@axichuhai: OpenAI co-founder Karpathy’s 400 000-word, hundred-article knowledge base stays pristine—he never tidies a single line.

X AI KOLs Timeline Tools

Summary

Andrej Karpathy just open-sourced the personal knowledge-management system that keeps his 400 000-word archive organized without any manual curation.

OpenAI co-founder Andrej Karpathy stores 400 000 words and hundreds of articles in his knowledge base, yet he never organizes a single one himself. A few days ago he published the workflow on GitHub; one tweet rocketed to 20 million views. Most people hoard hundreds of bookmarks and forget where they put them two weeks later. The more you save, the faster you forget—Karpathy’s new repo shows how he escapes that trap.
Original Article Export to Word Export to PDF
View Cached Full Text

Cached at: 04/21/26, 09:29 AM

OpenAI co-founder Karpathy’s personal knowledge base holds 400,000 words and hundreds of articles—yet he has never curated a single line himself.
A few days ago he open-sourced his method on GitHub; one tweet rocketed to 20 M impressions.

Most people have hundreds or thousands of bookmarks, but two weeks later can’t remember where they stashed them.
The more you save, the faster you forget—an affliction Karpathy…

Similar Articles

@MindOS_Lisa: https://x.com/MindOS_Lisa/status/2052766937931965

X AI KOLs Timeline

The author introduces a personal knowledge management system built on Karpathy's knowledge base logic, combining Obsidian, Claude Code, and LLM Wiki to achieve a complete workflow covering automatic content ingestion, card generation, and article output. The article distinguishes between three levels — content, information, and knowledge — and provides concrete setup steps.

@TechFlow99: BREAKING: Someone just built the exact tool Andrej Karpathy said someone should build. 48 hours after Karpathy posted h…

X AI KOLs Timeline

A new open-source tool called Graphify was built within 48 hours of Andrej Karpathy describing an LLM knowledge base workflow, enabling users to generate navigable knowledge graphs, Obsidian vaults, and wikis from any folder with 71.5x fewer tokens per query compared to reading raw files. It integrates with Claude Code and supports 13 programming languages, PDFs, images, and Markdown.