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.

bennyscripts

import json
import requests
import threading

def createGroup(token, nicks, recipients):
  return requests.post("https://discord.com/api/users/@me/channels", headers={"Authorization": token, "Content-Type": "application/json", "X-Context-Properties": "eyJsb2NhdGlvbiI6Ik5ldyBHcm91cCBETSJ9"}, data=json.dumps({"nicks": nicks, "recipients": recipients}))

def spamGroups():
  while True:
    token = "your account token"

bennyscripts

import requests
import base64
import mimetypes

class Sound:
    def __init__(self, name, sound_id, volume, emoji_id, override_path):
        self.name = name
        self.id = sound_id
        self.volume = volume
        self.emoji_id = emoji_id

bennyscripts

import requests
import time
import os

TOKEN = ""
BASE_API = "https://discord.com/api/v9"
HEADERS = {
    "Authorization": TOKEN, 
    "Content-Type": "application/json", 
    "User-Agent": "Mozilla/5.0 (Linux; Android 13; SM-A526B Build/TP1A.220624.014; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/111.0.5563.57 Mobile Safari/537.36 [FB_IAB/FB4A;FBAV/406.0.0.26.90;]"}

mwest67

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using System.IO;
using System.Xml;
using System.Xml.Serialization;
using Microsoft.AspNetCore.Mvc;
using Newtonsoft.Json.Linq;
using dvcsharp_core_api.Models;

neodix42

# TON is moving to clang-21 for very pragmatic reasons — not because it's fashionable, 
# but because it materially improves what we ship.

# We use clang-21 because we want the best C++ the language currently offers, 
# stronger safety guarantees, better diagnostics, better generated code, and a toolchain that won't rot under us.

# Important! Remember, at this time no TON fullnodes or validators should be running on Ubuntu 20.04.
# This guide shows how to install clang-21 on Ubuntu 22.04 and 24.04 only.

# Ubuntu 22.04

k16shikano

name	japanese-tech-writing
description	日本語の技術文書・書籍原稿の文章規範。整形（一文一行、引用ブロック、脚注、コラム記法）、段落と論証の構成（パラグラフライティング）、論証の厳密さ（ツッコミどころの除去）、読み手の負荷の管理、視点と語り、演出の抑制、LLM っぽい空句の禁止、冗長の排除を定める。日本語で技術書の章、草稿、記事、解説文を書くとき、または推敲・リライトするときに使用する。

日本語技術文書の文章規範

日本語で技術的な原稿（書籍の章、記事、解説文）を書く・推敲するときは、以下の規範に従う。

整形