Recently, there have been several news items worth noting:

- With the "help" of teammate Amazon, Anthropic's Fable 5 and Mythos have been regulated by the U.S. government, leading to the immediate offline status of these two models.

- Anthropic's Dario Amodei proposed strict regulation and prevention of AI in certain countries at the G7 meeting, receiving agreement from some countries like Canada.

- Anthropic and OpenAI will soon require individual users to provide stricter identity information.

Seeing these news items, I was not very surprised; I just felt that this decoupling has come much faster than expected.

From Anthropic's business model, its main users and major source of profit come from corporate users.

These severe verification measures are unlikely to affect corporate users much and mainly impact individual users. Therefore, these measures will not significantly affect its revenue.

However, for OpenAI, which primarily targets individual consumers, adopting similar measures indicates:

The top closed-source large model companies in the industry will be proactively or reactively subject to strict regulation in the future, regardless of ideological or national security considerations.

This regulation currently seems to be happening only in the United States, but it will inevitably occur in all countries and regions deploying sovereign large models.

Under this regulation, the most commonly affected will be individual users and small to medium-sized startups (like current one-person companies).

Although individual users and small businesses do not use AI as frequently as corporate users, and contribute less to the revenue of closed-source large model companies, this group is still significantly large on an absolute scale worldwide, and the market size is not small.

Moreover, this strict regulation and limitation will produce another problem. This problem may not be obvious now but will become more apparent in the future:

This regulation and limitation strictly confine the use of closed-source large models to the human user group. I believe that as we move into the future, AI will likely become the true major user of AI.

In the future, for ordinary people, most interactions with AI will likely involve humans sending language instructions to their (butler) AI Agents, which will then call upon different specialized AI tools to complete the instructions. There may even be multiple dedicated AI Agents under one AI Agent, with each dedicated AI Agent responsible for handling specialized tasks.

For these AI Agents, when calling AI tools, they cannot verify identity through facial or video recognition like humans.

Furthermore, in the future, the number of AI Agents is likely to greatly exceed the number of human users.

Therefore, from the perspective of individual users, small to medium-sized enterprises, and AI Agents, the currently leading closed-source large models may not be ideal tools for high-frequency use in the future, or they may not be the ideal tools for handling everyday tedious tasks.

What is likely to adapt to and meet the high frequency and urgent needs of this portion of users includes decentralized AI based on crypto asset payments.

Payments based on crypto assets eliminate the potential obstacles that AI Agents may encounter in the payment process.

Decentralized AI resolves differences in inference efficiency that may arise due to GPU bans; it can also effectively avoid regulatory barriers through open-source models; and it can switch nodes as needed, selecting stable service providers.

Is it even possible that, in the future, large language models will run directly on virtual machines at blockchain nodes, completely resolving the issue of model regulation?

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。