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.

turnercore

import argparse
import copy
import csv
import logging
import re
import shutil
import zipfile
from datetime import datetime
from pathlib import Path
import xml.etree.ElementTree as ET

turnercore

(() => {
  let isVideoHandled = false; // Flag to indicate if the video has been handled

  const checkVideoStart = () => {
    const video = document.querySelector('#movie_player > div.html5-video-container > video');
    if (!video) return;

    // When the video starts playing (currentTime changes from 0), pause it
    if (video.currentTime > 0 && !video.paused) {
      video.pause();

turnercore

#!/bin/bash

# Create directories and files
mkdir -p ./components/ui
mkdir -p ./lib
touch ./lib/utils.ts

# Prompt to move globals.css
echo "Would you like to move globals.css to styles/globals.css? (y/n)"
read -r -p "[Y/n]: " move_response