This didn't appear from nowhere. It's the convergence of two libraries I've spent years in — the why-and-systems thinkers and the data-architecture canon — applied to the AI era.
This didn't appear from nowhere. It's the convergence of two libraries I've spent years in — the why & systems thinkers and the data-architecture canon — applied to the AI era.
Lead with the why, not the what. The thesis starts from purpose; the formula is just how.
Don't plan — prototype. Build experiments, not five-year plans. The method behind every system here.
Productivity as a system with trusted structure — not willpower, not tips.
Dimensional modelling — structure that makes data answerable.
The architecture spine: source → integration → delivery.
Auditable, governed, scalable modelling — the banker's discipline in data form.
Principles before tactics; effectiveness as character + system.
Personal knowledge as an asset — capture, organise, express.
Linked notes & ideaverse — the connective tissue of a thinking system.
"PKM is a system, not an app." Plain-text, data you own, identity-first. We agree completely — SBM just adds the data-engineering spine underneath (rules, validation, git) that turns the system into infrastructure.
Long before AI, the question was the same: how do you organize a growing collection of documents so you can actually find and use it? Tiago Forte's PARA (Projects, Areas, Resources, Archives) and Nick Milo's Atlas/ACE gave me concrete answers — and Milo's Maps of Content (MOCs), hand-made index notes that gather and link related material, became one of my core tools. These are real, battle-tested ways to give a vault structure.
They matter more now, not less. Every one of them is a way of imposing structure on a document collection — and a well-structured vault is exactly what an AI can navigate, reason over, and act on. The organizing systems built for human retrieval turn out to be the foundation AI needs too. That's structure beats magic: the work the PKM world did on structure is precisely what makes the AI layer work.
A growth mindset is the precondition for all of this: skills and systems are built, not innate. The same belief underpins growth-hacking — measure, learn, iterate.
A long-standing love of efficiency, effectiveness and clever workflows — but reframed: real leverage comes from systems, not a collection of one-off tips.
Its foundation is the same as mine: a growth mindset, structured experiments, and data-driven iteration. The business cousin of personal life-hacking.
The slip-box: atomic, linked notes that compound into a thinking partner. My Content-Intelligence is "Zettelkasten for the AI age" — the same idea, with AI doing the linking and extraction.
A practical structure for organising knowledge by actionability.
Prototype, don't plan — the experimental mindset behind every block.
Plain-text notes you own — now genuinely powerful in combination with Claude Code in VS Code, where the AI can read and reason over the whole vault.
An AI that works inside your files and data — not a chat window beside them. Where structure meets capability.
For shared, structured content where a database-of-pages beats loose notes.
Some AI thinkers describe the goal as a digital clone — a second self that thinks and acts as you. I aim for something more honest, and more useful: an assistant grounded in your data — reliable because it's engineered, with you in the driver's seat. Not a magic copy of you; a system you can trust, inspect, and correct. Structure beats magic — and a grounded assistant beats a convincing fake.
Grounded assistant > digital cloneGathering and structuring data is only half the job. Before any real decision, you reach for mental models — inversion, second-order effects, opportunity cost, first principles, the map-is-not-the-territory. They're the reasoning frameworks that turn information into judgement (Munger, Farnam Street, and the latticework-of-models tradition).
And here's the link to AI that excites me: a model is a reusable reasoning pattern — exactly the kind of thing you can hand an AI as a rule. Give the system your data and the mental models you trust, and it can pressure-test a decision the way you would — "what's the second-order effect here? what am I not seeing?" Structure carries the facts; mental models carry the thinking; AI applies both, at scale.
Data + mental models → better decisionsA living list — drawn from a personal library of 271 books and counting. I build past these ideas: from organising notes to architecting a system that acts.