By then he was four years deep into one problem and nobody else was in there with him.
In 1988 Richard Lenski started a count in generations rather than prototypes. He put 12 identical populations of E. coli bacteria into 12 flasks in a Michigan State lab.
Day after day, someone has moved a drop from each flask into fresh broth.
The broth holds sugar the bacteria can eat, plus a second food called citrate they cannot.
For 30,000 generations none of the 12 populations could touch the citrate.
Around generation 31,500 one of them evolved the ability to eat it.
“It is not enough to be busy. So are the ants. The question is: What are we busy about?”
Henry David Thoreau
Two men lay wounded on the same field and the one with rank was carried first.
A man who could have been saved bled out behind a man whose wound could have waited.
Dominique-Jean Larrey served as surgeon-in-chief of Napoleon’s armies through 25 campaigns and 60 battles. He had more wounded men than surgeons, so the order was the only thing he controlled.
He threw rank out and sorted by injury instead, regardless of title or nationality:
The gravely wounded who might recover went first
Soldiers who could survive a wait, waited
Those whose wounds were likely fatal were treated last
Alexander Kotov wrote down a method for choosing a chess move in 1971. Before you move, build a list of the moves worth considering. Work through each one until you reach a definite conclusion on it.
His instruction before any move: look at the whole board, rather than the square where the last one landed.
So I pulled up the bets we could make, ranked against each other. Near the top was a PS: one line at the bottom of the emails we already send, going out millions of times a year. The playbooks needed building from scratch, so they landed well down the page.
Henry Singleton played blindfold chess at near-grandmaster level as an old man. He ran Teledyne without a five-year plan.
Teledyne started in 1960 on $225,000 of Singleton’s own money and compounded at 20.4% a year against the S&P’s 8%. He gave analysts no guidance and held no quarterly calls.
“My only plan is to keep coming to work. I like to steer the boat each day rather than plan ahead way into the future.”
Henry Singleton, to BusinessWeek
I re-rank my list often. Singleton steered each day against conditions he said nobody could predict.
The more options you rank, the better the one you pick.
He writes the first draft of each book by hand. A pen is slow enough that he can think a sentence all the way through. Then he types the whole thing up.
Once a month I write my principles out by hand. I copy the ones that still fit onto a fresh page, in order. The newest one lands at the bottom.
Lucretius wrote a poem saying the world is atoms and empty space, the soul dies with the body and no god is watching. It sat almost unread for more than a thousand years.
Poggio Bracciolini pulled a manuscript out of a monastery in 1417. He lent it to Niccolò de’ Niccoli, who copied the whole poem out by hand. The Swerve is Stephen Greenblatt’s account of the find.
Poggio’s manuscript is lost. Niccoli’s copy survived. More than 50 later manuscripts descend from Niccoli’s copy.
13 years later, my oldest principles are still on the page.
Peter and Rosemary Grant spent 40 years of summers on Daphne Major, an island in the Galápagos, measuring finch beaks. A drought hit in 1977 and much of the population died. The seeds left behind were big and hard, so the deeper a bird’s beak the better its odds. Jonathan Weiner wrote up their work in The Beak of the Finch.
The survivors’ chicks inherited the same deep beaks.
Chris Rock built a tour at a small club called the Stress Factory. He turned up unannounced 40 to 50 times carrying a yellow legal pad, ran 45 minutes a night and watched most of the jokes die. Peter Sims counted those nights in Little Bets.
A handful of those jokes made the tour.
The more a test costs to pass, the more its survivors are worth.
📝
Where could you add friction to see what survives?
In 2012 an investor turned down Jia Jiang’s startup.
Jiang set out to get rejected once a day for 100 days, to desensitize himself to the pain of it.
Day one he asked a security guard for $100, got a flat no and was terrified doing it.
Day two he got turned down again and made himself stay and joke with the man. Day three he walked into a Krispy Kreme and asked for doughnuts arranged like the Olympic rings. The shift leader made them and handed them over for free.
A UCLA exposure therapy protocol has you climb to the next item on your fear ladder once your rated distress on that item drops by half.
Rate “having a conversation with a stranger” at 80 out of 100 and work it until it reads 40.
Researchers at RMIT put 19 people through three virtual-reality public-speaking runs in one session. It took all three runs to bring the distress down.
The second run rated worse than the first.
Barbra Streisand walked out in front of 125,000 people in Central Park in June 1967 and blanked on her lyrics. She didn’t sing for paying audiences again for 27 years.
The fear started before that night, back in her Funny Girl run. She avoided the stage instead of returning to it. In her 2023 memoir, 56 years later, she writes that it’s still there.
In investing, two stocks are correlated when they move together. Ray Dalio found that a thousand stocks that move together protect you no better than five.
