grahamr

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.

General links mentioned in the talk:

Der Flounder
Virtually Ghetto
Installing OSX in Fusion
Create VMware OSX Install DMG
VMware Volume Deployment Package
Setting the license key from the command line
Using vmrun

	https://support.apple.com/en-us/HT213772
	Published Date: March 8, 2024

	Verification: https://regex101.com/r/GCfKMt/11

	1) Exact regex — Matches major model identifier numbers based on Apple's knowledge base article (more accurate):

	^(Mac(1[345]\|BookPro1[5-8]\|BookAir([89]\|10)\|Pro7)\|iMac(Pro1\|(19\|2[01]))\|Macmini[89]),\d+$

	2) Current or higher regex — Matches model identifiers based on Apple's knowledge base article and may match higher versions before this regex is updated (more future-proof).

	#!/bin/bash

	<<ABOUT_THIS_SCRIPT
	-------------------------------------------------------------------------------

	Written by:William Smith
	Professional Services Engineer
	Jamf
	bill@talkingmoose.net
	https://gist.github.com/2cf20236e665fcd7ec41311d50c89c0e

	#!/bin/bash

	# server connection information
	URL="https://talkingmoose.jamfcloud.com"
	userName="api-editor"
	password="P@55w0rd"
	command="UpdateInventory"

	# XML data to upload
	THExml="<mobile_device_command>

-#! /bin/bash
-# author: Zengwen Yuan
-# date: 2018-01-25
-# File location extracted from installation package https://support.apple.com/kb/dl1888?locale=en_US
-sudo rm -r /Library/Image\ Capture/Devices/HP\ M1130_M1210\ Scanner.app
-sudo rm -r /Library/Image\ Capture/Devices/HPScanner.app/Contents/PlugIns/HPAiOScan.bundle
-sudo rm -r /Library/Image\ Capture/Devices/HP\ Scanner\ 3.app/Contents/Frameworks/HPScanServices.framework/Versions/A/PlugIns/HPDOT4Scan.bundle/Contents/PlugIns/HPAiOTulip.bundle
-sudo rm -r /Library/Image\ Capture/Devices/HP\ Scanner\ 3.app/Contents/Frameworks/HPScanServices.framework/Versions/A/PlugIns/HPDOT4Scan.bundle
-sudo rm -r /Library/Printers/hp/laserjet/p1100_1560_1600series/hpRaster.bundle

	#!/bin/sh

	# Usage: get_lutron_cert.sh [bridge_ip] \| tee cert.pem

	function error() {
	echo "Error: $1" >&2
	exit 1
	}

	login_server="device-login.lutron.com"

	#!/usr/bin/env bash


	# See McAfee KBs:
	# https://kc.mcafee.com/corporate/index?page=content&id=KB88461 # Remove ENS for Mac (ENSM)
	# https://kc.mcafee.com/corporate/index?page=content&id=KB61125 # Remove McAfee Agent

	# Exit upon failed command
	# set -e

	#!/usr/bin/env python

	'''
	download macsysadmin videos

	requires:

	pip install requests
	pip install beautifulsoup4
	'''

	<?xml version="1.0" encoding="UTF-8"?>
	<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
	<plist version="1.0">
	<dict>
	<key>PayloadContent</key>
	<array>
	<dict>
	<key>PayloadContent</key>
	<dict>
	<key>com.microsoft.Excel</key>