Ideas you can touch
Reading about an idea is one thing; holding it is another. Forty-three machines in twelve wings, each an essay in disguise. Nothing here is decorative, and one of them will remember you tomorrow.
“THE MARQUEE”
The machines people talk about afterward: your whole life in squares, a game-show host with a secret, a casino wheel that devours doubling systems, paper folded past the Moon, the universe on one slider, and a bomber full of holes.
Your life in weeks
A 90-year life fits in the grid below, one square per week. Set your age. The dark squares are spent. The pulsing one is this week, the one you are currently spending on this website.
Every plan you are postponing lives in the cream squares. There are fewer of them than the feed wants you to notice.
The three doors
In 1990 a columnist named Marilyn vos Savant printed the correct answer to this puzzle and roughly ten thousand readers wrote in to call her an idiot. About a thousand of the letters were signed with a PhD. Every one of them was wrong. You are about to play the exact game they argued about, against a host who knows where the prize is and never forgets a round.
“Three doors. Behind one, a 100× return. Behind two, nothing. Pick, and I’ll show you a little secret about one of the others.”
Why switching wins, explained so a five-year-old nods
You pick door 1. Your door holds 1 chance in 3. The dashed box, doors 2 and 3 together, holds the other 2 chances in 3. Nobody argues with this part.
Monty opens door 3. Empty, and he knew it; he was never allowed to open yours or the prize. The box does not lose its 2⁄3 just because he showed you which half of it was junk. All of it now sits on door 2.
So the choice was never “two doors, fifty-fifty.” It is: keep 1 door, or take both of the others. Switching means taking both doors, and Monty has already opened the empty one for you.
Two rounds out of three, your first pick was wrong, and on exactly those rounds the door Monty leaves shut is the prize. When Marilyn vos Savant printed this in Parade in 1990, ten thousand readers wrote in to correct her, a thousand with doctorates. The letters stopped once people played the game. The tally above does the same job.
Still arguing? Play it with a hundred doors.
Same rules, bigger room. One prize, ninety-nine empties. Pick any door.
“Go on. Any of the hundred.”
Nobody who watches Monty tiptoe around one door out of a hundred still believes it is a coin flip. Three doors is the same trick with the volume turned down.
Six in a row
Monte Carlo, August 18, 1913. Black hits ten times running and the room leans in. By fifteen, gamblers are stacking francs on red, because after all that black, red is due. Black came up twenty-six times in a row that night, and the casino ended it millions of francs ahead, paid for by the law of averages. Below is that table, rebuilt: a European wheel, one green zero. The system every gambler swears by is armed and ready. Watch for six of one color, bet the other, and double after every loss so your first win puts you ahead. You have $1,000 and $10 chips. Find out who ends up with the money.
“Table’s open. The system needs six in a row before a single chip moves.”
Seven losses deep you have pushed $1,270 across the felt and the system asks for $1,280 more, all to win the ten dollars you wanted at the start. A $1,000 bankroll dies on rung seven. And the green zero belongs to nobody: not red, not black, not odd, not even. When it lands, every bet on the table loses at once.
After six blacks, the chance of red is 18 in 37. After twenty-six blacks, it is 18 in 37. The wheel does not know what it just did, and doubling your stake only chooses how you lose: often and small, or once and completely.
Fold to the Moon
Take a sheet of paper, a tenth of a millimeter thick. Fold it in half. Again. The slider goes to 42, and the answer to where that lands is genuinely hard to believe.
One sheet of paper. Every fold doubles it. This is the same math as your savings account, your habits, and a pandemic.
A billion seconds
You know the words million, billion, trillion. Your brain files them next to each other like shirt sizes. Count them out loud, one number per second, and see when you finish.
A million seconds ago was last week. A billion seconds ago, the nineties. That gap is the entire difference between a millionaire and a billionaire.
Powers of ten
One slider, forty-one orders of magnitude. Drag from the inside of a proton to the edge of everything, and notice where you live on the ruler.
You are here. Hello.