On 19 July 1989 a DC-10 left Denver for Chicago with 296 people aboard.
Over Iowa the rear engine’s fan disk blew apart and metal tore through the tail in dozens of places.
That plane carried three hydraulic systems. Losing one or even two still left you flying. McDonnell Douglas showed the FAA a design where:
“no single failure or malfunction or probable combination of failures will jeopardize the safe operation of the airplane.”
Debris cut hydraulic lines 1 and 3 where they ran through the tail. The departing disk tore out number 2.
Captain Al Haynes and his crew steered with engine thrust alone and put the plane down at Sioux City. 184 people walked away from that landing. 112 didn’t.
The design assumed those three systems would fail independently.
On 4 June 1996 an Ariane 5 lifted off from Kourou carrying two inertial reference computers, one live and one on standby.
37 seconds in, the standby computer shut itself down. The live one followed 0.05 seconds later.
Neither machine was broken. They ran identical code, so a single arithmetic overflow stopped both and the rocket tore itself apart.
A French mathematician named Jacques-Louis Lions chaired the inquiry board. His report reached Paris that July and named the cause: one piece of software, running in both places.
From the 1870s to the 1950s the banana you could buy was the Gros Michel, planted across two continents.
The Gros Michel looked like millions of separate plants. Each was a cutting from the same original, so the whole crop was one plant repeated. A soil fungus called Panama disease drove it out of the export trade by the early 1960s.
The industry replaced it with the Cavendish, immune to that strain. A head gardener named Joseph Paxton raised that variety at Chatsworth House in Derbyshire from a single plant shipped out of Mauritius in 1830. He named it after his employer, the Duke of Devonshire.
The Cavendish is also a clone.
A newer strain called Tropical Race 4 beats the Cavendish too and has reached more than 20 countries.
A single fungus could take millions of plants at once.
I could add a sixth health habit tomorrow. That’s a sixth cover on the one part of my life that already has five.
Ed Thorp made sure his investments could not all lose at once. He set his fund up so a loss on one side was matched by a gain on the other. Thorp made money whether the market rose or fell. His fund, Princeton/Newport, returned 19.1% a year for 19 years.
The first cover in an area with none beats the tenth in an area that’s already covered.
🗺️
What could you start in an area you don’t cover yet?
“Don’t it always seem to go, that you don’t know what you’ve got till it’s gone.”
Joni Mitchell
I stood in the cold shower this morning and felt nothing. When I started, it hit like a slap and woke my whole body. Today it’s just cold water and a clock.
Through the 1960s Egypt built the Aswan High Dam to stop the Nile’s yearly flood, generate power and store water for the dry season.
For thousands of years before that, the river rose each summer and spread black silt across the fields. The silt fed next year’s crops. Egypt worshipped the rising water as the god Hapi and timed festivals to it.
As the reservoir filled, it trapped the silt. The sardine catch off the delta fell from about 15,000 tons in 1964 to about 550 tons two years later.
The coastline started eroding into the sea.
I kept wanting to cut my meditation, because I stopped feeling it work.
Britain wanted to grow more of its own food. A hedgerow is a living fence: a dense line of hawthorn and blackthorn that marks a field’s edge and pens in livestock. After the war, the Agriculture Act of 1947 paid farmers to tear them out by the meter. Bigger machines needed bigger fields, so the small plots got merged.
Britain lost more than half its hedgerows, close to a million kilometers of them since the 1950s. The topsoil blew away with no windbreaks. Birds and pollinators lost their cover.
Now Britain pays farmers to plant hedges back, some working from old maps to set them where their grandfathers pulled them out.
I came close to dropping the fasting too, because I stopped noticing what it did.
For centuries, an Indian stepwell called a baoli dropped stone steps deep into the ground and gave a whole town water through the dry months. People met on the steps to draw it and talk. The deep stone stayed cool in the heat.
Then piped water arrived. A tap at home replaced the climb down for water. The wells looked unclean and outdated, so towns walked away, tipped garbage in, let the upkeep lapse. Now India talks about the day the taps run dry.
The revival has started: wells cleared out and put back to work, soaking rain into the ground. The pipes moved water to houses. The wells put water back under them, a job that stayed invisible until the pipes ran dry.
“Your scientists were so preoccupied with whether or not they could that they didn’t stop to think if they should.”
Ian Malcolm
“You can buy it. You can’t afford it.”
I said that to my partner years ago, mid-argument, over a car I thought they shouldn’t get.
Their case was simple: I can afford it. And strictly speaking, they could. The money was there. So was a big monthly payment.
I use the same argument against myself now.
Each day I run a batch of experiments, things I don’t yet know will work. I count the ones I finish.
My floor is the minimum number of experiments I hold myself to on the worst day.
