Arian van Putten arianvp

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.

SSH Keys Best Practices

(VERY rought draft, still a work-in-progress)

SSH Key Seperation

Authentication and signing keys serve different purposes, especially on platforms like GitHub where code integrity and access control are crucial. Here's a comparison based on their functions and best practices:

Authentication Keys:

Purpose:

The idea is to have nginx installed and node installed. I will extend this gist to include how to install those as well, but at the moment, the following assumes you have nginx 0.7.62 and node 0.2.3 installed on a Linux distro (I used Ubuntu).

In a nutshell,

nginx is used to serve static files (css, js, images, etc.)
node serves all the "dynamic" stuff.

So for example, www.foo.com request comes and your css, js, and images get served thru nginx while everything else (the request for say index.html or "/") gets served through node.

nginx listens on port 80.

	CREATE OR REPLACE FUNCTION decode_cbor(cbor bytea)
	RETURNS TABLE (
	item integer,
	map_item_count integer,
	text_string text,
	bytes bytea
	)
	LANGUAGE sql
	AS $$
	/*

	{ stdenv, fetchgit, fetchurl, writeText, automake, autoconf, aria, coreutils, curl, cvs, gcc, git, python, which, bison, flex, gperf, help2man, libtool, ncurses, texinfo, wget, file }:

	stdenv.mkDerivation {
	name = "xtensa-esp32-elf";
	version = "1.22.x";
	src = fetchgit {
	url = "https://github.com/espressif/crosstool-NG.git";
	# branch = "xtensa-${version}";
	rev = "6c4433a51e4f2f2f9d9d4a13e75cd951acdfa80c";
	sha256 = "03qg9vb0mf10nfslggmb7lc426l0gxqhfyvbadh86x41n2j6ddg6";

	{ pkgs ? import <nixpkgs> {}}:
	let
	lib = pkgs.lib;
	makeImageFromDebDist =
	{ name, fullName, size ? 4096, urlPrefix
	, packagesList ? "", packagesLists ? [packagesList]
	, packages, extraPackages ? [], postInstall ? ""
	, extraDebs ? []
	, QEMU_OPTS ? "", memSize ? 512
	, createRootFS }:

	import urllib2
	import re
	import sys
	from collections import defaultdict
	from random import random

	"""
	PLEASE DO NOT RUN THIS QUOTED CODE FOR THE SAKE OF daemonology's SERVER, IT IS
	NOT MY SERVER AND I FEEL BAD FOR ABUSING IT. JUST GET THE RESULTS OF THE
	CRAWL HERE: http://pastebin.com/raw.php?i=nqpsnTtW AND SAVE THEM TO "archive.txt"

	Latency Comparison Numbers (~2012)
	----------------------------------
	L1 cache reference 0.5 ns
	Branch mispredict 5 ns
	L2 cache reference 7 ns 14x L1 cache
	Mutex lock/unlock 25 ns
	Main memory reference 100 ns 20x L2 cache, 200x L1 cache
	Compress 1K bytes with Zippy 3,000 ns 3 us
	Send 1K bytes over 1 Gbps network 10,000 ns 10 us
	Read 4K randomly from SSD* 150,000 ns 150 us ~1GB/sec SSD