2025-09-03 22:03:56
本文作者通过构建个人投资系统的实践，深入对比了2025年的三款主流AI编程工具：Cursor、Augment 和 Claude Code。
作者将这三者生动地比作“没有感情的代码修改机器”（Cursor）、“优秀的实习生”（Augment）和“可以托付任务的初级工程师”（Claude Code），从用户体验、上下文工程能力和工作流整合等角度，剖析了它们各自的设计哲学与优劣。
文章的核心观点是，当前AI编程工具的竞争已超越了单纯的模型能力，关键在于如何将智能Agent无缝融入开发者的真实工作流中，从而重塑代码编写的范式。
点击阅读原文
2025-09-03 22:03:56
This article provides an in-depth comparison of three leading AI programming tools in 2025—Cursor, Augment, and Claude Code—based on the author practical experience building a personal investment system.
The author vividly portrays them as an "emotionless code modification machine" (Cursor), an "excellent intern" (Augment), and a "junior engineer you can entrust with tasks" (Claude Code), analyzing their respective design philosophies from the perspectives of user experience, context engineering, and workflow integration.
The core thesis is that the competition among AI programming tools has moved beyond the raw power of the underlying models. The focus now is on how effectively these tools can seamlessly integrate intelligent agents into a developer real-world workflow, ultimately reshaping the paradigm of how we write code.
点击阅读原文
2025-08-25 23:17:14
这篇文章讲述了作者受市场焦虑和个人投资需求驱动，利用 Gemini 等 AI 工具学习投资知识，并指挥编码助手开发了一款个人投资顾问 AI 的经历。
作者通过这个过程，从一个最初因数据获取困难而放弃项目的人，转变为一个不写一行代码就能构建系统的“超级个体”，并分享了在AI辅助开发中关于上下文、模型控制和规范重要性的感悟。
点击阅读原文
2024-04-21 13:25:23
详细介绍 ChatGPT 的训练过程以及 Transformer 的输入输出过程和其核心 attention 的原理。让你对 ChatGPT 以及其架构原理有更深层次的认知。
点击阅读原文
2024-04-02 23:04:23
分享自己对 AI 的学习和理解过程。详细阐述了 llm 神经网络的工作原理，包括神经元如何接收输入、计算输出以及权重如何影响网络的学习过程。通过这些基础知识，作者旨在帮助读者更好地理解AI Agent的开发和大模型的工作机制。
点击阅读原文
2022-12-13 17:01:23
Before I became a Tech Lead, I had already seen a number of great Tech Leads in action and had some ideas about how I would do the job, deal with customers, plan for and solve problems, and so on.
I have slowly learned what makes a good tech lead, so I am writing this blog now while I am still motivated, so I do not get lost later.
点击阅读原文
