Investment researcher · 用 AI 重新设计研究工作流

对冲基金研究员，在做两件事 —— AI 辅助投研 和 个人知识库维护。

在我看来这其实是同一件事：投研本来就是知识工作，追踪公司、假设、行业脉络，和追踪文章、概念、自己的认知迭代，需要的是同一套脚手架。两条线共享同一条原则：零代码，让完全没有编程基础的人也能用 AI 帮自己成长。我自己就是零代码出身，做出来的东西默认是给"和我一样不会写代码"的人用的。

I'm an investment researcher at a hedge fund, working on what looks like two projects but is really one: AI-augmented investment research, and personal knowledge curation. Investment research is knowledge work — tracking companies, hypotheses, and industry threads needs the same scaffolding as tracking articles, concepts, and your own evolving thinking. Built on one principle: zero-code, so people without any coding background can use them too. I have zero coding background myself — every tool I build is designed for people who, like me, can't write a line of Python.

karpathy-claude-wiki — A Karpathy-style personal wiki template for LLMs. Markdown + frontmatter, no vector DB, no RAG. Built for Claude Code, works with any file-editing AI agent.

灵感来自 Karpathy 那条推 —— 让 LLM 维护个人 wiki。零代码、纯 markdown、刻意不用向量库。门槛降到"会双击 PowerShell"。

This is the first one I've open-sourced — more coming.

这是开源出来的第一个，后面还有。

• 让 0 代码基础的人也能搭建和维护自己的 AI 工作流（投研 / 知识库 / 写作）
• AI 在投资研究里真正能加速的环节，vs 只是看起来能加速的环节
• 个人知识库写给 LLM 看，和写给未来的自己看，哪些约定要重新定

_{🐱 Profile picture is my cat 豆包. He's the chief approver. / 头像是我家猫豆包，首席审稿人。}