What I study when I am not writing.
This is what I study when I am not writing: founders read as case files, investors read like philosophers, six hundred years of people who paid for the future, and the paintings that taught me more about conviction than any pitch deck.
I’m Pedrum. I’ve built companies and just finished an MBA. Almost everything I believe came from people who built things, people who funded them, and paintings I did not understand until suddenly I did. This room holds all three, plus the reading that feeds them.
Nothing here is comprehensive. A founder, a book, or a painting makes the cut only if it changed how I work. Some things are hidden; the colophon has hints.
Chapter I
Twenty-five founders at the fork, and the twenty-three times someone said no.
Case Files · 25
01Twenty-five builders born between 1835 and 1990, from Rockefeller at sixteen with a ledger to the Collisons with seven lines of code. Each file ends with what I took from it. Open all twenty-five and the drawer gives up its secret.
A thirteen-year-old telegraph messenger who learned to read the wire by sound, then rode the railroads, then bet everything on the Bessemer process while rivals called steel a fad. His mills ran on obsessive cost sheets; he cared less what a ton sold for than what it cost, because cost was the number he controlled. In 1901 he sold to Morgan for $480 million and then executed the stranger half of his plan: giving nearly all of it away on a system, funding more than 2,500 libraries so the next messenger boy could read his way up too.
What I take The exit is the midpoint. What you do with the second half is the biography.
Before he owned the oil business he owned its arithmetic. As a sixteen-year-old assistant bookkeeper he treated the ledger like scripture, and he never stopped. Rivals guessed at their refining costs; Rockefeller knew his to the third decimal, per barrel, per barrel-hoop, per drop of solder on a can lid. In an industry running on rumor, he was certain about his costs. That certainty let him price rivals into selling and railroads into secret rebates.
What I take Knowing your numbers colder than anyone else is a weapon.
She could not afford advertising, so she gave the product away: a lipstick into a stranger’s hand, a free touch-up at the counter, a sample in every bag whether you bought or not. The industry called it ruinous. Lauder was building a distribution system out of generosity. Every woman who walked out with a miniature sold the brand at her own dinner table, and the gift with purchase she pioneered is now standard across cosmetics. It began as the strategy of someone too poor for billboards.
What I take When you can’t outspend, out-give.
The moving assembly line cut the time to build a Model T from about twelve hours to about ninety minutes, and the price fell year after year while wages went the other way: in 1914 Ford stunned the industry by roughly doubling pay to five dollars a day. Moralists called it charity and rivals called it madness. Turnover collapsed, the best mechanics in America lined up at his gate, and his own workers became his customers. He understood that a mass market has to be paid into existence.
What I take Speed comes from removing motion.
An orphanage seamstress who rebuilt women’s clothing out of jersey, a humble knit then used mostly for men’s underwear, because it was cheap, and because women could finally move in it. She borrowed from menswear, cut away ornament, and made simplicity a luxury. The cautionary chapter matters too: in 1924 she signed away ninety percent of her perfume business to partners in a deal she spent decades fighting, and only renegotiated her way to a fortune in 1947. She invented modern luxury and nearly gave away its best asset at the same desk.
What I take Subtraction can be a revolution. Read the deal twice, especially when the product is your own name.
Hollywood called Snow White “Disney’s Folly”: a feature-length cartoon that took about three years, ran far over budget, and was financed by a bet-the-studio bank loan. It became the highest-grossing film of its era. In the 1950s the banks laughed at an amusement park, so he borrowed against his own life insurance to start Disneyland, then rewrote the company’s vocabulary around it: employees became cast members, customers became guests, shifts became the show. He understood that language is architecture. Rename the work and you change how it’s done.
What I take Build the world, and name things like the names matter.
He flew a secondhand plane low over county roads to count cars in competitors’ parking lots, walked every rival’s aisles with a notebook, and stood in his own doorways asking customers what he’d gotten wrong. Everyone discounted. Walton’s edge was refusing to let any layer of management stand between him and ground truth. Data flowed up to Bentonville every Saturday morning while his competitors waited for quarterly reports, so Walmart turned its inventory faster than anyone could copy.
What I take Consensus is a lagging indicator. Parking lots are a leading one.
In 1955 Bulova offered to buy a hundred thousand of Sony’s transistor radios, an order bigger than the company itself, on one condition: Bulova’s name on the case. Morita, whose country’s exports were still a punchline for shoddiness, said no. He would not sell the name, the one asset that outlasts every product. Decades of Trinitrons and Walkmans later, “Made in Japan” meant precision, and that single refusal did as much as any product to change it.
