What annoys me about most self-help books is how much time they spend trying to convince you of their position. They wade through history, details, and backstory, all based on the conceit that you must know all of that (the boring stuff) before they can teach you any of this (the interesting stuff).

I’ll try not to make that mistake. Let’s get straight to the core idea:

I use AI to increase my intelligence and my capability. I use AI as a tool to assist me. That means I have a system for managing my knowledge—literally, what I know.

So, what is your personal knowledge management (PKM) system? If you don’t have one, you’re missing an enormous opportunity. Any insights you gain from AI, instead of nourishing your garden, will be lost as runoff.

This leads to my first rule for working with AI, based on a principle called the Illusion of Explanatory Depth. The idea is that we often think we know something just because information about it is readily available. With AI, this becomes: because the AI knows it, I know it.

But the real question is, do you still know it without access to the AI? Usually, the answer is no—unless you make a conscious effort to learn.

So, step one is to be intentional about:

  1. Where and how you store your work. This means not using the AI app as your only repository.
  2. How you will reuse your work so you can build upon it.
  3. How you will iterate on your work with AI.

Put simply, if your main way of working with AI is just chatting with it—and all that knowledge stays locked in the app—you aren’t getting everything you can out of it. You’ve offloaded the work of learning to the AI, counting on it to do the remembering for you.

Plenty of people talk about Digital Second Brains and Personal Knowledge Management, and I’ve written about it myself. Frankly, I don’t care which system you pick, but it must do two key things:

  1. Provide a system for working with what you know outside of the AI apps themselves.
  2. Include a version control system so you can experiment, iterate, and recover from mistakes.

My favorite tool for the first is Obsidian, because I can use AI not only to read my notes but also to add new learning artifacts directly to them. My favorite for the second is GitHub, because it’s a powerful tool for preserving context. Just copying backup folders isn’t good enough and will lead to long-term problems.

Whether you use Notion, Logseq, Evernote, or Obsidian, I don’t really care. I strongly encourage you to stop reading right now and decide on your system.

Once you’ve done that, go to GitHub.com and create a free account. (I pay for one to keep some projects private.) I’ll give you a simplified GitHub tutorial later in this series, but there are tons of great resources out there. Even better, we’ll use our new notes system, GitHub, and AI to learn these tools together.

When picking your note system, I strongly encourage you to choose one that uses markdown.

For this series, I’ll be using the PARA organization system (Projects, Areas, Resources, Archives) with Daily Notes. Every day, I start with a fresh, dated note that acts as an inbox—it’s an extraordinarily powerful habit. Obsidian handles this setup well, but I’ve built it in Notion before, too. You could probably make Apple Notes or Google Keep work, but I don’t recommend it. Just pick something!

Why Not Use Your Chat History?

While it’s cool that chat histories are saved, moving them between AI models is a pain. More importantly, if you start working in a command-line environment, your chats will not be saved, so you’ll need another system to preserve your work.

Besides, organizing information by the application that created it is a terrible system. You would never store a toothbrush, a toilet brush, and a hairbrush together just because they’re all “brushes.” In the physical world, we organize by context. Your kitchen holds all the different tools you need for cooking. Your bedroom contains what you need to sleep and get ready. We don’t organize by manufacturer in the real world, so stop doing it on your computer.

Stop organizing your information by what it is and start organizing it by what it is for. Your knowledge should serve your purpose, so organize it that way.

Ready?

Do you have a note system? Do you have a GitHub account? Great. Now we’re going to learn how to learn.