Author: Haotian

I recently had a common feeling after chatting with some entrepreneurs and VCs. Everyone's expectations for the AI ​​+ Crypto track are still firm, but they are a little confused about the narrative evolution of web3 AI Agent. What should I do? I have sorted out several potential changes in the subsequent AI narrative for reference:

1) AI Agent uses MEME to issue coins as a result of MEME, and is even "despised". If the project does not have PMF support and only a set of Tokenomics is idle, it will naturally be labeled as pure MEME hype, just a wolf in sheepskin, and has little to do with AI;

2) Originally, AI Agent > AI Framwork > AI Platform > AI DePIN implementation order may be adjusted. When the Agent market bubble bursts, Agent It has become the "carrier" after the formation of technologies such as big model fine-tuning and data algorithms. Without the core technical support behind it, it is difficult to have the opportunity to show off your muscles again with just one AI agent;

3) Some projects that originally made service platforms such as AI data, computing power, and algorithms will surpass AI agents and become the focus of attention. In other words, even if a new AI agent is launched, the agents created by these AI platform projects will be more convincing. After all, if you can run an AI platform project, your team and technical foundation are much more reliable than a Dev that is deployed at a low cost based on the framework; 4) web3 AI Agent can no longer fight with the web2 team, and you have to find the direction of differentiation in web3. The web2 Agent focuses on Utility, so the logic of deploying the development platform at low cost can work, but the web3 Agent focuses on Tokenomics. Overemphasizing low cost deployment will only stimulate the bubble of more asset issuance; there is no doubt that the web3 AI Agent should combine the blockchain distributed consensus architecture to innovate and develop (the article on the top of my homepage is described in detail);

5) The biggest advantage of the AI ​​Agent is "application pre-position", which belongs to the logic of "fat protocol and thin application", but how should the protocol be fat? How to mobilize idle computing resources and use distributedThe advantages of architecture-driven algorithms are to activate more vertical segmented scenarios such as finance, medical care, and education. And how to lose weight when applying? Let AI Agent manage assets independently, independently intend to trade, autonomous multimodal interaction, etc. are not achieved overnight. You cannot try to eat a big fat man in one bite, split the needs and slowly implement them. Otherwise, the mature standards of a DeFai scenario will have to be worn out for one or two years;

6) The MCP protocol in the web2 field and the Manus automated execution of multimodals, etc., are all inspiring for innovation in the web3 field. They can directly develop suitable web3 application scenarios based on MCP +Manus, or use a distributed collaboration framework to make enhanced business scenarios above MCP. Don’t just talk about subverting everything. It is enough to be able to properly optimize based on existing product protocols and give full play to the irreplaceable differentiation advantages of web3. Whether it is web2 or web3, they are in the process of this great innovation of AI LLMs. Ideology doesn’t matter, it is important to truly promote the development of AI technology.