What I take Never sell the compounding asset to finance the depreciating one.
A dyslexic farm boy who started by selling matches from a bicycle, then seeds, then pens, then furniture by catalog. The founding accident became doctrine in 1956 when a colleague sawed the legs off a table to fit it in a car: flat packing moved assembly, shipping, and warehousing onto the customer, and IKEA turned the difference into permanently lower prices. Kamprad flew economy, drove an old Volvo, and wrote frugality into A Furniture Dealer’s Testament. He understood that cost discipline is a culture, and a culture is the only cost program that outlives its author.
What I take Every cost you remove is pricing power you keep forever. Inconvenience is a business model.
Toyota rejected his first batch of piston rings: years of work, judged scrap. He went back to school for metallurgy, remade them, and won the contract; then war and an earthquake took the factory anyway. He started over with surplus generator engines bolted to bicycles. What never changed was the order of his loves: the machine first, the market second. He tuned engines by ear, and in 1954 he announced Honda would race the Isle of Man TT; the team arrived in 1959 and won in 1961. The market forgave every setback because the machines kept getting better.
What I take Obsession with craft survives disasters that strategy doesn’t.
Twenty-five years at Texas Instruments, passed over for the top job, and then a founding at an age when most executives write memoirs. Recruited to Taiwan, he started TSMC in 1987 at fifty-five on a model nobody wanted: make everyone’s chips, design none of your own, compete with no customer. The pure-play foundry looked like humble contract work. TSMC now makes most of the world’s most advanced chips.
What I take Being passed over can be the founding event. A new business model beats a new product.
Most people know the keynotes. Look at the detours: a calligraphy class with no plan, a firing, and the NeXT and Pixar years. Together they gave Apple typography, an operating system and storytelling. None of it looked like a career. All of it became the moat. Taste accumulates in detours, and Apple’s margins are what taste looks like on an income statement.
What I take The detour you can’t justify today is often the edge you’ll monetize in a decade.
In 1976 he wrote an open letter scolding hobbyists for copying software, insisting that code, though free to duplicate, still deserved a price. That stubborn idea became an industry. The bigger deal came in 1980: IBM needed an operating system, Gates bought one from a small Seattle shop for a modest sum, licensed it to IBM non-exclusively, and kept the right to sell it to every clone maker on earth. IBM got a product; Gates got the standard.
What I take Own the license. The margin lives in the rights.
He joined Apple in 1998, when it was months from irrelevance, and did the unglamorous surgery: closed the factories, outsourced with precision, and drove inventory from months of stock to days, treating logistics with the same fanaticism the design studio reserved for corners and radii. Then he took the job of following a founder-myth and refused to impersonate him. He ran the company as an operator: services, silicon, privacy as policy, and roughly a tenfold rise. Operations, done obsessively, is strategy.
What I take Don’t play the predecessor’s instrument. Master your own.
For its first decade, Blue Ribbon Sports was a company perpetually one bank meeting from death. Knight kept his accounting job to make payroll, begged letters of credit, and lived on float. The lesson buried in his memoir is duration. The dark middle of a company lasts years longer than any founding myth admits. The founders who make it arranged their lives so they could keep going while nothing worked.
What I take Structure your life to survive the middle. The middle is most of it.
In 1984 he bought a bankrupt textile conglomerate, mills and diapers and department stores, because buried in the wreckage sat one asset the accountants had no line for: Christian Dior. He sold most of the rest. The insight that built LVMH is that a great name is stored desire, and desire has software economics: near-zero marginal cost, pricing power that rises with scarcity, and no expiry date if maintained. He then applied the discipline luxury lacked, real logistics and real balance sheets, without ever letting the customer see the machinery.
What I take Inside every distressed asset, look for the one thing that compounds. Buy that; sell the rest.
The 1997 shareholder letter reads like a dare: we will be misunderstood for long periods, we will price for market leadership rather than quarterly optics, and we will make bold bets where the odds are good but the outcome uncertain. Then he reprinted it behind every letter that followed, an annual public contract with his own time horizon. Everything distinctive about Amazon, from AWS growing out of internal plumbing to a decade of “unprofitability” spent on reinvestment, falls out of one variable set differently: the clock.
What I take Most competition dissolves the moment you change the time horizon.
