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.

SamuelChojnacki

## Production : 

# Étape 1 : Construction
FROM node:20.11.1-alpine3.19 AS builder

# Définir le répertoire de travail
WORKDIR /opt/app

# Copier les fichiers nécessaires pour installer les dépendances
COPY package.json yarn.lock ./

hmbashar

<?php



//save theme active time
function my_theme_activation_init() {
  // Check if already saved the activation date & time
  // to prevent over-writing if user deactive & active theme
  // multiple time
	if(!get_option('mytheme_activation_time', false)){

GhazanfarMir

########## Install NGINX ##############

# Install software-properties-common package to give us add-apt-repository package
sudo apt-get install -y software-properties-common

# Install latest nginx version from community maintained ppa
sudo add-apt-repository ppa:nginx/stable

# Update packages after adding ppa

hasinhayder

<?php
function init_url_rewrite_rule(){
    add_rewrite_endpoint( 'edit',EP_PERMALINK | EP_PAGES | EP_ATTACHMENT );
    if(get_option("EDIT_REWRITE_RULE")!=1){
        flush_rewrite_rules();
        update_option("EDIT_REWRITE_RULE",1);
    }
}

function redirect_edit_url(){

sarim

#!/bin/sh

function print_bar {
    GDONE=$1
    GPROG='['
    for i in $(seq 1 1 $GDONE)
    do
       GPROG=$GPROG'#'
    done

paulund

<?php
add_filter( 'preprocess_comment', 'minimal_comment_length' );

function minimal_comment_length( $commentdata ) {
    $minimalCommentLength = 20;

	if ( strlen( trim( $commentdata['comment_content'] ) ) < $minimalCommentLength ) 
        {
		wp_die( 'All comments must be at least ' . $minimalCommentLength . ' characters long.' );
        }

zenorocha

(function() {
    
    var script,
        scripts = document.getElementsByTagName('script')[0];
        
    function load(url) {
      script = document.createElement('script');
      script.async = true;
      script.src = url;
      scripts.parentNode.insertBefore(script, scripts);