Description
转自 https://www.youtube.com/watch?v=OSZdFnQmgRw
Andrej Karpathy 分享了他的个人 Obsidian 知识库系统，实现了无向量数据库、无嵌入的轻量级 RAG 工作流。通过巧妙的文件系统结构，让大语言模型直接处理大规模文档。本视频演示如何搭建该系统，并将其与 Claude Code 结合，打造适合独立开发者和小团队的强力知识管理工作流。
Andrej Karpathy shared his personal Obsidian-based knowledge system that acts as a RAG pipeline with no vector database, no embeddings, and no complex retrieval setup. Using a clever file structure and Obsidian, it lets LLMs navigate large document collections efficiently. This video walks through the full setup and pairs it with Claude Code as a powerful workflow for solo operators and small teams.