Cody Stamps crstamps2

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.

	# Add this file to `spec/support/rails_log_splitter.rb`

	require "stringio"
	require "logger"
	require "fileutils"

	class RailsLogSplitter
	def initialize(all: false)
	@io = StringIO.new
	@logger = Logger.new(@io)

	# Sample Nginx config with sane caching settings for modern web development
	#
	# Motivation:
	# Modern web development often happens with developer tools open, e. g. the Chrome Dev Tools.
	# These tools automatically deactivate all sorts of caching for you, so you always have a fresh
	# and juicy version of your assets available.
	# At some point, however, you want to show your work to testers, your boss or your client.
	# After you implemented and deployed their feedback, they reload the testing page – and report
	# the exact same issues as before! What happened? Of course, they did not have developer tools
	# open, and of course, they did not empty their caches before navigating to your site.