A small universe
Conway’s Game of Life: four rules, no ruler. Paint some cells with your cursor, press run, and watch complexity do what it always does when nobody stops it. Markets, cities, and gossip all work this way.
Draw on it. The rules don’t care what you meant.
The armor problem
1943. Bombers come back from Europe covered in bullet holes, mapped below as dots. Steel is heavy, so you can armor only one part of the plane. Choose.
Diagram: Martin Grandjean, McGeddon, Cameron Moll · CC BY-SA 4.0
Look carefully at where the holes are. Then think hard about which planes you are looking at.
“THE BOARDROOM”
Four true stories you can play: a fifty million dollar email answered with a laugh, an airline saved by a stopwatch, a burger chain that is secretly a landlord, and the most expensive sip in history. You make the calls they made.
The fifty million dollar email
Dallas, autumn 2000. You run Blockbuster: nine thousand stores, sixty thousand employees, and a river of late fees. A money-losing DVD-by-mail startup flies in and offers to sell itself to you for fifty million dollars. Its name is Netflix. You are about to face every fork Blockbuster faced, in order, with the information they had. See if you can beat the timeline.
Every scene in this machine really happened. The laughter in round one is documented.
The ten minute turn
Texas, 1971. Southwest Airlines is one year old, owns four planes, and is out of money. To make payroll, the board sells the fourth plane. Then comes the decision that builds the most profitable airline in history: they keep flying the four-plane schedule with three planes. You are head of ground operations. A plane only earns money while it is in the air. Make the math work.
The industry takes 45 minutes to turn a plane at the gate. At that speed, your three planes fly 24 legs a day and the schedule promised 32. Drag the slider and find the airline.
You are not in the burger business
Chicago, 1956. Ray Kroc’s young McDonald’s is franchising fast and nearly broke; the $950 franchise fee and a 1.9 percent royalty barely cover the lights. A former Tastee-Freez finance man named Harry Sonneborn walks in with an idea so good it becomes the company. Before you meet it, take the quiz.
Where does McDonald’s actually make its money today?
The sip test
Atlanta, 1984. Pepsi is beating you on television with filmed blind taste tests, and Coca-Cola is losing share for the first time in memory. Your lab has a new, smoother, sweeter formula that wins blind sips against Pepsi and against your own century-old recipe. You have run more taste tests than any company in history. Run them yourself, then make the call they made.
The counters are waiting. Your researchers have the cups poured.
Every metric is a proxy. A sip is not a can, a can is not a habit, and a habit is not love.
Final round
The four stories above are calls somebody really made. A final round at a strategy firm is where they test whether you could: a few conversations that pretend to be about clients and are really about you. Here there are four: a cinema chain whose profit fell while its revenue rose, a European grocer deciding whether to cross the Atlantic, a buyout fund with a week to price a chain of animal hospitals, and a startup with no idea what to charge for the thing it built. You play the question back, build the issue tree, pull the exhibits, do the math in your head, and give the answer first. You are scored on structure, analytics, judgment and communication, and the scorecard tells you exactly why. The firm, the clients and the case numbers are fictional; the public facts in each case file are not.
Setting up the interview room.
Nobody at the firm needs you to know how cinemas work. They need to watch you find the one number that matters, say what it means, and stop talking.
“CAPITAL”
Machines about money: compounding, shrinking, streaks, herds, a short that can ruin you, and an option that melts while you hold it.
The compounding machine
Two people start today. One gets 1% better every day, one gets 1% worse. Drag the days and watch the gap become the whole game.
The curves are identical for weeks. Then they aren’t. That flat early stretch is where almost everyone quits.
The shrinking hundred
Inflation never mugs you; it pickpockets. Set a rate and a number of years, then watch what your hundred quietly becomes while it sits somewhere feeling safe.
Rule of 72: at 3%, prices double every 24 years. Your cash does not.
The honest coin
A perfectly fair coin, flipped a hundred times. Watch how long the streaks run, then remember every market chart you’ve ever been told a story about.
Press the button. The coin has no memory; you do. That’s the problem.
