LLM Wiki
AI Summary
Karpathy's April 2026 gist that went viral in 48 hours. The reframe: agent memory isn't a retrieval problem, it's a maintenance problem, and an LLM-maintained markdown wiki is a better substrate than a vector store because it cross-references, deduplicates, and compounds as it grows. His line — 'Obsidian is the IDE, the LLM is the programmer, the wiki is the codebase' — is the cleanest analogy anyone has written for what personal agent memory should look like. Short, opinionated, already spawning implementations.
Original excerpt
You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Star 5,000+(5,000+)You must be signed in to star a gist Fork 5,000+(5,000+)You must be signed in to fork a gist
* Embed Embed this gist in your website. * Share Copy sharable link for this gist. * Clone via HTTPS Clone using the web URL.
Clone this repository at <script src="https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f.js"></script>
Save karpathy/442a6bf555914893e9891c11519de94f to your computer and use it in GitHub Desktop.
Embed Embed this…
18 more articles in this vault.
Import the full Agent Memory Patterns vault to Burn 451 and build your own knowledge base.
Content attributed to the original author (Andrej Karpathy). Burn 451 curates publicly available writing as a reading index. For removal requests, contact @hawking520.