NVIDIA nearly died in 1996; Huang laid off most of the company and bet the remains on a single chip. Then came roughly a decade building CUDA, a programming platform for computation nobody was asking for, while Wall Street treated it as the expensive eccentricity of a video-game company. When deep learning arrived, the only shovels in the store were his. He explained the bet constantly. People found it boring to believe him.
What I take You can announce your decade-long bet out loud. Almost nobody will bother to front-run conviction.
The founding argument was arithmetic: a rocket’s raw materials cost a small fraction of its sticker price, so the price was a habit of the industry. Rivals that had priced in throwing the machine away ridiculed reusability. The stress test came in 2008. On September 28, after three failures, SpaceX’s fourth Falcon 1 reached orbit. On December 23 NASA awarded SpaceX a space-station cargo contract, and on Christmas Eve Tesla’s rescue financing closed. The method: price the physics, and size your commitment so survival itself is the moat.
What I take First principles are a demolition tool. Use them on the invoice everyone else accepts.
Previous file: Jensen HuangNext file: Larry Page & Sergey Brin
Two graduate students who treated the web like academia and links like citations: a page mattered if pages that mattered pointed at it. PageRank put a library-science idea behind a search box. The part I keep is the humiliation: in 1999 they tried to sell the whole thing to Excite for under a million dollars, and were turned down. The market’s no forced them to build the company instead, and the ad auction they later bolted onto search paid for everything that followed.
What I take Sometimes the market refuses you into greatness. Keep a list of your rejections; one of them is a gift.
He built an online snowboard shop, hated every e-commerce tool available, and wrote his own. The snowboards were the false start; the realization that the tooling was the company was the founding. Shopify’s strategy has stayed weirdly consistent ever since: arm the rebels. Give a million small merchants the infrastructure of a giant and take a sliver of everything that flows through. It’s the rare platform-era company whose incentives point the same direction as its customers’.
What I take The tool you built out of frustration for yourself is market research.
Previous file: Larry Page & Sergey BrinNext file: Zhang Yiming
Every Western platform assumed the social graph was the product: you see what your friends share. Zhang bet the graph was the bottleneck, and that a recommendation engine reading behavior directly could out-program your friends within a week of watching you. TikTok is that bet at planetary scale, and Western feeds copied it. I watch him the way one watches weather. My essay on machines occupying time is, in part, about what he proved.
What I take Interface decisions are destiny. Whoever chooses what you see next is the real editor-in-chief.
When nobody would fund air mattresses for strangers, he sold novelty election-themed cereal to stay alive, which sounds like a joke until you see what it signals: a founder who will manufacture survival out of anything. The deeper contribution came after. Airbnb sold trust: reviews, verified identities, professional photos, and a host guarantee. Together they made sleeping in a stranger’s home feel safe. He treated trust as a design problem with components.
What I take Trust can be designed and shipped like any other feature.
Previous file: Zhang YimingNext file: Patrick & John Collison
Online payments circa 2010 meant weeks of paperwork with a bank’s gateway partner. Stripe’s founding artifact was seven lines of code a developer could paste before lunch. The Collisons picked the least glamorous layer of the internet, money plumbing, and treated it with the care other companies reserve for logo redesigns: obsessive documentation, APIs as literature, infrastructure as a form of ambition. Then they kept widening the aperture until the company behaved like infrastructure for the GDP of the internet.
What I take Pick the boring layer everyone depends on and make it beautiful. Boring plus beautiful is a monopoly.
Five thousand one hundred and twenty-six failed prototypes of a bagless vacuum, built in a coach house between 1979 and 1984. The appliance giants declined to license a machine that would make their bags obsolete, so he manufactured it himself, kept full ownership, and let the failures become the company’s culture: engineers there are hired to be wrong quickly.
What I take Iteration is a moat because it’s miserable. Count your prototypes and keep going.
You opened all twenty-five files. That is studying, and studying is how every file above got started. This one is blank on purpose. The founders in this drawer had ledgers, prototypes, refusals, and dark middles. None of them knew their file would end up in someone’s study. Neither do you.
What I take Nothing yet. The file is open. Write something worth filing.
Museum of No · 23
02Decca said it to the Beatles, and Kodak said it to its own digital camera. Morita said it to Bulova and built Sony. What separates the two kinds of no is what the refuser understood about time.
1962 · London
Auditioned the Beatles on January 1, 1962, and passed. The line attributed to the label: “Guitar groups are on the way out.”
