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.

jmanhype

Agent Learning via Early Experience + ACE Integration

Production-Ready Framework for Continuous Agent Learning

A complete implementation of reward-free reinforcement learning through world modeling, exploration, and self-reflection, with full ACE (Adaptive Context Engineering) integration for knowledge curation and semantic deduplication.

---

What Is This?

marvinhagemeister

import { ServerResponse, type IncomingMessage } from "node:http";
import { Http2ServerRequest, Http2ServerResponse } from "node:http2";
import { isArrayBufferView } from "node:util/types";

const INTERNAL_BODY = Symbol("internal_body");

const GlobalResponse = Response;
globalThis.Response = class Response extends GlobalResponse {
  [INTERNAL_BODY]: BodyInit | null | undefined = null;

LukasKriesch

import regex as re
import requests

MAX_HEADING_LENGTH = 7
MAX_HEADING_CONTENT_LENGTH = 200
MAX_HEADING_UNDERLINE_LENGTH = 200
MAX_HTML_HEADING_ATTRIBUTES_LENGTH = 100
MAX_LIST_ITEM_LENGTH = 200
MAX_NESTED_LIST_ITEMS = 6
MAX_LIST_INDENT_SPACES = 7

m5r

import notifierQueue from "~/queues/notifier.server.ts";

export const loader = async () => {
    await notifierQueue.add("test", { emailAddress: "mokhtar@remixtape.dev" });

    return null;
};

pkcpkc

/*
Collect headers (h1, h2, ..., hx) from html and generates navigatable, stylable table of contents.
 
@param maxHeaderLevel int Define header level depth, defaults to 3.
@param styleItem function Function that accepts text:string, level:int and itemAnchor:string to style toc entry, default renderer is set already (check source for usage).
@return string HTML table of contents
*/
function generateTableOfContents(maxHeaderLevel = 3, styleItem = function (text, level, itemAnchor) {
    var spaces = " ".repeat(Math.max(0, (level - 1)) * 3);
    var tocEntry = spaces + '<a href="#' + itemAnchor + '">' + text + '</a><br/>';

vojtajina

// simple dsl just wrapping angular's dsl, just providing higher abstraction
angular.scenario.dsl('submitMessage', function() {
  return function(message) {
    // these dsl already register futures (add fn into the queue),
    // so you don't wrap them into addFutureAction
    input('modelValue').enter(message);
    element('button.submit').click();
  };
});