The patience market
A market that only exists on this page. Buy in, watch it breathe, and sell whenever your nerve runs out. One catch: after you sell, the chart keeps going without you, and it will tell you exactly what leaving cost.
Temperament, measured in seconds. Most people don’t last a minute.
Venture odds
Most startups return nothing; a rare few return everything. Choose how many bets to spread your fund across, then run a thousand parallel futures and see how your decade could have gone.
One bet is a lottery ticket. Twenty is a strategy. Same odds, very different sleep.
The herd
Two hundred and forty traders, each watching their neighbors. The dial controls how much they care what everyone else is doing. Turn it up slowly and watch a calm market learn to stampede.
At low herding, the crowd cancels itself out and the price breathes. Push past 70 and individuals stop existing.
The rich get richer
A network grows one newcomer at a time. In a fair world, newcomers connect at random. In the real one, they connect to whoever is already connected. Run both and watch what the second rule does to the map.
This one rule, preferential attachment, explains platform monopolies, citation counts, follower graphs and the size of cities. Nobody planned any of them.
The generous coin
I flip a fair coin. Heads, you win $150. Tails, you lose $100. One flip, real stakes, decide now.
The math says this coin pays you $25 per flip, on average. Your amygdala is not consulting the math.
The short squeeze
You are short 100 shares of a fashionable stock at $100, with $10,000 in the account. The trade is probably right. The word probably is about to cost you sleep. Cover whenever you like; your broker will cover for you if you wait too long.
A long position can lose what you paid. A short can lose more than you have. Feel the difference.
The option clock
You just bought a 30-day call option: the right to buy a $100 stock at $105. It cost $320, and most of that price is not the stock; it is time. Press start and watch what time does to it, day by day.
The gray part of the bar only shrinks. The seller of this option is at the beach, letting the clock do the work.
“THE MIND”
Your brain ships with known bugs. These machines introduce you to yours, personally.
The marshmallow clock
The famous test, adapted for adults with cursors. Press the button and hold it for thirty full seconds. That is the entire game. I will be here, talking, the whole time.
Four-year-olds who managed fifteen minutes did better at basically everything. You get thirty seconds. Begin whenever you feel brave.
The reaction tax
Click the box, wait for it to turn orange, click again as fast as you can. Then we’ll talk about what your number means for anyone who thinks they can out-trade a machine.
A fast human needs about a quarter of a second. Exchanges resolve orders in microseconds. Do the division before your next hot trade.
The gauntlet
You liked the bomber? Here are five more rooms like it. Five scenarios, each with one correct exit and several confident wrong ones. Your reasoning is the only equipment. Score at the end.
Room one is a pitch meeting. In you go.
“OTHER PEOPLE”
Trust, coincidence and contagion. Everything that spreads lives in this wing.
The trust machine
Ten rounds against a hidden opponent. Each round you both choose: cooperate or betray. Both cooperate: 3 points each. You betray a cooperator: 5 for you, 0 for them. Mutual betrayal: 1 each. The machine has a strategy. Find it.
Round 1 of 10. The machine moves when you do.
The birthday party
How many people until two of them probably share a birthday? Everyone guesses around 180. Drag the slider, watch the odds, then actually throw the party and meet the twins.
Twenty-three people. Coin-flip odds. Your intuition just filed a complaint; overrule it.
The threshold
A forest, a match, and one dial: how densely the trees are planted. Below a certain density, fires starve. Above it, everything burns. The scary part is how thin the line between those two worlds is. Set the density, plant, then click anywhere to strike the match.
Mathematicians call the line near 59% the percolation threshold. Epidemiologists call it R₀. Marketers call it going viral. Same fire.
“THE STADIUM”
Sport is a laboratory where cowardice has a scoreboard. A penalty you must guess, a goalie you must pull, and a fourth down the math already solved.
The penalty spot
You are the goalkeeper. Ten penalties are coming. Before each shot, commit: dive left, dive right, or hold the middle. Real shooters put roughly a third of penalties straight down the middle, and real keepers almost never stand still. Let’s see why.
The crowd is watching. That fact is about to make your decisions worse.