1876 · New York
Was offered Bell’s telephone patents for $100,000 and declined. The telegraph company judged the telephone a toy, and the toy ate the company.
$100,000
1975 · Rochester
Steve Sasson built the first digital camera in Kodak’s own lab in 1975. Leadership shelved it to protect film. Kodak filed for bankruptcy in 2012.
1999 · Silicon Valley
Declined to buy Google for under a million dollars. Excite@Home filed for bankruptcy in 2001.
under $1M
2000 · Dallas
Passed on buying Netflix for $50 million and reportedly found the pitch amusing. Blockbuster filed for bankruptcy in 2010.
$50M
1982 · Hollywood
Declined product placement for a small alien with a candy habit. Reese’s Pieces said yes, and sales jumped.
1955 · Tokyo
Refused Bulova’s hundred-thousand-unit order rather than put another company’s name on his radios. Case file F-08 tells the rest.
100,000 radios
2006 · Palo Alto
Turned down Yahoo’s $1 billion at twenty-two, when his board and common sense said take it. Six years later Facebook went public at a $104 billion valuation.
$1B
1997 · Cupertino
Returned to a dying Apple and killed most of the product line, drawing a two-by-two grid and refusing everything outside it. Focus is a list of funerals.
since 1948 · California
Nearly eighty years of refusing to franchise, to sprawl the menu, or to go public. The no is the brand.
2008 · Sunnyvale
Refused Microsoft’s $44.6 billion. Within a decade Verizon bought the core business for $4.8 billion.
$44.6B
2010 · Chicago
Turned down $6 billion from Google at the peak of daily-deal mania, then watched ninety percent of its value evaporate after the IPO. Size your certainty before you size your no.
$6B
2003 · Silicon Valley
Declined Google’s $30 million in stock, worth billions today, to stay independent. Then it lost the category to a Harvard dorm room.
$30M
1990s · Ohio
Refused every Calvin and Hobbes license: no plush tigers, no film, no cereal. He walked away from a fortune estimated in the hundreds of millions. Some assets keep their value by staying unsold.
Maranello · doctrine
The house rule attributed to the old man: always sell one car fewer than the market demands. Scarcity was the product.
since 1985 · Issaquah
The $1.50 hot dog combo has not moved since 1985. A departing chief executive reportedly told his successor that raising it would be a firing offense. A price can be a promise.
$1.50
2010 · Cupertino
Published “Thoughts on Flash,” a public letter refusing Flash on the iPhone, while much of the industry called it arrogance. Adobe ended Flash in 2020.
1957–2011 · Rome
Gave almost no interviews across half a century and let the work speak. The paintings aged better than any explanation would have.
after 1964 · Monroeville
Won the Pulitzer Prize for To Kill a Mockingbird in 1961, then declined nearly every interview for fifty years. The silence protected the book from its author, and the author from the noise.
2011 · nowhere
Built a system later worth more than a trillion dollars, then walked away from roughly a million untouched bitcoins and from the name itself. A network with no founder to trust was the design.
~1M bitcoins
a century and counting
Two of America’s largest companies, private since their founding. Time horizon is the one advantage a shareholder letter cannot grant.
since 1996 · Bellevue
No IPO, no acquisition, barely an org chart, while running Steam, one of software’s great money machines. Staying illegible is a strategy.
2022 · Ventura
Yvon Chouinard gave the company to a trust and a nonprofit and declared Earth its only shareholder. The last no, said to cashing out.
Chapter II
Six hundred years of people who paid for the future.
Ten plinths from 1389 to now. Nine carry the tool a financier handed down; the tenth is open. Pick one to walk to that frame in the lineage.
Lineage of Capital · 9
03Behind Donatello’s bronze David stands a banker. Behind the computer age stands a professor with a new kind of fund. Nine financiers in six hundred years, each handing the next a tool that did not exist before, and a tenth place still open.
The bank made the money, and Cosimo spent it on Florence. He backed Brunelleschi and Donatello the way a modern fund seeds moonshots: diversified bets on unproven genius, held through ridicule, with returns paid in a changed city.
Handed down · capital as a lever on culture
He financed the 1519 election of Charles V as Holy Roman Emperor, priced sovereign risk when the concept barely existed, and ran a courier network faster than any crown’s. In 1516 he founded the Fuggerei, housing for the poor that still stands. The rent is still one Rhenish guilder a year, about 0.88 euros.
