Enter your access password to view the lessons and prompt library.
Work through each module in order, or jump to any lesson as a reference. Every prompt is ready to copy.
Understand the difference between how most people use AI (starting from scratch every time) and how this system works: building up knowledge that stays and grows.
Most people treat AI like a search engine. Ask a question, get an answer, close the tab, forget it. That is called RAG (Retrieval Augmented Generation). Every time you ask a question, the AI searches your files or the internet, grabs relevant pieces, and gives you an answer. Then it forgets everything and starts over the next time.
Andrej Karpathy, co-founder of OpenAI and the person who coined the term "vibe coding," recently described something better. Instead of asking Claude to find and synthesize documents every time, you have Claude pre-compile everything into a persistent, interlinked wiki once. From that point forward, every new piece of information you add gets woven into what already exists.
| Attribute | Traditional RAG | Second Brain (This System) |
|---|---|---|
| Starting point each query | Raw documents, from scratch | Pre-built, synthesized wiki |
| Cross-document synthesis | Done over and over every time | Already done, baked into pages |
| Compounding value | None | Grows with every source you add |
| Contradiction handling | Ignored | Flagged and resolved during ingest |
| Adding a new source | Added to search index only | Updates 10 to 15 related pages |
Think of your raw files as ingredients. The AI is the chef. The wiki is the finished meal. Every time you add a new ingredient, the chef works it into the existing menu rather than starting a new kitchen from scratch. Karpathy found he didn't even need a vector database; the LLM is good at navigating a well-structured index file on its own.
Get your past AI conversation history exported and ready to feed into your second brain.
Most people have been using ChatGPT and Claude for over a year. That is hundreds of conversations: research sessions, business strategy, writing drafts, brainstorming, learning. It is all sitting locked inside those platforms doing nothing. Today you get it out.
conversations.jsonUse these for ongoing capture after the workshop. Install one now so it is ready.
Have Obsidian installed and a vault with the correct folder structure ready to receive your data.
second-brain. Pick a location you will remember. Everything is stored as plain .md files on your computer. No cloud lock-in. You always own your data.Have Claude Code running inside your vault folder and ready to process files, run compile operations, and write to your second brain.
sudo npm install -g @anthropic-ai/claude-code and enter your password when prompted.
Have Claude connected to your Notion workspace and able to create, read, and update pages as your second brain.
| Obsidian (Paths A & B) | Notion (Path C) | |
|---|---|---|
| Setup difficulty | Moderate | Easy: just OAuth |
| Compile speed | Fast (direct file access) | Slower (one API call per page) |
| Data portability | Files you own locally | Cloud-hosted, export needed |
| Offline access | Yes | No |
| Best for | Power users, large bases | People already in Notion |
/mcp; you should see notion listed as a connected server. Then test with: "Search my Notion workspace for any pages related to business ideas and summarize what you find."Run your first compile operation and watch your wiki start to take shape from your real conversation history.
This is the moment where it all becomes real. You have your exports, your vault is set up, and Claude is connected. Now you feed it your history and watch it build.
Start a local Claude Code session from inside Claude Desktop, point it at your vault folder, and run compile operations without ever opening a terminal.
.md in the right place in your vault. That way the content shows up as normal notes you can open and link in Obsidian.
.json files will not appear as readable, linkable notes the way .md files do. If your exports are still JSON, prompt Claude to convert them to Markdown before you rely on browsing the vault in Obsidian.Have a simple, sustainable routine for keeping your second brain growing without it feeling like a chore.
Every query you run, every output Claude generates: file it back into the wiki. Your explorations become permanent additions to your second brain. Over time, the wiki reflects not just what you have read, but how you think.
Lex Fridman feeds his second brain into voice mode and talks to it on long runs. He explains concepts out loud and asks the wiki what he is getting wrong or oversimplifying. This is the Feynman technique (learn by teaching) with a study partner who knows everything in your knowledge base.
These are the core operations that run your second brain. Save them, use them, adapt them.
Run this the first time to build your wiki from scratch. Replace [time period] with how long you have been using AI tools.
Run this whenever you add new files to your raw folder. Claude integrates them into the existing wiki without overwriting what is already there.
Run this once a month to keep your wiki healthy. Finds orphan pages, contradictions, stale claims, and gaps.
Ask Claude to reason across your entire knowledge base. This is where the compounding value shows up.
Your second brain tells you what it needs to grow. Run this to generate your next reading and research agenda.
Surface patterns in how you have made choices over time.
All 6 prompts as a Markdown file, ready to drop into your vault.