The pulled goalie
Hockey, down one goal, clock running out. Pulling your goalie for an extra attacker helps you score and gifts them an empty net. Tradition pulls with about a minute left. Choose your moment, then play five hundred games and compare yourself to tradition.
Losing by one and losing by two pay exactly the same: nothing. Once you believe that sentence, the goalie comes out earlier.
Fourth down
Three fourth-down calls, the kind NFL coaches face every Sunday. Make each call, then see what the expected points actually said. For fifty years, nearly every coach got these wrong on purpose. Your job is to find out why.
Your headset is on. Sixty thousand people will boo the correct answer.
“TIME”
The only budget that never refills. One memory test and one dial you control.
The vanishing number
A number appears for a few seconds, then disappears, and you type it back. Every round adds a digit. Science says most humans hit a wall around seven. Prove science wrong.
Phone numbers are seven digits for a reason. Your working memory was the spec sheet.
The delegation dial
Turn the dial and hand your daily choices to the machine, one at a time. Perfectly convenient at every notch. See how it reads at the end.
All yours. Heavy, isn’t it.
“TWENTY SLOTS”
One machine that keeps score for life. You had to walk the whole lab to earn it.
The punch card
You made it to the bottom of the lab, so here is the only machine that keeps score for life. Buffett tells students to imagine a card with twenty lifetime decisions. This one is yours. Punching a slot is permanent; this page will remember. Spend one only for something real you’ve decided this year.
No undo. That’s the point.
“THE PIT”
The lab was finished. Then somebody wheeled the roulette table back in after closing, swearing this time nobody would double anything; every bet would simply cover the damage plus ten. One machine. Watch the rule double itself anyway.
The covered ladder
Everyone who watches the doubling ladder in Experiment 03 die assumes the doubling was somebody’s reckless choice. This table lets you make a more sensible-sounding demand instead. Wait for six of one color, bet $10 on the other, and after every loss follow one iron rule: the next bet must cover everything lost so far, plus $10 profit. No doubling anywhere in that sentence. Now watch the sentence do arithmetic. Lose $10, so the rule asks for $20. Lose that, you are down $30, so it asks for $40. Then $80. Then $160, $320, $640. Every bet comes out exactly double the last one, because each bet roughly equals the whole hole, so the hole doubles, so the bet must. Nobody chose to double. Coverage chose it. The martingale is not a strategy someone invented; it is the cheapest possible ladder that keeps your promise, and this machine proves it to the dollar. Then visit the bankroll menu below the table. A bigger bankroll never fights the wheel; it only chooses which rung kills you, how rarely, and at what price, while you post all of it as collateral to chase ten dollars.
“After hours. Same wheel, same six-in-a-row rule, one iron demand: every bet covers everything lost, plus ten.”
Every rung keeps the promise: win anywhere and the whole ladder ends exactly $10 ahead. And read the stakes out loud: 10, 20, 40, 80, 160, 320, 640. They are the doubling ladder from Experiment 03, digit for digit, because covering the hole plus ten is doubling, whether you meant it or not. Doubling the bankroll buys one more rung: the busts get rarer, the craters get deeper, and the average gets slightly worse, because deeper ladders push more money across the felt and the wheel taxes every dollar of it. Staying power multiplies the sign of your edge, and on this felt the sign is negative. The green zero still belongs to nobody.
Demand that every bet covers the hole plus ten and you have written the doubling ladder in different words. There is no gentler ladder that keeps the promise, and no ladder of any shape that moves 18 in 37. The wheel takes 2.7 cents of every dollar from all of them.
“THE CORNER OFFICE”
Machines where you run the whole firm. A leveraged fund in the week the trade turns, and an acquisition engine hunting tired software. Being right is not enough.
Conviction: the margin call
You are the founder and chief investment officer of an AI-focused hedge fund. The thesis made you famous: long the machines that build intelligence, short the software it obsoletes. It returned triple digits and you levered it, because conviction plus leverage is how legends get built. It is Monday, before the open, and three things are waiting on your desk: an overnight market briefing, a memo from your risk officer, and a notice from a prime broker who has decided your positions deserve more collateral. You have six trading days. Being right is not enough; you also have to stay solvent, liquid, and in control long enough for right to matter. Inspired by a real fund’s reported July 2026 drawdown; everything playable below is a labeled simulation, not a trading record.