Handed down · information advantage as the core asset
From a ghetto coin shop he placed five sons in five cities: Frankfurt, London, Paris, Vienna, and Naples. The family’s couriers carried market-moving news ahead of governments. The structure itself, a distributed partnership with shared intelligence, was the innovation.
Handed down · the network as the fund
In the panic of 1907, with no central bank in existence, one private citizen locked the heads of New York’s banks in his library until they pledged the capital to stop the run. He financed Edison, consolidated railroads and steel, and showed capital’s strangest power: to be, for a moment, the state.
Handed down · the backstop
After the crash nearly ruined him, he wrote the discipline down. Security Analysis, published in 1934, turned investing into a craft with a textbook instead of a temperament. His Columbia classroom seeded the industry, and his best investment was a student from Omaha.
Handed down · the method, written down and taught
See allocator file A-01He took Graham’s discipline for buying existing value and pointed it at companies that did not exist yet. ARD, founded in 1946, was the first publicly owned venture capital firm, DEC its proof, and his Harvard students his real portfolio.
Handed down · institutional money for the unbuilt
See allocator file A-10When eight young scientists wanted to leave Shockley’s lab in 1957, Rock found the money that let them found Fairchild. He backed Intel on a page-and-a-half memo in 1968 and Apple in 1978, and is credited with popularizing the name of the job itself: venture capitalist. Talent, he showed, is the scarce asset, and capital follows it west.
Handed down · bet on teams, and move to where they are
The Sequoia founder’s heresy was “bet on markets, not founders”: a giant wave forgives a mediocre surfer. From one office on Sand Hill Road he funded Atari, Apple, and Cisco, sometimes installing the management himself. The market-first camp traces to him; the founder-first camp, to Rock.
Handed down · the market as the senior partner
He opened the door. Y Combinator replaced country-club introductions with an application form and an essay culture that treated starting a company as a teachable craft: small checks, big batches, advice in public. Thousands of companies later, the radical part is still the first part. Anyone can apply.
Handed down · the open application, and the essay as a tool of capital
Nine tools in six hundred years: patronage, the courier, the family network, the backstop, the method, the fund, the bet on teams, the bet on markets, the application form. The tenth has not been built yet.
Allocators · 10
04Every investor answers one question: what do you believe enough to pay for? Here are ten answers. Graham nearly lost everything in 1929 and turned the scar into a method; Simons banned storytelling from the building.
Margin of safety is humility expressed in arithmetic.
He was brilliant before 1929 and nearly wiped out anyway; the crash took most of his partnership’s capital while he watched. What he wrote down afterward became Security Analysis. Buy so far below conservative value that being wrong is survivable, treat the market as a manic counterparty rather than an oracle, and never confuse a good company with a good price.
Inactivity is a position, and most of the time it’s the correct one.
The public knows the compounding; the professionals study the sitting. Buffett’s edge is temperament: holding cash for years while being called a has-been (1969, 1999), then acting in size within days when the odds finally tilt. He tells students to imagine a punch card with twenty lifetime slots. The scarcity is the discipline. Almost everything, including almost everything good, has to be let go.
Invert, always invert. The exits are marked more clearly than the entrances.
A lawyer who read physics, biology, and psychology and refused to accept that finance had its own sealed logic. His latticework method hangs every decision on the big models of many disciplines and lets them argue. It is a defense system against the brain’s shortcuts, which he catalogued as twenty-five causes of human misjudgment. His favorite operation was inversion: don’t ask how to succeed. Ask what guarantees ruin, then avoid it.
You can’t predict. You can prepare.
His memos rest on one admission: you cannot know the future, but you can know where you stand. Marks reads markets like weather, never predicting the storm’s hour but noticing when everyone has stopped carrying umbrellas. First-level thinking says a great company is a great stock; second-level thinking asks what is already in the price. The pendulum between greed and fear never stops. His method is knowing which way it is swinging.
The instrument doesn’t make the return. The holder’s behavior does.
Thirteen years running Magellan at roughly 29% a year. The detail that haunts me is the finding, often repeated, that many of his own investors still lost money, because they piled in after hot streaks and fled after dips. The fund compounded; the behavior didn’t. His method was almost insultingly humble: know what you own, be able to explain it to a ten-year-old in two minutes, and notice what is in your own cart before Wall Street does.
Markets are participants in the reality they price.
