Andreas Motl amotl

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.

Using https://volkovlabs.io/plugins/volkovlabs-echarts-panel/ you can create a pretty nice graph showing both wind speed & direction.

> Write a blog post about this code trick:

import logging
import os

from funcy import monkey

from django.db.models.query_utils import DeferredAttribute

	# Docker build cache mount persistence
	#
	# Usage: add `extends: .docker-build` to your Docker build CI job, and set
	# environment variable `DOCKER_CACHE_MOUNTS` to a comma-separated list of
	# build cache mount IDs, then use `--mount=type=cache,id=...,target=...`
	# in your Dockerfile
	#
	# Adapted from: https://github.com/reproducible-containers/buildkit-cache-dance
	#
	# See bottom for license

	#!/usr/bin/env uv python
	"""
	CrateDB Settings Extractor

	This tool extracts settings from CrateDB's documentation and outputs them
	in either JSON or Markdown format, or the SQL statements to set the default value.
	It parses the HTML structure of the documentation to identify settings, their
	descriptions, default values, and whether they're runtime configurable.

	Author: wolta

	#!/usr/bin/env bash
	#
	# Try and squash unstaged changes into existing branch commits.
	#
	# This command examines each unstaged file and attempts to create a fix-up
	# commit to squash it into its natural parent in the current branch.
	#
	# - If it's able to do this for all modified files, the fix-up files are
	# automatically squashed in.
	#

	// I'm tired of extensions that automatically:
	// - show welcome pages / walkthroughs
	// - show release notes
	// - send telemetry
	// - recommend things
	//
	// This disables all of that stuff.
	// If you have more config, leave a comment so I can add it!!

	{

	# synthio_midi_synth.py - pretty usable MIDI-controlled synth using synthio in CircuitPython
	# 11 May 2023 - @todbot / Tod Kurt
	# Uses cheapie PCM5102 DAC on QTPY RP2040
	# Video demo: https://www.youtube.com/watch?v=N-PbbWWDE6k
	# Features:
	# - midi velocity controls attack rate (gentle press = slow, hard press = fast)
	# - notes have small random detune on all oscillators to reduce phase stacking
	# - adjustable number of detuned oscillators per note 1-5 (midi controller 83)
	# - five selectable waveforms: saw, squ, sin, noisy sin, noise (midi controller 82)
	# - vibrato depth on mod wheel (midi controller 1)

	import numpy as np
	from netCDF4 import Dataset
	import os
	import datetime
	import xarray as xr
	import pandas as pd
	import matplotlib.dates as mdates
	import matplotlib.dates as mdates
	from matplotlib import pyplot as plt
	from matplotlib.colors import LogNorm

	#!/usr/bin/env python
	import argparse
	import os
	import shutil
	import subprocess
	import packaging.version
	import grayskull.strategy
	from importlib.resources import open_binary
	import yaml