Do more than that on a good day and the extra counts toward tomorrow, so a good week puts me days ahead of schedule. Those days in hand are my habit bank.
When the bank has days to spare, a voice says raise the floor: you can clearly afford more. Think how much faster you can move.
The bank proves I can hold my current floor.
It says little about whether I can hold a higher number tomorrow.
Poker players face the same question.
Your bankroll grows until it’s big enough for the next level of stakes. The pros test first.
The rule they repeat: prove you can consistently beat the stakes you’re at before you move up.
The bankroll gets you into the bigger game. Beating the players there is a separate question.
Chris Ferguson was already a World Series of Poker Main Event champion when he started a new bankroll from nothing.
Starting from free online tournaments in 2006, the game theorist set out to run zero dollars up to ten thousand. He got there in September of 2007. His two rules:
Never sit down with more than 5 percent of the bankroll on the table.
Leave the game once more than 10 percent of it is sitting in front of you.
Ferguson held a world title and still obeyed his own two rules at each table.
An edge is a real advantage, the skill that wins money over the long run. Serious players break “consistently beating” into four tests you can run on a habit too: how big your edge is (margin), how much you’ve measured (sample), what the game takes off the top (rake) and how wildly results swing (variance). The four tests together are what I call the Edge Check.
1) Margin.
The big blind is the forced bet that sets a table’s stakes. Players measure winning in big blinds won per 100 hands. Win $100 over 100 hands at a table with a $10 big blind and you’ve won 10 big blinds per 100.
A long-term winner usually wins 1 to 3 big blinds per 100 hands. The best players at small stakes win 5 to 10 big blinds per 100 hands.
Beating a level means winning with enough room that a bad stretch is survivable.
Long-Term Capital Management had two Nobel laureates at the partner table and a genuine edge.
The partners borrowed about $25 for each dollar of their own, which left almost no room for a bad stretch. One bad four-month stretch in 1998 erased 4.6 billion dollars. 14 banks put up 3.6 billion more to take it over.
If your floor is 20 push-ups and you average 21, you hold a 5 percent cushion, one rough week from breakeven.
If you average 30, you hold a 50 percent cushion. Raise the floor to 25 and you still clear it on an ordinary day.
2) Sample.
One good poker session is noise. A poker player wants 50,000 hands before a win rate means anything and closer to 200,000 before they trust it.
The casino knows its rate exactly. It deals millions of hands.
A first-night blackjack winner walks out sure they’re good at the game on a sample of one night.
In a habit, one heroic Saturday counts for about as much as a single poker hand. Raise your floor, the minimum you hold on your worst day, only after weeks of ordinary days beat it, with the bad days counted.
3) Rake.
The house takes a cut of each pot. Online that cut runs about 5 percent and stops at a fixed dollar amount. Live it runs closer to 10 percent.
A typical winner wins 1 to 3 big blinds per 100 hands. At the smallest online stakes, the house’s cut adds up to 8 to 10 big blinds per 100 hands.
A player can beat each opponent at the table and still lose money.
Pros count their win rate after rake.
Barber and Odean tracked 66,465 brokerage households through the 1990s and found the same cut at work: the most active traders earned 11.4 percent a year while the market returned 17.9. The stock picks were roughly average. The trading costs did the damage.
An hour of weight training at the gym costs you 40 more minutes driving there and back. That drive is 40 of the 100 minutes the habit costs you. Raise your floor from one gym session a day to two and you pay the drive twice. A habit you can run from your living room adds almost nothing on top of the work itself.
Count what a habit gives you after overhead.
4) Variance.
A real winner is a player whose skill wins money over the long run. The skill might earn 2 big blinds per 100 hands while ordinary luck swings results by about 100 big blinds. Play 100,000 hands and that much luck still leaves a genuine winner behind about one time in four.
In 1999, Berkshire Hathaway’s book value grew 0.5 percent while the S&P 500 returned 21 percent. Barron’s ran a December cover story titled “What’s Wrong, Warren?”. Buffett called it the worst absolute performance of his tenure.
The dot-com crash arrived the next year and Buffett’s caution turned out to be right.
A Baird study of 155 funds that beat their benchmark over a full decade found 97 percent spent at least one three-year stretch behind it.
In a habit, variance is bad days arriving in clusters. A bad month comes even while the habit is genuinely working, so a raised floor, a higher daily minimum, has to survive one. Before raising, ask whether you’d hold the new floor through a stretch where each rep runs over budget.
Margin, sample, rake, variance. Pass all four and you’re ready for the next level.
The Edge Check answers one question: Can I move up?
A higher floor, a bigger daily minimum, drains the habit bank faster and costs more willpower each day. We rarely ask where that willpower should go. Spend it at the bottleneck: the one constraint holding you back right now.