He wanted to be a philosopher. Reflexivity, his idea that markets write reality as well as read it, because rising prices change the fundamentals that justify them, is a contribution to epistemology that happened to pay. In 1992, with conviction and setup aligned, he bet ten billion dollars against the Bank of England and broke the sterling peg. The philosophy chose the trade; the sizing made it history.
Conviction without sizing is just opinion.
Thirty years of roughly 30% annually and, by his account, no down year. The record rests on asymmetry of expression: tiny when uncertain, enormous when the setup is clear, and out at once when the thesis cracks. He credits Soros with the lesson that separates professionals from talented amateurs: what matters is how much you make when right and how much you lose when wrong.
The market owes your story nothing.
A prize-winning geometer who walked away from mathematics, hired physicists and codebreakers instead of MBAs, and banned narrative from the building: if the signal is real, trade it. Nobody needs to explain why. The Medallion fund’s returns, decades of them, are an uncomfortable dataset for everyone else in finance, because they suggest the stories most investors trade on are mostly decoration. I keep him around as a standing rebuke: maybe the narrative is the bug.
Time horizon is an asset class. So is working on something you love.
He left Wall Street for a university job that paid a fraction of what he was worth, then rewrote institutional investing: out of bonds, into venture, private equity, timber, and absolute return, harvesting the premium the impatient leave on the table because a university thinks in centuries. The Yale Model made billions for education, and his protégés went on to run endowments at Princeton, MIT, Stanford and others. What I underline is the career: he priced meaning into the compensation.
Back the person; the company is the byproduct.
A Harvard professor who institutionalized a new asset class: in 1957 his American Research & Development put about $70,000 into two engineers called Digital Equipment and watched it become $355 million, the first proof that a venture fund could work. His doctrine was stranger than the returns. He said he was building men and companies, in that order, and held positions for decades.
Numbers · 18
05Eighteen numbers I keep coming back to. Each one is a whole story: 5,126 prototypes, $800 for ten percent of Apple, 79 days of New Coke.
02 / 18
$70k → $355M
Dyson’s failed prototypes, 1979 to 1984, before the one that worked.
ARD’s 1957 stake in Digital Equipment, and what it became. The first proof that venture capital could work.
Lines of code in Stripe’s first integration. Cut the paperwork and you can take an industry.
Years to double your money at r percent. The pocket calculator for compound interest.
Try it in the LabSlots on Buffett’s imaginary lifetime punch card. Twenty decisions, so each one has to count.
Try it in the LabWhat Ronald Wayne accepted for his ten percent of Apple in 1976. Position sizing includes the positions you leave.
Medallion’s reported gross annual return, 1988 to 2018. The fund ran on models alone.
Kevin Kelly’s true fans: enough paying supporters to make a living, by his 2008 math.
The year an undertaker, Almon Strowger, patented the automatic telephone exchange because he suspected the human operator served someone else. My essays keep returning to it.
Chanel’s share of her own perfume after the 1924 deal. She spent two decades winning her name back.
Days New Coke survived before the company surrendered on live television. Two hundred thousand taste tests, each one a single sip.
Try it in the LabBlockbuster stores at the 2004 peak, and Blockbuster stores now. The last one is in Bend, Oregon.
Kroc’s service fee on McDonald’s sales, the sliver that nearly sank the company until Sonneborn found the rent underneath the restaurants. The true business hides under the obvious one.
Try it in the LabLetters to Marilyn vos Savant about three doors, about a thousand signed by PhDs. Most said she was wrong. She was right.
Try it in the LabConsecutive profitable years at Southwest, 1973 to 2019. Bought with a stopwatch and a ten-minute turn.
Try it in the LabThe Dutch East India Company sells shares to anyone in Amsterdam, and the shares start to trade. Thirty-five years later came the first famous bubble: tulips, 1637.
Employees at Instagram when Facebook paid $1 billion for it in 2012. About $77 million a head.
Folds of a sheet of paper 0.1mm thick. Fold 27 passes Everest. Fold 42 passes the Moon.
Try it in the LabChapter III
The paintings that taught me more about conviction than any pitch deck.
The Gallery · 10
06Critics said a child could do it. It takes most of a lifetime to get back to the mark a child makes.
American, 1928–2011 · Lexington, Rome
Untitled (New York City), 1968
Oil-based house paint and wax crayon on canvas
The patron saint of this website. He left America at the height of Abstract Expressionism, moved to Rome, and spent fifty years making paintings that look like a wall a poet leaned against: the names of Achilles and Apollo, loops of pencil like handwriting with the language removed, chalk-white lines on slate-dark grounds. Critics called it graffiti. In 2010 the Louvre gave him a ceiling. He showed that a gesture can carry as much intelligence as a sentence, and that knowing how little to put down is the last skill acquired.
