Manny Colon mannycolon

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.

Update package.json, set version to a prerelease version, e.g. 2.0.0-rc1, 3.1.5-rc4, ...
Run npm pack to create package
Run npm publish <package>.tgz --tag next to publish the package under the next tag
Run npm install --save package@next to install prerelease package

To remove a submodule you need to:

Delete the relevant section from the .gitmodules file.
Stage the .gitmodules changes git add .gitmodules
Delete the relevant section from .git/config.
Run git rm --cached path_to_submodule (no trailing slash).
Run rm -rf .git/modules/path_to_submodule (no trailing slash).
Commit git commit -m "Removed submodule "
Delete the now untracked submodule files rm -rf path_to_submodule

##Ubuntu Server

NodeJS

sudo apt-get install g++ curl libssl-dev apache2-utils git-core make
cd /usr/local/src
wget http://nodejs.org/dist/v0.10.28/node-v0.10.28.tar.gz
tar -xvzf node-v0.10.28.tar.gz
cd node-v0.10.28

Pointing a GoDaddy domain to a Heroku app

https://devcenter.heroku.com/articles/custom-domains http://thenomadicfreelancer.blogspot.com/2012/08/pointing-godaddy-domain-to-your-heroku.html

1) Custom Subdomains

For each custom subdomain use domains:add in the Terminal.

	// If the user has been on the page for over 12 hours, the next link
	// click will do a full page transition to get new code
	import React from "react";
	import { BrowserRouter, useLocation } from "react-router-dom";

	let hour = 3600000;

	export default function StaleAppRouter(props) {
	let [isStale, setIsStale] = useState(false);

	import { NextApiResponse, NextApiRequest } from "next";
	const twillio = require('twilio');
	const authToken = process.env.TWILIO_AUTH_TOKEN;
	const accountSid = process.env.TWILIO_ACCOUNT_SID;
	const myCellPhoneNumber = process.env.MY_NUMBER;
	const superSecretAPIKey = 'ABCD1234';

	export default async (req: NextApiRequest, res: NextApiResponse) => {
	const client = twillio(accountSid, authToken);
	const auth = req.headers.authorization;


	function Timer() {
	const startTime = React.useRef(performance.now());
	const [time, setTime] = React.useState(performance.now());
	React.useEffect(() => {
	const id = setTimeout(() => {
	ReactDOM.flushSync(() => {
	setTime(performance.now());
	});
	}, 2);

	/**
	* A utility for reading nested objects without pulling your hair out.
	*
	* Supported objects: List, Map, JSONObject, JSONArray
	*
	*/
	public class ObjectReader implements Iterable<ObjectReader> {
	private final Object map;

	/**

	// https://developer.mozilla.org/en-US/docs/Web/API/window.setTimeout

	var timeoutID;
	delayedAlert();

	function delayedAlert() {
	timeoutID = window.setTimeout(slowAlert, 2000);
	}

	function slowAlert() {