The journey of building an infra and an application for Web3 AI is a long marathon
Words: Haotian
Recently, I have chatted with many developers on the front line of web3AI Build, and found that doing things around web3AI infra is much more complicated than I imagined:
1) At present, most of the active AI projects in web3 are generally MEME, boasting a lot of stories that cannot be realized and implemented, the key is to attract most of the attention and liquidity by cutting into the market through rapid issuance, and the chicken feathers (negative EV) after the short-term bubble bursts. Mainly because the narrative of AI + Crypto is too sexy, and at the same time, its practical application is too challenging, it has naturally become the hardest hit area of the bubble that relies on narrative to issue coins;
2) The web3AI infra is essentially a refactoring of the web2 AI infra and is thankless most of the time. Just like when Crypto challenged centralization in the name of decentralization, for a long time, the decentralized network architecture was criticized for being meaningless in repeated construction, until the subsequent DeFi application scenarios found some value capture points.
The current dilemma of web3AI is the same as the original vision of decentralized crypto. Most people are still accustomed to saying "what's the use of web3AI"? But don't forget, decentralized computing power aggregation and distributed inference, distributed data annotation networks, etc., can all find entry scenarios in terms of training cost, performance, and practicability.
3) The construction and expansion of web3AI infra has a large cost of trial and error period, and needs strong rationalism support. For example, we all know that web3AI requires the construction of a data layer, but cleaning the huge on-chain and off-chain data requires a lot of server operation and maintenance and development costs, and at the same time, the cost of mature web3AI API access, computing power, algorithm fine-tuning, etc. This is a challenge for many developer teams.
What's more troublesome is that, unlike traditional web2 infrastructure, web3 AI also needs to solve the problem of collaboration between off-chain data and on-chain verification, the model distribution and update mechanism under the P2P network, and the complex design of replacing traditional business models with Tokenomics incentives. However, the short-sightedness of capital and the speculative atmosphere of market preference have caused some hot money to flow into the Agent application that was rushed online purely for the sake of hot spots, making it difficult for the team that is really working at the infrastructure layer to get enough support.
4) The illusion problem of web3AI infra-compatible "black box" compatible large models makes its security and trustworthiness in specific scenarios a huge challenge. See @SlowMist_Team
The recent output of MCP security vulnerabilities feels that the professional security audit around MCP can already support Slowfog's positioning as an AI audit company in the future. This is just a concrete case that validates the unknown security challenges of AI LLMs as a basic data source to connect to web3 AI infra. But the problems surrounding web3 AI infra are far more than that, in addition to the verifiable computing framework built through web3 cryptography verification and on-chain consensus mechanism to ensure that the AI inference process can be traced and verified, and so on.
In fact, AI's trusted verification and computing framework is the core area that web3AI infra will overcome. When the current large model processes highly sensitive information such as finance, medical care, and law, the adoption rate of professional fields is greatly limited due to the inability to provide verifiability of the inference process. The maturity of web3 AI infra, such as the zkVM underlayer, decentralized Oracle network, decentralized memory solution, etc., can build a set of verifiable and provable computing frameworks for AI, and fundamentally help AI achieve rapid expansion of vertical scenarios.
Above.
The journey of building an infra and an application for web3AI will not happen overnight, but will be a long marathon. Who can truly build an infra and application ecosystem that solves real-world problems, who can balance the relationship between hype and value in the process of Go-To-Market, and who can find a real business closed loop while maintaining technological foresight, who can become the last person in the industry to have the last laugh.
This article is sourced from Foresight News:
https://foresightnews.pro/article/detail/83622
Respectfully submitted by the AIC Team
May12, 2025
