TJ Singleton tjsingleton

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.

	require "csv"

	# Reads from STDIN or a file, converts \G mysql formatted results to csv, and writes to STDOUT
	#
	# STDIN Usage:
	# mysql ... \| ruby mysql-to-csv.rb
	#
	# File Usage:
	# ruby mysql-to-csv.rb results.txt
	#

 require 'time'
 require 'json'
 require 'pp'
 require 'open3'
 class DeleteShare
   def initialize
     @instances = Hash.new {|h,k| h[k] = [] }
     @public_ips = Hash.new {|h,k| h[k] = [] }
     @volumes = Hash.new {|h,k| h[k] = [] }

	# swap loops so we can exit early from the inner loop once we've searched the valid range

	require "pp"
	require "minitest/autorun"

	class StockTest < Minitest::Test
	attr_accessor :loops

	def pick_best_time_to_buy(stock)
	best_value = 0

	route: 66.249.64.0/20
	descr: Google
	origin: AS15169
	notify: noc@google.com
	mnt-by: MAINT-AS15169
	changed: noc@google.com 20110301
	source: RADB

	route: 66.249.80.0/20
	descr: Google

	#!/usr/bin/env bash

	usage (){
	echo "Usage: $0 new-branch-name"
	exit 1
	}

	[[ $# -eq 0 ]] && usage

	git fetch origin

	_ = nil
	scope = binding

	loop do
	print "> "
	code = gets.chomp
	_ = eval(code, scope)
	puts "=> #{_.inspect}"
	end

	last_line = nil
	index = 0

	ARGF.each_line do \|line\|
	index += 1
	next if index < 35
	puts last_line if last_line # skip last line

	line.gsub!("#{0x01}\\#{0x01}", "#{0x01}#{0x01}")
	line.gsub!("\\", "\\\\")

	require_relative './mylib'

	run ->(env){ [200, {}, ['']] }