Discover gists

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.

Since we're dealing with simpler AV1 encoding, that does mean we'll be eskewing aomenc-av1, since it requires 2-pass encoding to be able to take advantage of it, and I use it externally since I have access to a special build.

Now, let's get on with the simple guide.

You'll first need to be reasonably competent with command line builds or use a recent up to date ffmpeg GUI with support for SVT-AV1.

As such, I would recommend getting a master git ffmpeg build for the operating system of your choice right here:

https://github.com/BtbN/FFmpeg-Builds

🔐 GitHub SSH Cheat Sheet

# Generate SSH key
ssh-keygen -t rsa -C "betterthanbestest@gmail.com"

# Add private key to agent
ssh-add ~/.ssh/id_best

# List loaded keys

Windows XP/2003 Product Keys

These keys have been tested to work. These are a combination of keys from Chinese websites, trial keys inside ISO's, auto-activate keys in OEM ISO's, and directly from Microsoft's website, all aggregated for your convenience.

Usage

In order to use these keys, you need the right edition of Windows XP/2003. Not just Home/Pro/Enteprise/etc, whether it's a Retail, OEM or Volume version. There's a clear cut way to check this.

If you see this, you have a Retail copy.

Common Terminal Commands

Key Commands & Navigation

Before we look at some common commands, I just want to note a few keyboard commands that are very helpful:

Up Arrow: Will show your last command
Down Arrow: Will show your next command
Tab: Will auto-complete your command
Ctrl + L: Will clear the screen

The Direct Response Copy Skill

Write copy that converts. Landing pages, emails, sales copy, headlines, CTAs, social posts — anything persuasive.

This is an AI skill file. It turns any AI into a direct response copywriter trained on the frameworks of Schwartz, Hopkins, Ogilvy, Halbert, Caples, Sugarman, and Collier. Instead of getting generic AI copy, you get internet-native writing that sounds like a smart friend explaining something — while quietly deploying every persuasion principle in the book.

How to use this

Orchestrator Agent Creation Guide

This document provides a comprehensive guide for creating Orchestrator Agents in OpenCode. Based on the reference implementation (@agent/orchestrator.md), this guide details the structure, patterns, and best practices for building intelligent routing agents.

1. Overview

An Orchestrator Agent serves as a central dispatch system. Unlike standard agents that execute tasks, an orchestrator's sole purpose is to analyze user requests and delegate work to specialized subagents.

	# /// script
	# requires-python = ">=3.11"
	# dependencies = [
	# "pandas",
	# "openpyxl",
	# "click",
	# ]
	# ///

	"""