Why I believe AI Agents are more worth paying attention to than large models

Large models are like brains; AI Agents are more like "people who can get things done." Being smart isn't enough; execution is what adds value. Many companies have already realized that no matter how powerful the model capabilities are, if they can't be integrated into business processes, their commercial value is limited.
The significance of AI Agents lies in embedding intelligence into workflows. They can automatically read information, trigger actions, and generate results. In other words, they elevate AI from "answering questions" to "completing tasks."
I am optimistic about this, and there's a practical reason: the increasing complexity of enterprise digitalization makes it difficult for humans to manage all systems. Agents can serve as a coordination layer, helping humans manage the digital world.
In the short term, they are more like assistants; in the long term, they could become infrastructure. Just as operating systems are for computers, Agents might become the scheduling hub of the digital world.
Of course, there is some hype involved now, and many projects are just repackaging concepts. But those with genuine product strength will gradually emerge.
One key point: large models determine the upper limit, but Agents determine implementation.
Interaction: Do you think AI should be more of an advisor or an executor?#我看好的AIAgent