Erik OShaughnessy JnyJny

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.

	import re
	from collections import Counter
	from datetime import datetime

	import dateutil

	key_fmt = "{d.year}-{d.month:02d}"


	def parse_log(path: str, target_year: int = 0) -> tuple[str, str]:

	from __future__ import annotations

	import gc
	import sys
	import weakref


	class InstancesMixin:
	def __new__(cls, args, *kwargs) -> Class:
	instance = super().__new__(cls)

	"""Find import and import from statements
	"""
	import ast

	from pathlib import Path


	class ImportRecord:
	@classmethod
	def toplevel_imports(cls, root: Path, prune: list[str] = None) -> list[str]:

	""" short module description goes here.

	Longer module description goes here.
	"""

	from typing import List, Union


	__author__ = "[email protected]"

	#!/usr/bin/env python3

	""" a quick and dirty books query thing

	"""

	from pprint import pprint

	import requests

	#!/usr/bin/env python3

	from string import digits
	from typing import List, Tuple


	def make_number_lists(base: str = None) -> Tuple[List[str], List[str]]:

	base = base or digits[1:]

	{"caused_by":"macos","macos_version":"Mac OS X 11.3.1 (20E241)","os_version":"Bridge OS 5.3 (18P4556)","macos_system_state":"running","incident_id":"5D23D70E-B825-42B6-AA03-135B7361F5C4","timestamp":"2021-05-19 16:30:12.00 +0000","bug_type":"210"}
	{
	"binaryImages" : [
	[
	"3fa744a7-0f34-36da-aa6d-f6503a3176ed",
	18446744005107531776,
	"K"
	],
	[
	"00000000-0000-0000-0000-000000000000",

	# Makefile - Generate HTML, PDF, EPUB, MOBI from ASC

	# This is how you assign a string to a variable name in 'make'. The
	# variable name doesn't have to be all caps and the equal doesn't have
	# to be snugged up to the variable name; it's just how I write
	# Makefiles

	FILENAME= the_document

	# $(IDENTIFIER) is how you reference a 'make' variable, you can use

	import hid

	d = hid.device()
	d.open(0x2c0d, 1)