Starting From Scratch

Install the tools. Have your first conversation. This is where every AI second brain begins.

You've seen the demos. Someone on Twitter shows their AI reading their notes, managing their email, writing drafts in their voice, remembering decisions from three months ago. It looks like magic, and you want it.

Then you try to build it and hit a wall. Where do you even start? What tools? What structure? The demos never show the beginning — only the result.

This newsletter is the beginning.

Every week, I'm going to walk through how I built an AI-augmented second brain from nothing — the actual steps, tools, and mistakes, not the theory. By the end of this series, you'll have a working system. Not a copy of mine, but one shaped by your own work and needs.

But first, we need to talk about what a second brain actually is and why AI changes everything about it.

Your brain is a terrible hard drive

You have ideas all day. Connections between things. Solutions to problems you weren't actively thinking about. But your brain stores almost none of it reliably. By tomorrow morning, most of today's insights are gone.

A second brain is just an external system where your thinking accumulates instead of evaporating. Notes, documents, bookmarks, project files, daily journals — anything you might need to find again, stored in one place with enough structure that you can actually find it.

This isn't new. People have been doing this with notebooks, filing cabinets, and apps for decades. Zettelkasten, Getting Things Done, PARA — all systems for organising knowledge outside your head.

The problem has always been the same: maintaining the system takes work. Filing, connecting ideas, finding things again — every step demands effort that competes with the actual thinking you're trying to do. Most people set up a note-taking app, use it enthusiastically for a month, and quietly abandon it because the overhead exceeds the benefit.

That's where AI changes the equation.

What AI actually does for a second brain

Forget "AI assistant" in the chatbot sense — visiting ChatGPT to ask questions, copying text back and forth. That's useful but it's not what we're building.

What we're building is AI that lives inside your knowledge system. It can read every note you've written. It can search across your entire vault. It can create new files, edit existing ones, and run automations — all while understanding the context of your projects, your writing, and your history.

The practical difference:

The second brain goes from a filing cabinet you maintain to a system that works alongside you — one that extends your thinking instead of replacing it.

What I built (and what you're going to build)

In November 2025, I started with an empty Obsidian vault and Claude Code running in a terminal. Three months later, the system — I call it Cerebro — has:

I'm not showing you this to impress you. I'm showing you where we're headed. You won't build all of this — you'll build what you need. But you'll have the foundation to build any of it.

This series walks through the journey from empty vault to working system. Each issue adds one layer. By issue four, you'll have a functional second brain with AI integration, daily notes, and your first automation.

Why Obsidian and Claude Code

You need two things: a knowledge base and an AI that can access it.

Obsidian is the knowledge base. It's free, it stores everything as plain markdown files on your computer, and it doesn't lock your data into a proprietary format. Your notes are just files in a folder. If Obsidian disappeared tomorrow, your data is still there.

That last point matters more than it sounds. When AI can read and write your files directly, you need those files to be in an open format on your local machine — not trapped inside Notion's database or Apple Notes' sync system.

Claude Code is the AI layer. It's Anthropic's command-line tool for Claude. It runs in your terminal and — this is the key part — has direct access to your filesystem. It reads, writes, edits files, runs commands, and searches your vault, all from the same terminal where you'd run any other tool. Not a browser chatbot — an AI operating inside your working environment.

You'll need:

That's it. No other tools, no complex setup, no cloud services to configure.

Your first session

Let's get the foundation in place.

1. Install Obsidian and create a vault.

Download Obsidian, open it, and create a new vault. A vault is just a folder — Obsidian will ask you to name it and pick a location. Call it whatever you want. I called mine Cerebro. Yours might be "Brain" or "Vault" or your name. Doesn't matter.

You now have an empty folder that Obsidian treats as a knowledge base.

2. Install Claude Code.

Open a terminal and run:

(If you don't have Node.js, install it first from nodejs.org. Version 18 or later.)

3. Start Claude Code in your vault.

Navigate to your vault folder in the terminal and run claude:

The first time, it'll ask you to authenticate with your Anthropic account. Follow the prompts.

4. Have your first conversation.

You're now talking to Claude inside your vault. It can see every file in the folder. Try these:

That's it. You have AI inside your second brain. It's not doing much yet — but it can see your files, create new ones, and respond with awareness of what's in your vault. Everything we build from here is layering capability on top of this foundation.

What's coming

Here's where this series goes over the next few weeks:

After that: agents, memory systems, content workflows, CRM integration, MCP servers, and whatever else the build produces. Each issue adds one layer, building on the last.

If you've been watching other people's setups and wondering how to get there — this is the path. Start here. Build one layer at a time.

See you next Friday.

Jim

Second Brain Chronicles is published every Friday. You're receiving this because you subscribed via Kit.com.