Case file · the real story and where the fiction begins
Reported facts: Reuters reported that Situational Awareness, the AI-focused fund founded by Leopold Aschenbrenner, lost about 67% in July 2026 during an AI stock rout while remaining up roughly 80% for the year, sold substantial public equities with Citadel among the buyers, and removed its leverage (Reuters, July 31, 2026; Reuters on the Citadel purchase).
Everything else here is fiction: the tickers, prices, brokers, characters, dialogue, financing terms, deadlines, and every number the game generates. The starting book, margin rules, and market paths are labeled model assumptions, not a reconstruction. Your playthrough is a counterfactual; the game does not force the reported 67% on you, and no outcome here says anything about what the real fund should have done.
Model assumptions
Synthetic book: $2,800M long across six AI-infrastructure names, $600M short software, $100M private, $100M central cash, $600M short proceeds held at brokers, $2,000M financing debt, NAV $1,000M. Gross public exposure 3.4x NAV, net 2.2x, debt 2.0x; these are three different numbers. Positions, proceeds, and debt split across two fictional prime brokers. Illustrative maintenance: 20% on longs, 30% on shorts, add-ons disclosed when applied. Sale proceeds count as collateral at the executing broker immediately but transfer elsewhere next day. Prices follow a seeded two-factor model (market and AI sentiment) with regime shifts; your orders pay spread and impact that grow with size, scarcity, and stress, and large positions cannot exit in one tick. Same seed plus same actions reproduces the same week.
The trade and the fund are different organisms. The trade needs to be right eventually. The fund needs to be alive on Thursday at 3 p.m.
The acquirer
There is a company in Milan that got rich doing the least glamorous thing in software: buying beloved, tired digital products that their owners stopped investing in, and running a ruthless playbook on each one. Replatform onto shared central technology. Cut the costs nobody will miss and some that they will. Raise the price on the users who stayed because they love the thing. Repeat, with the cash flows funding the next deal. You now run a fictional version of that machine. Eight years, $50M of your own money, a debt facility, and a market full of faded icons. The levers are real trade-offs: every price hike taxes loyalty, every cost cut risks quality, every replatform costs money now to make money later, and every new asset makes the shared platform cheaper for all the others. Build the empire without letting the debt build a case against you.
On the block this year Your portfolio · set the playbook, then close the booksCase file · the real playbook this is modeled on
The real story: Bending Spoons, founded in Milan in 2013 and led by CEO Luca Ferrari, buys mature digital products with loyal audiences that their owners under-invest in, then applies a standard playbook: migrate the product onto its central technology platform, cut operating costs deeply, optimize pricing and paywalls, and run product, marketing, and technology through centralized teams. Acquisitions include Evernote (2023), Meetup, WeTransfer (2024), komoot and Brightcove (2025), Vimeo (agreed 2025, about $1.38B), and AOL (agreed 2025), with deals financed substantially by debt facilities; the company raised funding in 2025 at a valuation around $11B. Everything you can click above is fictional: the companies, numbers, elasticities, events, and outcomes are model assumptions built to teach the shape of the strategy, not a record of any real transaction.
Everyone wants to build the next new thing. The Milan playbook is a bet that the last thing, bought right and run hard, compounds faster. The catch is printed on every lever above: the playbook works because the users love the product, and every pull of the lever spends a little of the love.
“THE FOUNDRY”
Machines where you build the firm before you get to run it: two Gilded Age fortunes made on the cheapest gallon and the cheapest ton, a startup with twelve weeks and one number, and an AI lab one model behind the frontier.