Identical outside. No two interiors alike.
American, 1928–1994
100 untitled works in mill aluminum, 1982–1986
Mill aluminum, each 41 × 51 × 72 in
Identical outside, and no two interiors alike, set in two former artillery sheds. He wrote against decoration, bought ranch land and buildings in Marfa, and installed the work for good.
In the model, the interiors are drawn from 12 illustrative types, and the split between the two sheds is approximate.
American, 1903–1970
No. 61 (Rust and Blue), 1953
Oil on canvas
Two rectangles of color, and grown adults cry in front of them. He insisted the paintings were about tragedy, ecstasy, doom, not color relationships, and hung them low and close so you would fall in.
Canadian-American, 1912–2004
The Tree, 1964
Oil and pencil on canvas, 6 × 6 ft
A grid ruled by hand in pencil: perfection attempted and never quite reached, on purpose. From across the room it dissolves into weather. She lived alone in the New Mexico desert and called the work discipline in the service of joy.
American, 1960–1988
Untitled, 1982
Acrylic, spray paint and oilstick on canvas
He crossed out words so you would read them harder, his own trick, stated openly. Anatomy diagrams, crowns, lists of kings: the canvases read like a mind annotating itself at full speed. He died at twenty-seven. In 2017 this skull sold for $110.5 million.
American, b. 1974
COMPANION, first made 1999
Painted vinyl, later bronze and inflatables
He started by repainting bus-shelter ads without permission. The crossed-out eyes moved from bus shelters to vinyl toys to museum floors without changing.
OCT. 12, 2026
Japanese, 1932–2014
Today series, 1966–2013
Liquitex on canvas
Nearly five decades of Date Paintings: the day’s date, hand-lettered in white on a monochrome ground, finished by midnight or destroyed. No expression, no image, just the fact of another day recorded. What is your daily mark, and would you make it public? The date on this label is today’s. It repaints itself for every visitor.
French, 1928–1962
IKB monochromes, from 1957
Dry pigment in synthetic resin on canvas
He mixed a blue so saturated it seems to hum, registered the formula as International Klein Blue in 1960, and sold canvases of nothing else. The color was the brand, and the brand was the argument: ownership of a feeling. He also sold empty space for gold and threw half the gold in the Seine.
American, 1923–2015
Window, Museum of Modern Art, Paris, 1949
Oil on wood and canvas, two joined panels
He found his shapes in the world’s leftovers: the shadow under a bridge, the curve of a window, a shirt on a line. Then he painted just the shape, one color, no story. Seeing is a form of editing, and the confidence to show one shape is rarer than the ability to show many.
German, b. 1932
Abstraktes Bild (809-4), 1994
Oil on canvas
Photographs painted and blurred as if memory smudged them; then, from the same hand, abstractions dragged into being with a squeegee the size of a door. Refusing to pick a style was the style.
Cabinet of Taste · 8 places
07Places worth the trip, films about ambition, and the music these essays were written to.
Eight places, seven for art and one for capital.
Ten films about building things and what it costs.
Eight recordings for long paragraphs, mostly instrumental and patient.
The Canon · 18 books
08Eighteen books. Under each, the idea I kept.
An operating system for decisions: Munger’s speeches, his checklists, and his twenty-five causes of human misjudgment.
The idea I keep: borrow the big models from every discipline and let them argue about your decision before you make it.
Chapters 8 and 20 carry the book: the market as a moody business partner, and the margin of safety.
The idea I keep: price is a quote. You’re allowed to hang up the phone.
Rockefeller as a system: ledgers, silence, Baptist thrift and ruthlessness. Chernow, 1998.
The idea I keep: mastery of detail scales into mastery of industry. There is no shortcut through the arithmetic.
Compounding as autobiography: money, yes, but also relationships, reputation, and regret, all rolling downhill and gathering mass. It does not hide the costs.
The idea I keep: everything in a life compounds, including what you neglect.
The founder memoir with the anesthesia turned off: a decade of near-death, debt, and doubt, written by someone who remembers being terrified.
The idea I keep: the dark middle is most of the story. Build a life that can afford to stay in it.
