karpathy

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

candideu

Here are some quick tips on how to transfer your Spotify playlists to your Navidrome instance using tunesynctool, assuming you're using Windows:

Download Python

Download Python to your computer: https://www.python.org/downloads/

When installing, make sure to add Python to PATH:

flaksp

BitWarden to Apple iCloud Keychain passwords converter

This Python scripts allows you to move your passwords from BitWarden to Apple iCloud.

You need to know:

• It ignores secure notes, credit cards and other types that are not passwords.
• It ignores BitWarden entries without usernames, passwords and URLs.
• It also ignores URLs that do not start with http:// or https://.
• It normalizes all TOTP tokens, e.g. wskg vtqa h5kl bhb4 v4v2 ybyo woc6 qme2 will be converted to otpauth://totp/example.com:dude@foo.bar?secret=WSKGVTQAH5KLBHB4V4V2YBYOWOC6QME2&issuer=example.com&algorithm=SHA1&digits=6&period=30.

kevin-smets

Default

Powerlevel10k