The titan’s ledger
Two Gilded Age fortunes on one engine. Play John D. Rockefeller from a Cleveland refinery in 1868 to the Supreme Court in 1911, or Andrew Carnegie from the steel works he bet everything on in 1872 to the price he wrote down for J.P. Morgan in 1901. Almost every crossroads is a real one, and sourced: the secret rebates, the Cleveland buyout, the pipelines, the trust, Homestead, the Mesabi ore, the sale. The levers are the ones they pulled (cost, scale, price, integration, rebates, labor, law), and the roads they did not take are open too. The machine keeps two ledgers, and only one of them is money.
Opening the ledgers.
Kerosene fell to about six cents a gallon and rails to under twenty dollars a ton, and every line in the other ledger is also true. History keeps both books open. So does this machine.
The batch
Y Combinator gives every company in a batch $500,000 on two SAFEs, three months, and one number to watch every week: how fast you grew. Here you get an idea (three to choose from), a cofounder (or not), twelve weeks, and a group partner who has watched every mistake you are about to make. Talk to users or build, launch ugly or launch late, take the early check or wait for Demo Day, have the hard conversation or let it ride; then raise, hire, and find out whether you are default alive. The program facts and the dilution math are YC’s own and sourced; every company and person in it is fiction. Make something people want, and do the arithmetic before you sign.
Clearing a desk for the batch.
Twelve weeks is long enough to build the wrong thing beautifully, and long enough to find ten people who would be very disappointed if you vanished. Demo Day only grades the second one.
Frontier: the scaling bet
You run Aldermoor, a fictional AI lab one model behind the frontier, with $5.0B in the bank, 40,000 rented GPUs, and eight half-years to decide what it becomes. The levers are the industry’s real ones: how much compute goes to the training run and how much to the research that makes the next run cheaper, whether to train a bigger model or a leaner one that is cheaper to serve forever, which compute contracts to sign, what to charge per million tokens, how much safety work each new capability level demands, and whose money to take on what terms. The rivals do not wait and the scaling laws do not negotiate. Staying at the frontier is a spending decision; staying trustworthy there is a series of them. The lab, its rivals, and every number are fiction; the case file keeps the real history of OpenAI, Anthropic, and the rest separate, sourced, and neutral.
Booking the compute.
The scaling law keeps every promise it makes: tenfold the compute and it hands you the same step, every time. It never says who pays for the tenfold, or what you promised before you got there.
“THE LAST MILE”
Two machines where winning the argument is only the start: a campaign that needs its votes in the right states, and a therapy made from each patient’s own cells that still has to survive the factory, the price and the drive to the hospital.
Two seventy
Two invented parties, one real electoral college: 538 electors apportioned by the 2020 census, 270 to win, Maine and Nebraska splitting theirs by district. You run one ticket through the nine weeks before Election Day with a finite war chest, a candidate who can be in three places a week, and polls that are honest, noisy, and sometimes wrong together in ways nobody can see until the count. You choose between persuasion that fades in days and turnout that takes weeks to build, survive the debates and the October story, and then sit through an election night where the early lead mostly tells you who counted first. You do not need more votes than the other side; you need them in the right places, and the right places keep moving. The parties, candidates and map politics are fiction; the rules, the arithmetic and the history in the case file are real.
Printing the map.
The college never asks how many people agreed with you. It asks where they live, one state at a time, and stops listening at 270.
The living drug
A CAR-T therapy is not a pill. It is a batch of one patient’s own T cells, collected in a hospital, flown to a factory, given a new receptor, grown, frozen, flown back, and returned to the vein they came from; and the therapy only exists if every step in that sentence holds. You run the program: choose the antigen, design the receptor one domain at a time, build the factory, escalate the dose through a phase 1 trial where the toxicity you cause is the same biology as the cure, then argue about the price and try to reach the people who actually qualify. Every lever below is a trade-off somebody has really argued about in a design review, and none of them has an obviously right answer. This is educational only: it is not medical advice, the numbers in the model are labeled assumptions, and nothing here should inform a decision about a real person.
Warming the incubator.
A pill is a molecule. This is a living thing with a delivery schedule, and it is usually the schedule that kills the program.
That’s every machine.
You walked all twelve wings.
The essays these toys came from live at the front door, and the rest of the collection is in the world. Come back any time; the chalk keeps, the punch card remembers.