Karpathy proposes replacing on-the-fly retrieval with an LLM-maintained markdown wiki that ingests sources, compiles structured pages, and self-updates through ingest, query, and lint cycles. This approach builds a persistent, compounding knowledge base without vectors or re-retrieval, though it currently lacks enterprise controls.

• Karpathy's "LLM Wiki" replaces vector-based RAG with an LLM that reads sources once and maintains a persistent, self-updating markdown wiki instead of re-retrieving chunked fragments per query.
• The system runs three operations—ingest (add new material), query (answer from the wiki, then save new answers back into it), and lint (scheduled scans for outdated info, broken links, and gaps)—so every interaction compounds the knowledge base rather than starting fresh.
• It directly targets two RAG failure modes: stateless retrieval that repeats embedding work every query, and chunking that destroys document context.
• The idea echoes Vannevar Bush's 1945 Memex concept of an associative, hyperlinked personal knowledge library, though it currently lacks enterprise-grade controls.

This guide walks you through creating a personal wiki using just three folders and text files. You automate content collection with a CLI scraper, write a simple schema file, then let an AI compile, update, and query your organized knowledge base.

• Three flat folders (raw/, wiki/, outputs/) plus a CLAUDE.md/AGENTS.md schema file replace databases and apps like Obsidian for personal knowledge management.
• agent-browser (a free Vercel Labs CLI tool) automates scraping JS-heavy or login-gated pages straight into the raw/ folder.
• Claude Code reads the schema and raw files to auto-generate an INDEX.md and per-topic markdown pages in wiki/, linking sources and summaries.
• A monthly AI-driven health check catches contradictions and unbacked claims before compounding errors degrade the wiki.