The Whole Toolbox — Mental Models

Give a small child a hammer and, the old line goes, everything they meet suddenly needs pounding. The lamp, the cat, the coffee table — all of it gets read as a nail, because a nail is the only thing a hammer knows how to see. It’s funny in a toddler. It is not funny in a grown, intelligent adult who has exactly one good idea about how the world works and quietly bends every situation they meet until it fits that idea. And here is the uncomfortable part: that adult is the natural endpoint of studying models one at a time. Learn only supply and demand and every problem becomes a market. Learn only psychology and every failure becomes a bias. Learn only game theory and everyone is a strategic player. The single sharpest model in your kit, held alone, doesn’t make you clear-sighted. It makes you a person with a hammer.

This course is the cure, and it’s the reason all the earlier courses existed. You didn’t collect inversion, base rates, feedback loops, incentives, supply and demand and game theory so you could pick a favourite. You collected them so that when a real, unlabelled situation lands on your desk — a bank rumour, a job offer, a product that’s suddenly everywhere — you can run several of them over it at once and build one judgment from what they collectively say. That interlocking mesh of models, read together, is what Charlie Munger called a latticework, and using it is the last and highest skill this whole site is built to teach.

The one idea to take away

Before five lessons unpack it, here’s the whole model in a line:

The one-sentence version

A latticework is a set of mental models from different disciplines, held at once and run over the same situation, so that they check each other (one vetoes another’s blind spot), complete each other (each explains a different part), and sometimes stack the same way (a lollapalooza) — turning many partial lenses into a single, better judgment.

Three verbs are doing the work in there, and the whole course hangs on them. Models check each other: a thrilling story runs into a cold base rate and the base rate wins. Models complete each other: supply and demand explains the price while game theory explains the players — different parts of one picture. And models stack: when several independent models all point the same way, their forces multiply into the lollapalooza effect you met last course — except now you’re the one holding the models, not the one being carried off by them. The skill is knowing which of the three is happening, because each means something completely different for what you should do next.

Before you read — take a guess

A friend is about to sink their savings into a hot new crypto coin. They say: 'Look, everyone I know is buying it, it's up 10x this year, and I already put in a little and it's working.' You want to think about this well. What's the LATTICEWORK move?

Since several forces point toward buying, that agreement alone proves the coin is a good buyRun several models over it at once and see how they line up — social proof and momentum push toward buying, but base rates (most 10x coins later collapse) and incentives (the people hyping it profit from the hype) push the other way; the clash between them is the thing to investigateWhichever way the most vivid, emotionally gripping argument points — that's usually the right readPick the single model you trust most and apply only that one, ignoring the rest

Watch models agree — and clash — on one decision

Reading “models check and complete each other” is one thing. Watching it is another. Below is a decision desk. Pick a scenario — a bank rumour where a queue is forming, or a viral product your company could bet on — and you get a panel of model-lenses. Switch them on one at a time. Each lays its own reading on the desk and nudges a needle toward one of two poles.

Two things are worth hunting for. First, switch on lenses that all point the same way and watch the needle pin hard to one pole — that’s a stacked, lollapalooza reading, where many independent models converge. Second, switch on lenses that point opposite ways and watch the needle fall back toward the middle while the desk flags a tension. That flagged clash is not a failure of the method — it is the method working. The disagreement between your models is the most valuable thing the desk can show you, because it points straight at what you don’t yet understand.

The decision desk

One situation, several lenses — do they agree or clash?

Switch on one model-lens at a time. Each lays its own reading on the desk and nudges the needle. Watch what happens when several agree — and what happens when they pull against each other.

A rumour spreads that your bank is in trouble. A queue is forming outside. Do you rush to pull your savings out now, or hold?

Where the models, together, lean

HoldJoin the run

The desk readsNo lens is on yet. Turn one on to see what a single model says — then add more and watch them agree or clash.

