeed making steady progress in terms of talent echelons, innovative ideas, pragmatism, etc., which is in line with the market's expectations for the development of AI, that's all.
I have indeed imagined how great it would be if such innovation appeared in the web3 AI track. There would be no need to compete with the web2 team. It would be enough to find a differentiated route and make some web3-specific narratives and technical implementations. At least it would be worthy of the technology believers Holders who bought at the bottom in the previous wave.
Unfortunately, expectations have been dashed time and time again. Watching the price of coins of those mainstream projects approaching zero, and watching new and old believers sell at a loss, I am really angry. Are you not happy with the narrative hype that cannot keep up with web2 in technological differentiation and innovation, and web3's expertise?
2) Sitting on Crypto, there is really no such grand mission as AI to revolutionize human technology. From the beginning, I understood that AI is an opportunity for the Internet, and Crypto also needs this opportunity. The real meaning of AI + Crypto is actually to give Crypto a new soul of AI, so that the "old narratives" that Crypto once hyped and fell due to excessive bubbles can be revitalized one by one. We only need to regard AI as a new productivity catalyst.
Based on this premise, whether it is new web2 developers entering the market to do AI, or original web3 developers who want to innovate with new narratives, they should all anchor on one goal: use web3's Tokenomis + decentralized flexibility and resource advantages to innovate the web3 Native narrative.
Web3ers must admit that the key to web2 AI Agent's success at this stage lies in the leapfrog innovation experience it brings to C-end users through "monopoly + scale". This is an advantage that web3 AI Agent cannot have. If it tries to compete head-on, web3AI is no match at all.
However, it is precisely some upstream distributed framework layers, some distributed trust frameworks and incentive network designs that web2 is unwilling to touch, and component designs such as interactive communication on-chain outside the Agent computing layer that are worth focusing on for web3. That is also the point that "future AI" cannot avoid when it develops to a certain stage, and it is also the differentiated advantage of web3AI.
