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.

xeoncross

import $ from "jquery";
import "bootstrap";
import "bootstrap/dist/css/bootstrap.css"; // Import precompiled Bootstrap css
import "@fortawesome/fontawesome-free/css/all.css";
import "./offcanvas.css";
import Holder from "holderjs";

$(() => {
  "use strict";

louisguitton

version: '3'

services:
  nginx:
    image: nginx:1.15-alpine
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./nginx:/etc/nginx/conf.d

asyd

from flask import Flask
from flask_sqlalchemy import SQLAlchemy

def create_app(config: str=None):
    app = Flask(__name__, instance_relative_config=True)
    if config is None:
        app.config.from_pyfile('dev.py')
    else:
        logger.debug('Using %s as configuration', config)
        app.config.from_pyfile(config)

crittermike

# One liner
wget --recursive --page-requisites --adjust-extension --span-hosts --convert-links --restrict-file-names=windows --domains yoursite.com --no-parent yoursite.com

# Explained
wget \
     --recursive \ # Download the whole site.
     --page-requisites \ # Get all assets/elements (CSS/JS/images).
     --adjust-extension \ # Save files with .html on the end.
     --span-hosts \ # Include necessary assets from offsite as well.
     --convert-links \ # Update links to still work in the static version.

ziadoz

#!/usr/bin/env bash
# https://developers.supportbee.com/blog/setting-up-cucumber-to-run-with-Chrome-on-Linux/
# https://gist.github.com/curtismcmullan/7be1a8c1c841a9d8db2c
# https://stackoverflow.com/questions/10792403/how-do-i-get-chrome-working-with-selenium-using-php-webdriver
# https://stackoverflow.com/questions/26133486/how-to-specify-binary-path-for-remote-chromedriver-in-codeception
# https://stackoverflow.com/questions/40262682/how-to-run-selenium-3-x-with-chrome-driver-through-terminal
# https://askubuntu.com/questions/760085/how-do-you-install-google-chrome-on-ubuntu-16-04

# Versions
CHROME_DRIVER_VERSION=`curl -sS https://chromedriver.storage.googleapis.com/LATEST_RELEASE`

stephenhardy

-- Remove the history from 
rm -rf .git

-- recreate the repos from the current content only
git init
git add .
git commit -m "Initial commit"

-- push to the github remote repos ensuring you overwrite history
git remote add origin git@github.com:<YOUR ACCOUNT>/<YOUR REPOS>.git