Lonly

19 | Finance\Creator

[Lonly @0xLonly]

Elon Musk started SpaceX expecting to make exactly $0, and the 4th rocket launched with the last money he had.

Watch the frame. June 2012. Caltech commencement. Musk at the podium in a black gown with a red hood, reading carefully, almost shy. The TV banner under him lists the companies: SpaceX, PayPal, Tesla Motors, SolarCity.

Then he tells the graduates how SpaceX actually began. It wasn't a business plan.

After PayPal, he assumed nobody could start a rocket company. So the first idea was smaller. Send a cheap greenhouse to Mars. Land some seeds in dehydrated gel, add water, and beam back one photo: green plants on a red planet. First life on Mars. The public would go wild, he figured, and NASA's budget would go up.

He put the money math plainly. The expected financial return was zero. Anything above zero was upside.

To get it there, he flew to Russia 3 times in 2001 and 2002 to try to buy refurbished intercontinental ballistic missiles. The cheapest rockets on earth were old nukes. He walked into the Russian rocket forces and asked for 2 of their biggest. They could keep the warheads. They thought he was crazy.

He had money, so it was fine. But the price kept being the problem.

On the 3rd trip back, he saw it. People had plenty of will to go to space. What they didn't have was an affordable way to get there. If space looked impossible, or like it would break the federal budget, nobody would fund it.

So he decided to fix the cost of the ride itself.

Almost everyone he talked to told him not to. One friend sat him down and made him watch videos of rockets exploding.

The friend wasn't far off. SpaceX failed 3 times. The way Musk puts it, with rockets the passing grade is 100%. You can't test the whole thing in space before you fly it, so the first real run has to have zero bugs. After the 1st launch, he was picking up pieces of his rocket near the pad.

The 4th flight reached orbit in 2008. It was paid for with the last money the company had.

If it had failed, there would have been no 5th.

4 years later, a Dragon capsule had docked with the space station and come back. And he was standing in a gown telling 20-year-olds that anything works on PowerPoint, so build the real thing.

He went in expecting $0.

He came out with the cheapest ride to orbit on the planet.

[Lonly @0xLonly]

@Ox_Cheetah Yeah, $80 was the theoretical floor

[Lonly @0xLonly]

Elon Musk was told batteries cost $600 per kilowatt-hour and always would, so he priced the metals inside and got $80.

Watch the frame. A Tesla factory floor so clean it shines like glass. Rows of red machines in the background. 2 men on stools in matching black t-shirts. Kevin Rose asks the question he asks everyone at the end. What's the one piece of advice you'd give a new founder?

Musk gives 2.

The first is uncomfortable. Go looking for negative feedback, and listen hard. When friends try a new product, he tells them to skip what they liked and only tell him what they didn't. A friend won't volunteer the bad list. He wants to stay your friend. So you have to pull it out of him. Treat praise like water off a duck's back. Overweight the criticism.

The second is the one that built the company.

Most people, he says, reason by analogy. We do things because they look like something that was done before, or because it's what everyone else is doing. It's easier. It takes less mental energy.

The other way is first principles. Boil the problem down to what you know is true, then build back up from there.

Then he shows exactly what that looks like with money.

The industry said battery packs cost about $600 per kilowatt-hour. That's what they had always cost, so that's what they would always cost.

Musk asked a different question. What is a battery actually made of? Cobalt. Nickel. Aluminum. Carbon. Some polymers for separation. A steel can.

Then: what would each of those cost if you bought them on the London Metal Exchange?

The answer came to roughly $80 per kilowatt-hour.

The gap between $600 and $80 wasn't physics. It was habit. It was everyone pricing the battery by looking at the last battery. The metals were cheap. The process of turning them into a cell was the expensive part, and processes can be redesigned.

He compares it to the people who said nobody would want cars. Horses are great, they eat grass, grass is everywhere, and there's no gasoline to buy. People really said that.

Then a line every founder should tape to the wall. A car has several thousand unique parts. 90% are fine. 5% are slightly problematic. 3 to 4% are problematic. 1% are extremely problematic. And unlike software, you can't ship a car that's 99% done.

