Oleg Korol olegkorol

LLM Wiki v2

A pattern for building personal knowledge bases using LLMs. Extended with lessons from building agentmemory, a persistent memory engine for AI coding agents.

This builds on Andrej Karpathy's original LLM Wiki idea file. Everything in the original still applies. This document adds what we learned running the pattern in production: what breaks at scale, what's missing, and what separates a wiki that stays useful from one that rots.

What the original gets right

The core insight is correct: stop re-deriving, start compiling. RAG retrieves and forgets. A wiki accumulates and compounds. The three-layer architecture (raw sources, wiki, schema) works. The operations (ingest, query, lint) cover the basics. If you haven't read the original, start there.

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.

Sometimes, companies are worried about "reusing personal GitHub accounts", but contributing source code or reviews with an already existing account does NOT mean that

existing ssh keys or personal access tokens would be valid to access your enterprise organization (all credentials would have to be authorized: https://help.github.com/en/github/authenticating-to-github/authorizing-an-ssh-key-for-use-with-saml-single-sign-on)
email notifications would go to any personal email (you can enforce the use of a company email domain: https://help.github.com/en/github/setting-up-and-managing-organizations-and-teams/restricting-email-notifications-to-an-approved-domain)
contributions could not be marked to belong to your organization explicitly (you can explicitly mark contributions to be done on behalf of your company: https://help.github.com/en/github/committing-changes-to-your-project/creating-a-commit-on-behalf-of-an-organization)
you could not enforce any merged commit to satisfy your coding standards an

FWIW: I (@rondy) am not the creator of the content shared here, which is an excerpt from Edmond Lau's book. I simply copied and pasted it from another location and saved it as a personal note, before it gained popularity on news.ycombinator.com. Unfortunately, I cannot recall the exact origin of the original source, nor was I able to find the author's name, so I am can't provide the appropriate credits.

Effective Engineer - Notes

By Edmond Lau
Highly Recommended 👍
http://www.theeffectiveengineer.com/

What's an Effective Engineer?

A curated list of AWS resources to prepare for the AWS Certifications

A curated list of awesome AWS resources you need to prepare for the all 5 AWS Certifications. This gist will include: open source repos, blogs & blogposts, ebooks, PDF, whitepapers, video courses, free lecture, slides, sample test and many other resources.

OS X Screencast to animated GIF

This gist shows how to create a GIF screencast using only free OS X tools: QuickTime, ffmpeg, and gifsicle.

Instructions

To capture the video (filesize: 19MB), using the free "QuickTime Player" application:

	#!/usr/bin/env bash

	set -euo pipefail

	DESCRIPTION="Convert an input file of {csv, parquet, json} format to {csv, parquet, json} format"
	DUCKDB_VERSION="1.1.0"
	VERBOSE=false

	usage() {
	echo "Usage: $0 <input_file> -f <output_format> [-o <output_file>] [-v]"

	{
	"firestore": {
	"rules": "firestore.rules",
	"indexes": "firestore.indexes.json"
	},
	"emulators": {
	"firestore": {
	"port": 8080
	}
	}

	function checkForSWs() {
	console.log('I will look for registered SW\'s now...')
	navigator.serviceWorker.getRegistrations()
	.then(sws => {
	if (sws[0]) {
	console.log('Found a registered SW:', sws[0], '\n Will proceed to unregister...');
	sws[0].unregister();
	console.log('Done.')
	} else {
	console.log('No registered SW\'s were found.')

	// just an example. A cleaner way is to wrap the showLine-stuff in a dedicated component
	<template>
	<div>
	<my-line v-if="showLine" :data="lineData" :options="options">
	</div>
	</template>

	<script>
	export default {
	data () {