Dark Factor @DarkFactorr
Financial research, machine intelligence, systematic strategies, and model risk. Joined July 2018-
Tweets11
-
Followers8
-
Following7
-
Likes46
Bill Ackman teaches balance sheets, income statements, cash flow, debt, equity, dilution, and return on capital without beginning with a single public company. He builds a lemonade stand instead. Ackman founded Pershing Square and spent his career analyzing real businesses, but in this lecture he reduces the machinery of finance to $750. An investor contributes $500 for one-third of the company. A lender adds $250 at 10% interest. The money is converted into a $300 stand, $200 of inventory, and the cash required to begin operating. From there, every financial term becomes a visible event. Buying lemons creates inventory. The stand becomes a fixed asset and depreciates as it wears out. Selling 800 cups creates revenue. Paying workers creates an operating expense. Reinvesting the cash produces more locations. By the fifth year, seven stands generate almost $8,000 in revenue, while profitability rises from 1.3% to 28.6%. Then Ackman uses the same business to explain why the lender earns only 10% while the equity investor can earn more than 100% on the original purchase price. The lender receives the first claim on the assets if the business fails. The shareholder accepts the possibility of being wiped out in exchange for everything left after the debt is paid. It is one of the cleanest visual explanations of corporate finance because every accounting line can still be traced back to a physical cup, stand, loan, or ownership claim. Watch the sequence where an imaginary street business gradually becomes a complete financial model.
The CEO of a company on pace for $8.15 billion in annual revenue went on camera and explained who should own the money created by AI. frontier labs spent the last three years pretending the answer was already settled. Alex Karp didn’t pitch Palantir. he didn’t talk about
The CEO of a company on pace for $8.15 billion in annual revenue went on camera and explained who should own the money created by AI. frontier labs spent the last three years pretending the answer was already settled. Alex Karp didn’t pitch Palantir. he didn’t talk about benchmarks. he described a future where OpenAI and Anthropic control the intelligence layer, enterprises surrender their data and margins, and millions of workers are told that losing their jobs is simply the cost of progress. the man making the argument is not watching AI from the sidelines. Palantir just reported $1.94 billion in quarterly revenue, up 93% in one year. its US government business reached $809 million, commercial revenue hit $764 million, and the company signed $3.37 billion in contracts in a single quarter. Karp said he has spent enough time with Dario Amodei and the effective altruism circle to understand the worldview: people own less. businesses become less profitable. jobs disappear. the labs controlling the models capture the upside. then the interviewer asked whether Chinese labs were stealing from Anthropic through distillation. Karp answered with one question: “how do you think the models got their value?” his point was that frontier labs already distilled value from code, research, intellectual property, and enterprise knowledge across the entire economy. they discovered property rights only when the extraction began moving in the opposite direction. the part nobody talks about: Karp gave an exact price for a fair AI economy. if a model creates $1 billion for a company, the builder should earn $300 million. that still creates one of the most profitable products in history while leaving $700 million with the enterprise that owns the employees, customers, capital, and consequences if the deployment fails. this matters while Anthropic is valued near $965 billion and OpenAI around $852 billion. Wall Street is debating which lab reaches $1 trillion first. Karp is asking whether the companies underneath them remain worth owning at all. a Fortune 100 executive told me this is the first AI interview his board circulated without an internal summary. not because Karp explained the technology. because he explained who gets the money.
A self-made millionaire who went from poverty to a $1M net worth in 15 years explained how to buy a profitable business without paying for it upfront. The opportunity starts with owners who spent 20 or 30 years building local plumbing, mechanic, HVAC or construction companies. Now they want to retire. Their children do not want the business. There is no buyer ready. So a company with customers, equipment, staff and cash flow may simply close. His solution is to structure the deal around the business itself. Agree on a price with the owner, keep them involved as a mentor, then pay them from future profits perhaps 10% each year instead of handing over the full amount on day one. The seller gets retirement income without running the company. The buyer gets an operating business without taking on a huge upfront risk. And because the former owner stays involved, the knowledge that normally disappears at retirement is transferred with the asset. This is the part most people miss: you do not always need enough money to buy the business. You need a deal where the business can help pay for its own acquisition. The seller owns an asset they no longer want to operate. The buyer owns time, energy and the willingness to continue it. The money appears when the structure connects both sides.
THIS PROFESSOR UNDERSTOOD ONE THING MOST COMPANIES GET WRONG, AND GOOGLE TURNED IT INTO A $74.5 BILLION-A-QUARTER MACHINE His name is Tim Roughgarden. He became one of the leading researchers in algorithmic game theory—the mathematics of what happens when computers, markets and selfish people collide. In one quarter, Google Search and YouTube produced $74.5 billion in advertising revenue. Roughgarden explains the principle beneath businesses like this in an old Stanford lecture filmed in front of an ordinary chalkboard. Most companies design systems for the people they wish they had. Honest customers. Rational employees. Advertisers who reveal what they really want and follow the spirit of the rules. Roughgarden starts from the opposite assumption. Everyone will pursue their own advantage. The question is whether your rules turn that selfish behavior into the result you want. Google cannot negotiate personally with every company that wants an ad above a search result. It lets them compete inside an auction whose rules transform hidden valuations into bids, rankings and prices. That field is called mechanism design. The secret is not convincing millions of strangers to cooperate. It is making cooperation emerge while everyone is still optimizing for themselves. Google did not build one of history’s largest advertising businesses by asking people to stop being greedy. It designed a machine where greed became an input. Read the article below to understand how auctions convert private incentives into billions in market revenue.
