The living budget
Scrub the numbers inside the sentence and watch it re-reason — pin a goal and let income flex to keep the books balanced.
Open the explorable →Explorables
Most charts you look at. An explorable you think with. Every number — on the chart and in the sentence — is a handle: drag it, and the whole thing re-reasons in front of you. Nothing here is fixed text; the understanding is something you reach by playing, not by reading.
It's a small library in three parts — Components, Dashboards and Essays. Each is a working example and a teachable pattern: a vocabulary you — or an AI building the next one — can learn from.
Components
The building blocks — one idea, one interaction. Reach for these first: a reactive document, a dependency network, an uncertainty curve, a trade-off frontier, a budget trend — plus a catalogue of fifteen more. This is the vocabulary every dashboard and essay is assembled from.
The critical path
A project's tasks and dependencies. Drop a delay onto one and watch the finish slip — and the bottleneck jump to a path that was fine yesterday.
Open the explorable →How long, really?
Shape an estimate's low / likely / high and see your real chance of beating a deadline — the area past the line is the risk.
Open the explorable →Which option?
Five ways to deliver, plotted on cost vs time. Drag one and watch it cross the frontier of real choices — or get beaten on both.
Open the explorable →Are we over budget?
The spend curve climbing toward the budget line — drag the months, or the budget itself, and see exactly when you cross it.
Open the explorable →Fifteen more, in miniature
A growing kit of smaller component ideas — waterfalls, risk & priority maps, sankeys, calendars, sensitivity, capacity grids — each a quick live demo to drag.
Browse the catalogue →Dashboards
Many components wired into one coordinated model: a single hero lever drives every panel at once. Each is built from public data and video transcripts, and cites its source so you can check it.
The Federal Reserve
One rate dial drives inflation, unemployment, the corridor and the ripple at once — the dual-mandate "oven". Built from a USAFacts explainer.
Open the dashboard →The US national debt
Budget → deficit → debt, and the interest, debt-to-GDP and ceiling that follow. Project the years forward and watch it all move.
Open the dashboard →The NZ economy
The Reserve Bank's OCR, inflation vs the 1–3% band, the oil-shock outlook and a bank stress test — built from six months of RBNZ videos.
Open the dashboard →Essays
Long-form writing where the prose itself is interactive — change a number, or your whole point of view, and the argument re-reasons around you. Includes short guided explorables that pose a question and walk you to the answer — sometimes by letting you reach for something and discover you can't have it.
Bitcoin, over time
Four little machines you can drive: the 21M supply, the difficulty feedback loop, the four-year cycle, and "what if you'd bought?"
Open the treatise →Moving the floor
Retraining the job market as one living system — change any number, or your perspective (worker, employer, Treasury, society), and the whole essay re-reasons around you.
Open the essay →Low inflation and low unemployment?
Drag the central bank's one lever and discover the corner everyone wants — and that no setting of the dial can ever reach.
Open the explorable →Can the government just print money?
Your wallet vs a loaf of bread. Print all you like — the loaves you can afford won't budge. Find out why, and what actually does lift them.
Open the explorable →Jobs, or just inflation?
Unemployment isn't one thing — it's three, stacked. Print money and watch which band shrinks, and why the floor underneath won't move.
Open the explorable →How they're built
All three libraries obey the same laws — the direction for building a new one, whether you're a person or an LLM agent:
- You reach the insight by manipulating it — strip the words and the point still lands.
- Every value is a handle, whether it's a bar, a node, or a number in the prose.
- It reasons — backward to the cause, forward to what happens next — and the words change as you do.
- It stays honest: no fabricated data, no axes that clip, and no flashing (motion is steady and respects reduced-motion).
- The same idea handles simple and complex data alike.
- Guided discovery: a question → a first-move nudge → a payoff that finds you → an insight you can't miss.
Read another way, these are a reference an LLM can learn from — the patterns, the rules and the kill-criteria that turn raw data into something you can think with. Pick a component, compose a dashboard, or write an essay: the library teaches the move.