The article examines the LLM Wiki concept introduced by Karpathy, explaining how a Markdown‑based wiki maintained outside the LLM context can persist and evolve model understanding, compares it with RAG, note‑taking tools and traditional knowledge bases, and outlines architectural components, risks, and practical guidelines.
Traditional RAG pipelines forget everything after each query, but the LLM Wiki mode proposed by Andrej Karpathy compiles source material into a version‑controlled, cross‑referenced Markdown wiki, enabling knowledge to compound over time, reduce query costs, and provide a transparent, human‑readable knowledge base for AI engineers.
This article explains how Andrej Karpathy's LLM‑plus‑Obsidian workflow transforms large language models into continuous knowledge engineers, detailing a three‑layer architecture, core operations, practical setup steps, and open‑source tools that enable a self‑maintaining, compounding personal wiki.
This article analyzes why building a layered knowledge wiki tied to workspace and Git submodules dramatically reduces context entropy for AI coding, outlines the five knowledge categories, progressive disclosure design, multi‑agent initialization workflow, and the measurable productivity gains and governance benefits achieved in practice.
Karpathy outlines a detailed workflow for using large language models to automatically collect, organize, and continuously enrich personal research materials into an interlinked Markdown wiki, highlighting tools, architecture, and future directions for a self‑improving AI‑powered second brain.
Karpathy demonstrates a three‑layer LLM‑driven Wiki that ingests raw papers, code and datasets, automatically maintains structured markdown, and continuously improves through ingest, query and lint cycles, offering a compounding knowledge base that differs fundamentally from traditional RAG retrieval.
The article examines how unrestricted wiki editing in companies leads to duplicated, outdated, and low‑quality documentation, contrasts this with the disciplined PR‑based workflow of open‑source projects like Python, and argues that effective documentation governance depends on clear processes and cultural support rather than just tools.
