Start here · free
Give your assistant a memory it owns
Twenty minutes, one file, nothing but a terminal. At the end, a brand-new session will tell you something true about your own life, from memory, without you re-explaining a thing.
Every AI chat starts from zero. It forgets what you decided last week and hands specific problems generic answers. The fix is not a better prompt; it is a small folder of plain files your assistant reads at the start of every session. This is the first piece of that, and it stands completely on its own.
The trick that makes it painless: you do not build it by hand. You hand a setup file to the same AI assistant you already use, and it interviews you, builds the piece on your machine, and proves it works. That is how the whole course installs, one piece at a time. This piece is free.
Do this now
- Download the setup file: quickstart.md.
- Open your AI CLI assistant (Claude Code, Codex CLI, or any agentic assistant that reads files and runs commands) in a terminal.
- Point it at the file and say: “Read this file and set me up.” Answer its questions. It does the rest.
What you will have
A git-tracked folder of markdown that is your assistant’s durable memory. A one-line kernel that loads it every session. One real document about a topic from your life, dated. And the proof: a cold session answering a question about you from the file alone.
That was one of nineteen. The rest install the same way: a task store so “what’s next” returns one confident answer, senses so your assistant can read your real calendar and inbox, rituals so every session ends with the record updated and committed. When you want the whole bench, here is the story and the course.