The price on the label is what the world has always paid.

The price of the metals is what it actually costs.

[Lonly @0xLonly]

@0x_Squirrel Clean rooms, formation, yield losses none of that shows up on the London Metal Exchange. The gap was real, just not permanent

[Lonly @0xLonly]

@Plioric89 True

Even at ~$100 in 2020 it was still sitting above the pure metals floor he pointed at

[Lonly @0xLonly]

@softarox Wouldn’t give him the farm just because of the last name

[Lonly @0xLonly]

@softarox Elon knows his stuff when it comes to this

[Lonly @0xLonly]

@softarox True, step three took longer than the original plan

[Lonly @0xLonly]

Elon Musk sat on a Silicon Valley stage in April 2009, months after almost losing Tesla, and explained that the whole company came down to the price of 1 battery.

Look at the frame. A small stage, 2 armchairs, a wooden side table with 2 bottles of water. A blue banner behind him repeating CHURCHILL CLUB. Musk in jeans and a dark jacket, legs crossed, hands moving the whole time. Across from him, an older moderator in a red tie with notes on his knee.

It doesn't look like the stage of a man who nearly went broke 4 months earlier.

But he had. In late 2008 Tesla was weeks from running out of cash. Musk had put most of what he made from PayPal into Tesla and SpaceX. The funding round that saved Tesla closed on Christmas Eve.

And here he is, calmly explaining why the car only works if the battery gets cheap.

The trick was almost boring. Every carmaker wanted to design a custom battery from scratch. Tesla didn't. The Roadster ran on about 6,800 ordinary lithium-ion cells, the same kind used in laptops. Billions of them were already made every year. The factories, the supply chains, the price drops were all paid for by the laptop industry.

Tesla just wired them together.

That's the money logic underneath everything he says on that stage. A battery is not a technology problem. It's a cost-curve problem. Every year those cells got cheaper, and every dollar cut from the pack was a dollar off the car.

So the plan ran in order. Build an expensive sports car for rich buyers, around $109,000. Use that money to build a cheaper sedan. Use that to build a cheaper car after it. 2 weeks before this talk, Tesla had just shown the Model S for $57,400, about half the price of the Roadster.

Same cells. Bigger factory. Lower price.

Most people saw an electric car company.

Musk saw a bet on 1 number going down every year, and he had bet nearly everything he owned on it.

The rich paid for the first car.

The cost curve paid for the rest.

[Lonly @0xLonly]

@NitePov Exactly

[Lonly @0xLonly]

A retired couple from a town of 1,900 people won around $26 million from the lottery, and the math took Jerry Selbee 3 minutes.

Watch the frame. A small convenience store in Evart, Michigan. Shelves of cigarettes, a wooden counter, a cash register. Jerry and Marge ran it for years. Jerry had a math degree, worked a factory job before the store, and had a habit nobody else in town shared. He read the fine print.

In 2003 he picked up a brochure for a new state game called Winfall. Most people read the jackpot. Jerry read the rules. One of them stopped him cold.

When the jackpot passed $5 million and nobody hit all 6 numbers, the money didn't roll over to next week. It rolled down. The entire pot got split among the smaller prizes, the people who matched 3, 4 or 5 numbers.

Jerry did the arithmetic at the kitchen table. A normal week, a $1 ticket is worth maybe 50 cents in expected prizes. That's how every lottery works. The state keeps the rest.

But on a roll-down week, millions of extra dollars suddenly sat in the small tiers. Prizes that usually paid a few dollars paid hundreds. Spread across enough tickets, the math flipped. Every $1 spent was expected to bring back more than $1.

Not luck. Not a system of lucky numbers. A flaw in where the prize money went.

So he tested it. He bought a few thousand dollars of tickets on a roll-down week and came out ahead. He did it again, bigger. It held.

Then he scaled it. Jerry set up a company and let friends and neighbors in. Look at the old men in the film, sitting at a checkered table with coffee and pie. Those were the investors. 32 of them, regular people from a small town, splitting the profits.

