Agentic AI

通过模拟购车场景实测OpenAI Deep Research，展示其如何大幅提升调研效率、解放脑力进行更深层思考，以及重新定义问题边界的认知炼金能力。

A hands-on exploration of OpenAI Deep Research through a simulated car-buying scenario, demonstrating how AI elevates research efficiency and enables deeper strategic thinking about negotiation and decision-making.

不要从AutoGen、LangGraph等框架入手学习Agentic AI——它们会锁定你的世界观。在领域快速演变期，应该从第一性原理出发自己构建。

Don't start Agentic AI with frameworks like AutoGen or LangGraph—they lock you into a worldview. Build from first principles while the field is still evolving rapidly.

通过分离Planner和Executor、强制文档沟通、用o1当Planner三项改造，解决Cursor鬼打墙问题，实现质量显著提升的多智能体系统。

How separating Planner and Executor roles, enforcing document-based communication, and using o1 as Planner transformed Cursor from a simple assistant into a multi-agent system.

Agentic AI为什么常常只能做到七八成：自我迭代的反馈环被打破，缺乏感知能力和主观评价标准，以及如何补足这些能力缺口。

Why Agentic AI often achieves only 70-80% completion: broken self-iteration loops due to missing perception and subjective standards, and how to fix them.

在做完上篇文章介绍的Cursor魔改之后，将Agentic AI用"我说你做"的模式进行使用，已经成为了我对AI利用的主要方式。而这又进一步启发我：有没有可能把Cursor或者类似的编辑器作为使用AI的通用入口呢？

After completing the previous article introducing the modifications to Cursor, using Agentic AI in a "I say, you do" mode has become my primary method for leveraging AI. This further inspired me: is it possible to use Cursor or a similar editor as a universal entry point for AI?