I built myself an AI second brain. It's the most useful thing I've ever made.

Every paper I read, every article I bookmark, every project note, every technical learning — indexed, embedded, and searchable through a RAG interface.

When I'm designing a new system and think "I read something about this approach six months ago," I don't dig through bookmarks. I ask my second brain: "What approaches have I saved for handling multi-tenant vector databases?" And it retrieves the relevant notes with context.

The tech stack is simple: Obsidian for notes, a Python script that chunks and embeds them nightly, Qdrant for the vector store, and a simple chat interface built with Streamlit.

Total cost: the time to set it up (one weekend) and the discipline to capture notes consistently.

What makes it genuinely useful vs a gimmick: the notes have to be in YOUR words. Saving raw articles doesn't work as well as saving your annotations and takeaways. Because your future self doesn't need the article — they need your interpretation and why you found it relevant.

If you're an engineer who reads a lot and forgets where you read things, build this. It's a two-day project that becomes more valuable every week as your knowledge base grows.

The best AI application is one that makes YOU smarter.