Bruno Bigras bbigras

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.

NoteAfterNote-10
Testing: Termux, vmtest, and QEMU
Published: June 19, 2024
Link: https://gist.github.com/NoteAfterNote/7614b0137ac6959e3bba35df66eaa75a
Updated: November 22, 2024

Observation: vmtest-QEMU-8.0.2 is faster than vmtest-QEMU-8.2.5 on Smartphone-1-Is-Android-10

Solution: qemu-system-x86_64

Remove

* -drive if=none,id=vm,format=raw,file=$VM -device virtio-blk-pci,id=vmvirt1,drive=vm

	import pandas as pd
	from dagster import asset, ResourceParam
	from sqlalchemy import create_engine

	# Define the PostgreSQL resource
	class PostgresResource:
	def __init__(self, connection_string):
	self.connection_string = connection_string

	def get_connection(self):

	// Helper to pause execution
	const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

	function getNextOldCatppuccinStyle() {
	// Grab ALL entries on the page
	const entries = document.querySelectorAll(".entry");

	for (const entry of entries) {
	const nameLink = entry.querySelector(".style-name-link");
	const styleInfo = entry.querySelector(".style-info");

	{
	"input": {
	"blocklist": [],
	"compressor#0": {
	"attack": 15.0,
	"boost-amount": 0.0,
	"boost-threshold": -72.0,
	"bypass": false,
	"dry": -80.01,
	"hpf-frequency": 10.0,

	blueprint:
	name: Low Battery Notifications & Actions
	description: >
	# 🪫 Low Battery Notifications & Actions

	Version: 3.4


	🚀 Stay Charged, Stay Smart! Let's automate and take charge of your battery maintenance!🔋⚡

	module "vpc" {
	source = "terraform-aws-modules/vpc/aws"
	version = "~> 3.0"

	name = "talos-demo"
	cidr = "10.0.0.0/18"

	public_subnets = ["10.0.0.0/24"]
	azs = ["eu-central-1a"]
	}

	#!/data/data/com.termux/files/usr/bin/bash

	# https://github.com/orgzly/orgzly-android/issues/8

	orgzly_sync() {
	# if you are using orgzly-revived use the following instead:
	# am broadcast -n com.orgzlyrevived/com.orgzly.android.ActionReceiver -a com.orgzly.intent.action.SYNC_START
	am broadcast -n com.orgzly/com.orgzly.android.ActionReceiver -a com.orgzly.intent.action.SYNC_START
	}

	blueprint:
	name: Bathroom Humidity Exhaust Fan
	description: >
	# 🚿 Bathroom Humidity Exhaust Fan

	Version: 2.7


	Step into the future of freshness - customize it your way and experience the convenience of automated humidity control! 🌿🚿