When Michigan shut Winfall down, Jerry found the same rule in Massachusetts. He and Marge drove across the country and checked into a motel for days at a time, feeding play slips into the machine at a local store for hours. Hundreds of thousands of tickets on the right weeks. Nothing on the wrong ones.

They weren't alone. A group of MIT students found the same hole and started playing it too. When a newspaper finally exposed the game, the state investigated. The verdict: everything they did was legal. The game was simply built wrong.

Look at the binders on his dining table. Every ticket logged. Every roll-down tracked. 9 years of it.

The lottery is designed so the average ticket loses. That's the whole business. Jerry never fought that. He just waited for the weeks when the average turned the other way, and bought every one of them.

Everyone else bought a ticket and hoped.

He bought the week when the math was on his side.

[Lonly @0xLonly]

Bill Benter made close to $1 billion betting on horses, and the house never once lost.

Look at the photo in the film. A cramped desk. A boxy computer, 2 lamps, stacks of binders. And on the corner of the desk, a roulette wheel. That’s where it started.

Benter began in Las Vegas as a blackjack card counter in the late 1970s. Casinos banned him. So he went looking for a game where they couldn’t.

He found it in Hong Kong. At Happy Valley and Sha Tin, the Jockey Club doesn’t bet against you. It pools every ticket, takes its cut, and pays the rest to the winners. You aren’t playing the house. You’re playing the crowd.

And the crowd is wrong in predictable ways.

So Benter built a model. Past finishing times, weight carried, the draw, the jockey, the distance, how a horse ran its last race. Years of results fed into an old Kaypro machine, every race scored, every horse given its own probability of winning. Then he compared his odds to the odds on the tote board.

When the crowd priced a horse at 10 to 1 and his model said 5 to 1, he bet. When it didn’t, he didn’t.

That’s the whole trick. Not picking winners. Finding the gap between what the crowd thinks and what the numbers say.

The bar was brutal. The track keeps around 17% of every pool. A model that’s merely good still loses. It has to beat the crowd by more than the cut, every race, for years.

His did. The screen in the film shows it plainly. PLACE BET 1. PLACE BET 2. All the way to 12. Hundreds of bets a night, sized by math, not hunches.

Most gamblers try to beat the house.

Benter just let the crowd pay him for being wrong.

[Lonly @0xLonly]

Don Johnson took $6 million from the Tropicana, then gave an interview on its casino floor, standing next to its CEO.

Watch the frame. 3 men under the departure-board screens of an Atlantic City casino. On the left, Don Johnson, black jacket, relaxed. On the right, Tony Rodio, President and CEO of Tropicana. Between them, a reporter with a mic. Slot machines blinking behind all 3.

By April 2011, Johnson had won roughly $15 million from 3 casinos. Tropicana close to $6 million. Borgata $5 million. Caesars $4 million. He bet up to $100,000 a hand.

No card counting. No hidden earpiece. No team in the parking lot.

He won before the first card was dealt.

Casinos were desperate for whales after 2008. Johnson knew it. So before he sat down, he negotiated. A 20% rebate on his losses. Softer rules at the table. Dealer stands on soft 17. Split and double down more freely.

Run the math. Lose $500,000 in a session, get $100,000 back. The house edge on blackjack is already thin, around 0.5% with good rules. A 20% rebate on losses flipped it. The game was now tilted toward the player, and the player was betting $100,000 a hand.

The casinos priced the deal for an average high roller who would lose over time. Johnson made sure he wasn’t average. Every term they signed shifted the expected value a little further his way.

That’s the whole edge. Not beating the dealer. Beating the contract.

And then he walked back onto the floor, next to the man whose casino paid for it, and smiled for the camera.

Most gamblers ask how to win the hand.

He asked what the house would pay him to play.

[Lonly @0xLonly]

Every stock price you track is secretly 2 hidden forces wearing 500 disguises, and 1 matrix pulls off the mask.