Model-lenses — switch each one on

IncentivesIf enough others withdraw, the bank fails and late withdrawers lose — so each person is personally rewarded for pulling out first.Social proofA visible, growing queue signals "sensible people are getting out," and under uncertainty we copy the crowd.Game theoryA bank run is a coordination trap: "everyone withdraws" is a self-fulfilling equilibrium — if you expect others to run, running is your best response.Feedback loopEach withdrawal weakens the bank and lengthens the queue, which triggers more withdrawals — a reinforcing loop that feeds on itself.Base ratesMost bank rumours fizzle out and deposits are usually insured, so the base rate of a genuine collapse from a rumour is low.Second-order thinkingIf everyone reasons "just pull mine out to be safe," the run itself causes the collapse that no one wanted — the cure is the disease.

Turn on lenses that all lean the same way and the needle pins to one pole (a stacked, lollapalooza reading — strong, but a cue to slow down). Turn on lenses that pull against each other and the desk flags the tension — the disagreement is the signal to investigate, not something to average away. That is a latticework thinking out loud.

Carry two things out of that panel. First, a latticework doesn’t hand you a verdict — it organises the argument. When lenses stack, you get conviction and a warning. When they clash, you get a map of exactly where to look. Either way you’re better off than with one model shouting alone. Second, which lenses you switched on was a choice — and a decision you don’t know how to read (you’re outside your circle of competence) is one where you can’t tell which models even apply. That meta-skill of choosing the lenses is what separates a real latticework from a person flipping switches at random, and it’s what lesson 4 is about.

Why one model always feels like enough (and why that’s the trap)

Here’s the psychological catch that makes this hard. From the inside, one model never feels like too few. It feels like clarity. The person who explains every event through incentives isn’t aware of forcing anything — they genuinely see incentives everywhere, vividly, and the explanation feels complete every single time. That’s the danger of the hammer: it doesn’t announce itself as a hammer. It announces itself as insight. The single model you love most is precisely the one you’ll over-apply without ever noticing, because a good model is genuinely illuminating right up to the edge of where it stops applying — and it gives you no warning at that edge.

The trap in one line

Your favourite model will feel like enough on every problem, including the ones it’s wrong about — because a model that fits most of the time gives no signal when you’ve wandered past its edge. The defence isn’t a better single model; it’s the deliberate habit of asking, out loud, “what would a different model say about this?” — which is exactly why the latticework has to be run on purpose, not trusted to instinct.

The map of the course

Five teaching lessons build the capstone from the ground up, then one exam you can’t undo. The route:

Why One Model Is Never Enough — the man-with-a-hammer tendency in full: how a single model distorts, why the model you know best is the one you’ll over-apply, and how your circle of competence tells you when a situation is one your models can even read.
The Latticework — Munger’s actual idea: models from many disciplines as an interlocking mesh you hang experience on, why cross-disciplinary breadth beats depth in one field, and the handful of big models that pay their way across almost every domain.
Three Ways Models Combine — the engine of the whole skill: models that stack the same direction (a lollapalooza), models that check each other (base rate vs vivid story), and models that complete each other (price, players, dynamics — one situation, several lenses).
Running the Lattice: A Case Study — a full worked decision (a bank run, read live through incentives, game theory, feedback loops and social proof at once) plus the meta-skill of choosing which models a novel situation needs.
The Checklist — and Where the Lattice Lies — turning the latticework into a written checklist and pre-mortem you run on purpose, and the honest limits: forcing a model, double-counting one effect under two names, confirmation-shopping for the flattering model, and paralysis-by-lattice.

Then a Final Exam — graded, one question at a time, one-way: once you answer, it locks. No back button, no retries, 70% to pass.

How to use this course

One habit does most of the work: before you read our analysis of any scenario, run your own lattice first. Ask what three different models say about it, out loud or on paper, then read on. And live in the decision desk — switch lenses on until you can feel the difference between a stacked reading (all agreeing, needle pinned) and a flagged tension (clashing, needle centred, something to investigate) in your hands. A latticework you’ve watched organise an argument sticks far better than one you’ve only read described.

Next up: lesson 1, where we look hard at the failure this whole course exists to fix — the man with a hammer, and why the single model you’re proudest of is the one most likely to lead you off a cliff.