Peter Thiel helped build PayPal, was the first outside investor in Facebook, backed Palantir and SpaceX, and says most startups are dead before they launch. his test is brutal: if you are building the 4th online pet-food company, the 10th solar-panel company, or the 1,000th restaurant in Philadelphia, someone else would do it if you didn’t. which means you are probably not building anything important. bookmark this. it is one of the cleanest filters for whether your idea has real venture upside or is just another resume project. Thiel’s point is that great businesses are not born from competition. they escape it. he says there are really only 2 kinds of companies: businesses trapped in crazy competition, and one-of-a-kind businesses that behave like monopolies. that is the real difference between a startup that might become worth billions and one that spends five years fighting for scraps. the goal is not “be an entrepreneur.” that is as meaningless as saying “I want to be rich.” the real question is: what important problem are you solving that nobody else is solving? what do you know that other people are missing? that is where venture-scale value comes from. not copying what already works. but building the thing that would not exist without you.
On August 23, 1991, a concrete offshore platform sank into a Norwegian fjord and erased nearly $1 billion. the steel was not the first thing that failed. the simulation was. engineers had used numerical software to estimate the stresses inside the platform’s load-bearing walls. the model underestimated them. the structure was lowered into the water, one wall cracked, and the entire 57,000-ton platform disappeared beneath the surface. Oxford professor Endre Süli uses the disaster to explain the one rule every quant, engineer and model builder eventually learns: a precise number is worthless when the approximation underneath it is unstable. he starts with a derivative. replace an infinitely small change with a finite step. turn the differential equation into numbers a computer can calculate. repeat the process across thousands or millions of points. that is numerical analysis. the same machinery simulates airflow around aircraft, cracks moving through concrete, heat spreading through materials and prices evolving through time. but every shortcut introduces error. make the computational steps too large and the answer becomes fast but wrong. choose an unstable method and tiny errors grow until the simulation describes a world that cannot exist. this is the hidden problem inside quantitative finance too. a backtest can be exact to six decimal places while the model is discretizing volatility, liquidity and tail risk incorrectly. the code runs. the chart looks clean. the bridge still falls. Oxford teaches the mathematics behind reliable simulations in its third-year course. hedge funds pay quants fortunes for the same judgment: knowing when a model is approximating reality, and when it is only producing numbers. the lecture is free. the wrong approximation cost $1 billion
In April 2002, AOL Time Warner traded at $21.85. For $0.85, anyone could buy the right to dump the stock at $20 one month later. That tiny contract controlled more than twenty times its own price. Robert Shiller brought the newspaper clipping into a Yale classroom and explained the one idea traders spend years pretending is more complicated than it is. The lecture is still online. almost nobody reaches the part that matters. Shiller begins with a farmer. A developer finds land beside two highways and thinks it could become the perfect place for a supermarket. Buying the entire farm today would be expensive. Walking away could mean losing the location forever. So he pays the farmer for one thing: the right to decide later. The farmer takes the money and accepts the obligation. The developer keeps the choice. If the land rises, he buys at the old price. If the project dies, he loses only the premium. That asymmetry became a global derivatives market. In Shiller’s newspaper clipping, AOL traded at $21.85. A one-month $20 call cost $2.55. A $20 put cost $0.85. More than 2,000 of those puts changed hands that day. The stock was one price. The market had already created dozens of other prices for what might happen next. Then Shiller walks to the board and draws the whole industry as one broken line. Below the strike, the call is worth zero. Above it, the payoff rises dollar for dollar with the stock. Ordinary portfolios are straight lines. They absorb every move. Options insert a corner where the exposure changes. That corner is why a £0.85 contract can protect a £21.85 asset. It is also why mortgages, insurance, employee compensation and structured products are all versions of the same trade. The first organised options exchange opened in Chicago in 1973. By the time Shiller gave this lecture, derivatives had become a market independent of the assets underneath them. People could buy the option, sell it, write it naked and close the position without ever touching the stock. No terminal. no code. one old newspaper clipping and a line with a bend in it. The formulas price the contract. The bend decides who survives the move.
Furiokao @furiokao
130 Followers 30 Following silence before the strike knowledge that cuts through noise
Rossst.03 @Rossst_03
5K Followers 214 Following Where prediction markets meet AI. I hunt mispriced odds on Polymarket. @zscdao member
Goshawk Trades @GoshawkTrades
75K Followers 396 Following Algorithmic trader | 7 years live | 10+ algos in production Helping traders go algorithmic (300+ so far). Co-founder @AskEdgarIO (AI agents for funds).
venus @RitOnchain
3K Followers 480 Following engineering alpha through systematic infrastructure in quant & ai
Livsun @L1vsun
4K Followers 180 Following US Entrepreneur × Quant Trading × Finance Tech For collab & business - tg: @l1vsunYou might like