An MIT lecturer, Choongbum Lee, starts with something that looks boring. A grid of numbers. Rows are companies, Apple, Morgan Stanley, the rest. Columns are dates. Each cell is a price. Just a spreadsheet.

Then he flips how you see it. That grid isn't only a table. It's a machine that stretches and bends space. And every such machine, no matter how messy, breaks down into a few clean directions and how hard it pulls on each one.

The tool is the singular value decomposition. A = UΣVᵀ. Any matrix at all, even a rectangle, even a mess, splits into 3 honest pieces. The directions going in. The directions coming out. And a short list of numbers, the singular values, saying how much each direction actually matters.

Here's why it hits. You think you're watching 500 stocks move independently. You're not. Pull the SVD and a handful of those singular values are big and the rest are basically 0. Translation: 500 wiggling prices are really driven by a few underlying forces. The whole market in 1 move. A sector in another. Everything else is noise dressed up as signal.

That's the quiet idea under every "factor model," every risk engine, every quant desk. Not predicting the next tick. Finding the 2 or 3 real dimensions hiding inside hundreds of fake ones, and throwing the rest away.

And the compression is brutal. 500 companies, 365 days, a giant wall of numbers. Keep the few singular values that matter and you can rebuild almost the entire thing from a sliver of the data. The structure was never in the 100,000 prices. It was in the 5 numbers behind them.

Lee is careful, and so should you be. SVD doesn't predict a price or hand you a trade. It shows you the skeleton under the data, nothing more. But once you've seen that a market of hundreds runs on a handful of hidden axes, you stop confusing motion with information.

He put the entire lecture online for free. MIT 18.S096. Chalk, a grid, and the math under every quant fund.

The wall of prices was never the thing.

A few hidden numbers always were.

[Rossst.03 @Rossst_03]

Most on-chain bots lose by being fast but wrong.

The winners do three things before the block closes: detect, simulate, include.

In this article I show EXACTLY how to build an on-chain HFT bot that trades every block, from scratch. https://t.co/HYApyAT9ha

[Lonly @0xLonly]

1 credit boom made Thailand both the 1997 crash and one of Asia's fastest-growing economies.

An MIT economist, Robert Townsend, put India and Thailand on a screen. India grew slow and safe. Thailand opened the credit taps, boomed, then blew up in 1997. Currency broke. Banks failed.

Over the long run, the country that crashed still ended up far ahead.

Not because the crash helped. Because the boom and the bust came from the same thing. Credit.

Fund the businesses nobody would fund, and output jumps. So does the risk. It's 1 lever with 2 outputs, and they ship together. You don't get the growth without the odds of ruin.

That's the equation nobody prints. The move that makes you rich and the move that wrecks you are the same move.

He gave the whole lecture away for free on MIT's site.

The same lever built the miracle and the wreckage.

[Lonly @0xLonly]

A stock that doubles then halves averages +25% a year, and leaves you with exactly $0 in profit.

Run the numbers once and it never leaves you. Up 100%, then down 50%. The "average" of those 2 years is plus 25%. Sounds like a winner.

But your money doesn't add returns. It multiplies them. Double it, times 2. Halve it, times one half. 2 times one half is 1. You are exactly where you started. Zero profit, hiding behind a 25% average.

Thomas Cover, a Stanford professor, puts this on a screen and explains the fix. For anything that compounds, the plain average lies. The honest number is the geometric mean, and here it's 0. That's 1 of the 5 equations from that finance post, the one almost nobody checks. A fund can advertise a 25% average while the people in it made nothing at all.

Then he does something that looks impossible.

Take that same going-nowhere stock. Put half your money in it, half in cash. The cash grows nothing. The stock compounds to nothing. Two dead assets. Now rebalance back to 50/50 every period, selling whatever went up, buying whatever went down. In the clean version, before fees, that combination compounds at about 6% a period over the long run. He builds growth out of 2 things that each lead nowhere. The volatility you were told to fear turns out to be the fuel.

