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.

OmerFarukOruc

AI Coding Agent Guidelines (claude.md)

These rules define how an AI coding agent should plan, execute, verify, communicate, and recover when working in a real codebase. Optimize for correctness, minimalism, and developer experience.

---

Operating Principles (Non-Negotiable)

• Correctness over cleverness: Prefer boring, readable solutions that are easy to maintain.
• Smallest change that works: Minimize blast radius; don't refactor adjacent code unless it meaningfully reduces risk or complexity.

J-Moravec

error404 = paste0(
    "<!DOCTYPE html>",
    "<html lang=\"en\">",
    "<head>",
    "    <meta charset=\"UTF-8\">",
    "    <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">",
    "    <title>Resources not found</title>",
    "</head>",
    "<body>",
    "    <div class=\"main\">",

kmboehm

import pandas as pd
import torch
import h5py
from openslide import OpenSlide
import os
import numpy as np
from joblib import Parallel, delayed
from features import extract_features_single_tile
import json
from scipy.stats import mannwhitneyu

hrbrmstr

#!/usr/bin/env Rscript

library(httr, include.only = c("GET", "add_headers", "content"))
library(jsonlite, include.only = c("fromJSON", "write_json"))
library(glue, include.only = c("glue"))

GITHUB_TOKEN <- Sys.getenv("GITHUB_TOKEN")

JSON_FILE <- "webr-experiments.json"
HTML_FILE <- "webr-experiments/index.html"

AlbertRapp

library(dplyr)
library(ggplot2)
library(patchwork)
library(ggiraph)

dat <- gapminder::gapminder |> 
  janitor::clean_names() |> 
  mutate(
    # ID that is shared for boxplots (this one uses factors, i.e. numbers, as ID instead of continents)

krlmlr

`<-` <- function(lhs, rhs, envir = parent.frame()) {
  assign(as.character(substitute(lhs)), rhs, envir)
  lockBinding(substitute(lhs), envir)
  invisible(rhs)
}

JarrydLong

<?php
/**
 * This recipe will redirect you to the levels page after registration
 * 
 * You can add this recipe to your site by creating a custom plugin
 * or using the Code Snippets plugin available for free in the WordPress repository.
 * Read this companion article for step-by-step directions on either method.
 * https://www.paidmembershipspro.com/create-a-plugin-for-pmpro-customizations/
 */
function my_pmpro_redirect_after_register( $user_id ){

domenicrosati

z3tt

#### 1. Sign up at GitHub.com ################################################

## If you do not have a GitHub account, sign up here:
## https://github.com/join

# ----------------------------------------------------------------------------

#### 2. Install git ##########################################################

## If you do not have git installed, please do so: