Holographic model cards

Put most of what you have where almost nothing can hurt it, and a small remainder where the upside is unlimited. Avoid the deceptive middle ground that is neither safe nor rewarding.

The shape caps your worst case while leaving your best case open — you can only lose the small side, but you can win with it enormously.

— after Nassim Taleb, Antifragile (2012)

Knowledge of what is wrong is more durable than knowledge of what is right. So subtract first: remove the harmful, the fragile, the unnecessary — before reaching for anything new.

Health by removing the toxin, clarity by cutting the clutter, strength by deleting the weakness. The absence of a negative often beats the presence of a positive.

— after Nassim Taleb & classical theology

In some systems, perceiving the thing changes the thing. Belief moves the fundamentals, the moved fundamentals confirm the belief, and cause and effect run in a loop instead of a line — the two spiralling together until the loop snaps.

A company seen as the leader draws cheaper money, better hires and more trusting customers, and so becomes the leader. Past a threshold, the story is not noise; it is load-bearing. The same loop runs just as fast in reverse.

— after George Soros, The Alchemy of Finance (1987)

Everyone has a domain where their judgment is genuinely reliable, and a larger one where it merely feels reliable. The circle is the border between them, and the skill is not widening it — it is knowing precisely where it ends.

The expensive mistakes never come from what you know you do not know. They come from the ground where you think you know and do not. So ask of any belief: did I earn this, or absorb it?

— after Charlie Munger & Warren Buffett

Behaviour follows incentives, not intentions. To understand why a person, a company or a system does what it does, stop reading the stated goals and read the reward structure. The output matches the incentive almost every time.

The uncomfortable half of the model is turning it inward. What do your own rewards actually pay you to do — and is it the thing you say you want?

— after Charlie Munger, on reward superresponse

The cost of a choice is never what you spent. It is the best thing you did not do instead. Time and attention are finite, so every yes is a silent no to something else.

Munger measures comparatively: do not weigh an opportunity against doing nothing, weigh it against the best option already on your table. If it cannot beat that, it is not an opportunity.

— after Charlie Munger & Warren Buffett

An engineer expecting ten-thousand-pound trucks rates the bridge for thirty thousand. Not pessimism about the trucks — honesty about her own measurements. Leave that same gap between what you pay and what you believe it is worth.

The margin is sized to how wrong your estimate could be, never to how confident you feel. Confidence is precisely the thing it is insuring against.

— after Benjamin Graham; kept by Munger & Buffett

When two, three, four forces all push in the same direction, you do not get a bigger result — you get a non-linear one. The effects multiply instead of adding. Unbreakable moats, sudden crashes and unshakeable habits are all built this way.

Read it both ways. As a warning: how many biases are pushing me toward this at once? As a design: stop over-tuning one factor and stack several that point the same way.

— after Charlie Munger, Elementary Worldly Wisdom (1994)

Most thinking is reasoning by analogy: it worked for them, so a modified version should work for me. Fast, cheap, and right most of the time — but it can only ever produce variations on what already exists.

The alternative is to break the problem down to the few things you know are true, the facts that cannot be reduced further, and rebuild upward from there. It is expensive, so spend it where the stakes are high and the standard answer is old.

— after Aristotle; sharpened by Elon Musk