How a defeated country’s engineers rebuilt its name one transistor at a time, and a long lesson in refusing short money that costs you the long asset.
The idea I keep: a brand is a nation-sized story told through the smallest details of a product.
Amazon’s DNA sequenced: regret minimization, customer obsession as a forcing function, and the strange power of a company that treats its own margins as a public menu for competitors.
The idea I keep: your margin is someone’s opportunity. Ideally yours, before it’s theirs.
A contrarian catechism, built to be argued with. Nobody leaves neutral.
The idea I keep: competition is a tax on sameness.
Management writing that starts where the frameworks end: layoffs, demotions, fear at 3am. Horowitz names the struggle, and naming it takes away some of its power.
The idea I keep: there is no formula for wartime. There is only the practice of deciding anyway.
Pixar’s president on why every Pixar film was bad for years first. The machinery that protects ugly early drafts from premature judgment is the studio.
The idea I keep: build a braintrust that critiques the work, never the person, and let candor outrank hierarchy.
Twenty short essays arguing that money is mostly behavior.
The idea I keep: wealth is what you don’t see: the cars not bought, the exits not taken.
The grave dancer’s memoir: a career built on buying what everyone else was fleeing, from distressed real estate to entire industries, told with zero interest in being liked.
The idea I keep: when everyone is looking right, look left.
The bug report for the human mind, filed by the person who found the bugs. Read slowly; you will catch yourself committing each chapter in real time.
The idea I keep: the confident voice in your head is a press secretary.
Written about television in 1985; reads like a code review of the feed. Postman saw that the medium becomes the culture.
The idea I keep: every medium smuggles in an epistemology. Ask what a technology makes easy to believe.
The prequel to everything I write about: two centuries of the harvest and resale of human attention, from penny press to platform. The pattern repeats so cleanly it’s almost funny.
The idea I keep: wherever attention pools, an industry forms to drain it. Guard the reservoir.
A fable short enough for an afternoon and heavy enough for a lifetime: the revolution succeeds, and then the commandments on the barn wall start getting edited at night. Orwell shows power capturing the language before it captures anything else.
The idea I keep: whoever controls the definitions controls the farm. Reread it before reading any regulator’s press release.
Why things catch on, in six parts: social currency, triggers, emotion, visibility, usefulness and story.
The idea I keep: word of mouth can be engineered. Design the retelling.
A head of state journaling at night, for himself, with no audience intended, which is why it works. Nearly two thousand years of executives have found their own worries answered in advance.
The idea I keep: you don’t control events; you control the account you give of them.
The best business school is free. Five archives, and where to start in each.
The essay culture this site descends from. The archive is free and long; these are the six I hand people first. The notes are mine; the essays are his.
Buffett’s letters, nearly sixty years of them, in plain English. The free archive starts in 1977 on a site that has barely changed since 1997. Start with these five.
The books are the polished lectures; the memos are the lab notebook, written to clients as cycles turned. Buffett has said a Marks memo is the first mail he opens.
Twenty-four shareholder letters, with the 1997 letter reprinted behind each one that followed. Read them in order and you watch a bookstore become an everything store, then infrastructure.
The writers and rooms I return to on a schedule. Not endorsements; dependencies.
Chapter IV
09 The Desk
Twenty-seven lines that never made it into an essay, filed under five subjects.
On attention01 / 27
Attention is the last scarce resource. Making and sending content now costs almost nothing, so everyone competes for the same few hours of yours.
When a decision will not simplify, I reach for a word. Most are older than any framework.
The essays are arguments I finished. These are the ones still open. If you have an answer, write to me.
Experiments · 47 machines
10Each machine teaches one idea by making you decide.
New in the Lab
Four of these rooms were games. They live in the Lab now.
One floor per wing, Wing I at the top. Choose a machine to go straight to it.
The Blackboard
11In the late sixties Twombly painted loops of white line on slate-dark grounds, handwriting with the language removed. Here is a small board of your own, with white chalk, orange chalk, a date stamp, and a wipe. Nothing is saved unless you keep it.
Colophon
12P.S. stands for postscript, and for Pedrum Shokouhi. The essays are the letter. This room is what comes after.
Five things are hidden here. Type a famous video-game code, or type Twombly’s name. Open every case file. Look in the developer console. Leave the tab and watch its title. On a phone, press and hold the P.S. mark instead of typing the code.
If you read one thing here, make it an essay. This room is the margin around the essays.
The company analyses live in Investing.
