Thoughts on AI, knowledge management, and building a second brain.
Research is bottlenecked by retrieval. You read a hundred papers over a year; you remember a few dozen specific arguments and forget the rest. When you sit down to write a literature review, you re-Go
Capture is the load-bearing habit in personal knowledge management. If the friction is more than a few seconds, you won't capture, and a system you don't capture into is a system that doesn't compound
Most people don't start with one note-taking app. They start with five or six: a Notion workspace that grew unwieldy, an Obsidian vault for personal writing, Apple Notes for quick captures, browser bo
Most AI conversations start at zero. You re-paste the same project context, the same prior decisions, the same research, and lose most of the compounding value of working with a model over months. Bui
"Local-first" became a load-bearing phrase in note-taking around 2020 — the moment people realized Evernote could revoke their data, that Roam Research could disappear overnight, that they had ten yea
Obsidian is the closest thing to a community standard for markdown-vault note-taking. It has a huge plugin marketplace, devoted users, and a sane storage format (plain `.md` files on disk). What it do
Retrieval-augmented generation — RAG — is the technique that turns a generic language model into one that answers from your data. It's the foundation under every "chat with your docs" product, every e
Switching note-taking apps is the single most expensive decision in personal knowledge management. The migration looks like a weekend of work. It's actually a six-week trust-rebuilding exercise. Every
The "how do I organize my notes" question is the single most-asked question in every PKM community on the internet. Folders versus tags is the old argument. Folders versus tags versus wikilinks is the
Niklas Luhmann published more than 70 books and 400 papers while running a famously simple system: index cards, numbered, cross-linked, kept in a wooden box. The "Zettelkasten" — German for slip-box —
If you want your AI to know what you know — without re-pasting context every conversation — you need a knowledge base your AI can actually read. That's a smaller category than it looks. Most "AI featu
Personal knowledge management isn't a tool. It's a practice — capturing, organizing, and revisiting what you know so it compounds. The tool just has to stay out of your way. This piece walks
A second brain isn't a notes app. It's a system you trust enough to outsource memory to. You stop trying to remember; the system remembers, and your conscious attention is free for thinking.
The promise of an AI knowledge base is that your conversations start with full context. The reality is that most AI-meets-notes setups produce one of two failure modes: the AI can't actually read
Research notes have a specific failure mode: you collect a hundred sources for one project, half of them are gone six months later when you start the next one. The links rot, the highlights aren'
Folders have one address per note. Graphs have many. That's the whole pitch for knowledge graph note-taking. You stop forcing every note into exactly one home and start letting your links carry t
Every AI conversation starts from zero. You re-explain the same context, re-establish the same patterns, and lose your best thinking between sessions. Meanwhile, your notes scatter across half a dozen