A morning run and a daily marketing block can both pass the Edge Check. Raising the run is still the wrong raise.
Each unit of willpower the run burns was needed on marketing, the habit where you’re actually stuck.
The Edge Check answers Can I? The bottleneck answers Should I?
♠️
What can you “afford” to do that you shouldn’t do?
The bicycle was a mess of guesses for most of the 1800s. Riders perched high above the road on a giant front wheel, the penny-farthing, one pothole from pitching over the bars.
The honeycomb landed on one shape too. A honeycomb is wall-to-wall hexagons and the hexagon is the cheapest way to divide a surface into equal cells: least wax, most honey.
People suspected this for more than 2,000 years. Marcus Terentius Varro wrote it down in 36 BC and Pappus of Alexandria is credited with it too.
The proof held out until 1999, when the mathematician Thomas Hales finally showed the hexagon wins.
The bicycle got its shape by trial and so did the honeycomb.
Hives varied from one generation to the next. The ones that spent less wax on walls left more descendants. The hexagon is what came through that filter.
In April 1985 Coca-Cola replaced a recipe that stood for 99 years.
Roughly 190,000 blind taste tests said the new one tasted better. They pulled the old recipe off the shelf.
Then the letters and the phone calls started, 8,000 calls a day at the peak.
The old recipe came back as Coca-Cola Classic and you can buy it today.
New Coke still ranks among the biggest product flops in business history. The recipe was finished, with nothing left worth changing. They changed it anyway.
Your keyboard never got finished and it stayed anyway. It has been the standard since the 1870s and it was never the fastest layout.
Christopher Sholes spread the common letters apart to keep the typebars of a mechanical typewriter from clashing, then the world kept his layout long after the jams were gone.
Her best proof that her hair formula worked was her own bald head. She carried before-and-after photographs into church basements and told the story herself.
The hair loss came from years of scrubbing other people’s clothes for $1.50 a day. The regrowth came from a formula she mixed herself.
White-owned newspapers would not carry her advertising.
White-owned beauty schools would not teach her.
She went door to door, bought standing weekly ads in the Black press, opened a factory in Indianapolis and founded her own school.
By 1917 about 20,000 women worked as Walker agents.
Each bought product at wholesale, sold it at retail and kept the difference. Each could recruit and train her own agents and earn on their sales.
In 1912 she asked Booker T. Washington for a speaking slot at his National Negro Business League convention. He ignored it. She traveled to Chicago anyway, then rose from the audience as the last scheduled speaker sat down.
“Surely you are not going to shut the door in my face. I am a woman who came from the cotton fields of the South.”
“Don’t sit down and wait for the opportunities to come,” she said. “Get up and make them.”
In college I rebuilt Reddit in a few lines of Ruby. The man who invented Ruby shared the clone. Ruby Weekly, the newsletter its programmers read, ran it too.
I was sure one rule stood between me and a developer job: intern as a developer first.
My internships were in another field and my college sat in a small town. I learned to code alone. So I built the proof myself, a portfolio of my own projects.
The portfolio got me the interview.
Then it got me the job.
Trends.vc came next, my market research newsletter, started with no credentials as a market researcher. I published weekly reports anyway. Readers subscribed and stayed.
The other move is to manufacture the start: make the job, invent the internship, assign yourself the project, build the portfolio.
I read a biography each day.
Sara Blakely was selling fax machines door to door after failing the law school admissions test twice.
She cut the feet off her own pantyhose, called the result Spanx and filed the patent herself with a book she bought at Barnes & Noble. Her pitch to Neiman Marcus was changing into the prototype in the store restroom, in front of the buyer.
Tyler Perry put his $12,000 in savings into renting a 200-seat Atlanta theater to stage his own play.
The play drew 30 people. He slept in his car for stretches and kept restaging it until it sold out 8 straight nights in 1998.
A designer named Aaron Draplin taught himself Illustrator, the drawing software, on the floor of a snowboard company. He hand-printed 200 memo books, mailed them to friends and sold 13 copies on launch day. Those memo books grew into Field Notes, the notebook brand.
The portfolio was built before anyone hired me.
The reports went out before a single reader subscribed.
Blakely made the prototype before Neiman Marcus said yes.
“If we are not aware that we are happy, we are not really happy. When we have a toothache, we know that not having a toothache is a wonderful thing. But when we do not have a toothache, we are still not happy. A non-toothache is very pleasant.”
Thich Nhat Hanh
I dug the same grave 5 times. All 5 climbed back out, some more than once.
Survival is information: a habit that outlives my judgment is one worth keeping.
A crocodile has worn the same body for 200 million years. That’s older than grass. It lived through the asteroid that killed the dinosaurs. Its design works.
Age proves a design. Same for a reptile, same for a habit. The trouble is a habit’s benefit stops registering.