He shows it with real stocks too. Over about 20 years, 2 wild, barely-related names grew roughly 4 times and 8 times on their own. A rebalanced 50/50 mix of the exact same two grew about 52 times. The best split cleared 70 times. He's upfront that he hand-picked them to make the graph scream, but the mechanism underneath is real.

The thread under all of it is simple and expensive to ignore. The average return in the headline is almost never the return you keep. And once you measure growth the right way, it changes what you'd even buy.

He gave the whole proof away at a lectern, for free.

The average was always the story.

The geometric mean is the money.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Lonly @0xLonly]

The kids who grabbed 1 marshmallow instead of waiting for 2 weren't weak, they'd learned the adult might not come back.

You know the experiment. A child, a table, 1 marshmallow. Wait about 15 minutes without eating it and you get a second one. Walter Mischel ran it at Stanford in the late 1960s, tracked the kids for more than a decade, and the ones who waited seemed to do better at almost everything later. The takeaway spread everywhere. Success is self-control. Teach a kid to wait and you've handed him his future.

A teacher named Jiang Xueqin stands at a board and flips it over.

A child only waits for the second marshmallow if he believes the second marshmallow is real. If the adults in his life have promised things before and not delivered, then eating the one in front of him isn't a failure of willpower. It's the rational move. You take the sure thing, because experience taught you the future is a coin flip you tend to lose.

And researchers found exactly that. When the experimenter had already broken a small promise minutes earlier, the kids waited a fraction as long. Not because their self-control vanished, but because the signal was clear. Don't trust this one. Trust turned out to be a quiet ingredient the famous version left out.

Here's why it lands on money. Every finance equation says the same thing in the end. Wait. Let it compound. The patient win. But the math assumes something it never prints on the page. That the future will actually pay out. On paper, patience is free. In real life it runs on trust, on stability, on having grown up believing tomorrow shows up when it says it will.

None of this makes a rough start a sentence. People climb out, and Jiang is clear that they do. It means the ability to wait is partly something that was built in you, or wasn't, by whether the world kept its word. And that's a thing you can rebuild on purpose, one kept promise to yourself at a time.

The equation says wait.

It never asks whether you can afford to.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Lonly @0xLonly]

@recketino Every cycle has a moment where Buffett looks completely outdated

[Lonly @0xLonly]

@furiokao You are never fair when you compare

[Lonly @0xLonly]

@levifizg The number on the sign and the number that actually lands in your pocket are two different things

[Lonly @0xLonly]

An MIT professor proved a $1,000,000 lottery prize is really worth $607,906, and he did it on a chalkboard.

You've seen the deal. You "win a million," but nobody hands you a million. They pay it out as $50,000 a year for 20 years. Same number on the sign. Nowhere near the same value.

Tom Leighton walks a room of MIT students through exactly why.

A dollar you get next year is worth less than a dollar today, because today's dollar can be put to work. At 6%, a dollar arriving in a year is worth about 94 cents right now. A dollar arriving in 10 years is worth about 56 cents. So each of those future $50,000 checks is smaller than it looks, shrunk by how long you have to wait for it.

Then he does the thing most people never see. He lines up all 20 payments, pulls each one back to today's value, and adds them together. The long ugly sum collapses into one short formula, live, in chalk. Plug in $50,000, 20 years, 6%, and out comes $607,906.

So the "million" is worth about $608,000 today. He makes it concrete. Offered $700,000 in cash instead, you take the cash. Offered $500,000, you take the stream. The number printed on the check was never the value of the prize.

And it all swings on one hidden number. Lower the interest rate and the future payments are worth more. Raise it and your million shrinks further. The exact same prize is worth wildly different amounts depending on a figure that appears nowhere on the ticket.

This is 1 of the equations from that finance thread, present value, drawn out in full. It's the quiet line sitting under every pension, every legal settlement, every loan, every lottery on earth. The people who know it never get tricked by a big number handed over slowly.

He put the whole derivation online for free. MIT. Lecture 12. Chalk, a board, and the truth about your million.

A million paid slowly was never a million.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw