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.

kyrylo

# frozen_string_literal: true
# config/initializers/colorized_logger.rb

# This initializer adds color to the Rails logger output. It's a nice way to
# visually distinguish log levels.
module ColorizedLogger
  COLOR_CODES = {
    debug:   "\e[36m",  # Cyan
    info:    "\e[32m",  # Green
    warn:    "\e[33m",  # Yellow

keijiro

using UnityEditor;
using UnityEngine;
using UnityEngine.LowLevel;
using System.Linq;
using System.Threading;

namespace EditorUtils {

//
// Serializable settings

yasirkula

using UnityEditor;
using UnityEngine;

public class SnapRectTransformAnchorsToCorners : MonoBehaviour
{
    [MenuItem("CONTEXT/RectTransform/Snap Anchors To Corners", priority = 50)]
    private static void Execute(MenuCommand command)
    {
        RectTransform rectTransform = command.context as RectTransform;
        RectTransform parent = rectTransform.parent as RectTransform;

n-studio

Notes

This guide uses Kamal 2.5.3

Motivation

Kamal was designed with 1 service = 1 droplet/VPS in mind.
But I'm cheap and I want to be able to deploy multiple demo/poc apps apps on my $20/month dedicated server.
What the hell, I'll even host my private container registry on it.

rain-1

Purpose

Bootstrap knowledge of LLMs ASAP. With a bias/focus to GPT.

Avoid being a link dump. Try to provide only valuable well tuned information.

Prelude

Neural network links before starting with transformers.

vpfrimmer

using System.IO;
using UnityEditor;
using UnityEditor.Android;
using UnityEngine;

/// <summary>
/// Used as a workaround for Unity 2022.x who isn't able to set the installlocation value to "auto" in the AndroidManifest.xml file.
/// Should be put in the Assets/Editor folder, otherwise you'll get errors on build.
/// The issue can be followed here : https://issuetracker.unity3d.com/issues/android-install-location-changes-when-exporting-project
/// </summary>

khskekec

LibreLinkUp HTTP-Dump

This document describes the HTTP communication of LibreLinkUp which functions as follower app to receive cgm data. Some data in the responses were masked.

Context

This dump was created on an android device with LibreLinkUp app. Capturing was done with HttpToolkit over adb.

API Url

bompus

// silence dart-sass / npm sass DEPRECATION WARNING: Using / for division outside of calc() is deprecated and will be removed in Dart Sass 2.0.0.

const silenceSomeSassDeprecationWarnings = {
  verbose: true,
  logger: {
    warn(message, options) {
      const { stderr } = process;
      const span = options.span ?? undefined;
      const stack = (options.stack === 'null' ? undefined : options.stack) ?? undefined;

StagPoint

using System;
using System.Runtime.InteropServices;
using System.Runtime.CompilerServices;

using Unity.Burst;
using Unity.Mathematics;

public static class CameraMath
{
	/// <summary>