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.

ljw1004

#!/bin/bash
set -euo pipefail
trap 'echo "at line $LINENO, exit code $? from $BASH_COMMAND" >&2; exit 1' ERR

# This is a Claude Code hook to stop it saying "you are right".
#
# Installation:
# 1. Save this script and chmod +x it to make it executable.
# 2. Within Claude Code, /hooks / UserPromptSubmit > Add a new hook (this file)
#

solarkraft

# This script automatically handles Syncthing conflicts on text files by applying a
# git three-way merge between the previously synced version and each divergent version.

# It depends on the watchdog package and git.

# For automatic dependency installation when running with ´uv run --script deconflicter.py´:
# /// script
# requires-python = ">=3.10"
# dependencies = [
#     "watchdog",

veekaybee

ChatGPT Resources

Context

ChatGPT appeared like an explosion on all my social media timelines in early December 2022. While I keep up with machine learning as an industry, I wasn't focused so much on this particular corner, and all the screenshots seemed like they came out of nowhere. What was this model? How did the chat prompting work? What was the context of OpenAI doing this work and collecting my prompts for training data?

I decided to do a quick investigation. Here's all the information I've found so far. I'm aggregating and synthesizing it as I go, so it's currently changing pretty frequently.

Model Architecture

zsteinkamp

#!/bin/bash
  
# Shell code to run in the container
cmd () {
  echo "Hello, ${OUTER_USER}. I am $(whoami)."
  cat /etc/issue
}

# Dockerfile definition
read -r -d '' DOCKERFILE <<EOD

sindresorhus

Pure ESM package

The package that linked you here is now pure ESM. It cannot be require()'d from CommonJS.

This means you have the following choices:

1. Use ESM yourself. (preferred)
   Use import foo from 'foo' instead of const foo = require('foo') to import the package. You also need to put "type": "module" in your package.json and more. Follow the below guide.
2. If the package is used in an async context, you could use await import(…) from CommonJS instead of require(…).
3. Stay on the existing version of the package until you can move to ESM.

rmcdongit

macOS 10.15 System Preference Panes

Below are a list of System Preference pane URLs and paths that can be accessed with scripting to assist users with enabling macOS security settings without having to walk them through launching System Preferences, finding panes, and scrolling to settings. Not all panes have an accessible anchor and some are OS specific.

To find the Pane ID of a specific pane, open the System Preferences app and select the desired Preference Pane. With the pane selected, open the ScriptEditor.app and run the following script to copy the current Pane ID to your clipboard and display any available anchors:

tell application "System Preferences"
	set CurrentPane to the id of the current pane
	set the clipboard to CurrentPane

bonniss

Github Search Cheat Sheet

GitHub’s search supports a variety of different operations. Here’s a quick cheat sheet for some of the common searches.

For more information, visit our search help section.

Basic search

Pyronerd62

<style lang="scss" scoped>
.tinyMce-container{
    margin-top: 3%;
    margin-bottom: 3%;
}
</style>
<template> 
    <div> 
        <pre v-html="description"></pre>  
        <editor v-model="description"></editor>   

jaredly

module StringContext =
  Context.MakePair({
    type t = string;
    let defaultValue = "Awesome";
  });

let component = ReasonReact.statelessComponent("Tree");

let make = _children => {