Tema Ultra

May 2007: Apple had sold 100 million iPods, the Zune not yet 1 million. Bill Gates says the Zune team loved Apple’s “gigantic market.” Steve Jobs sat beside him with an unreleased iPhone. In the video below Gates tells it himself: "In a few cases like the Zune, that if you go over to that group, they think of Apple as a competitor." He says they're going to "come in and contribute something." Asked what he carries, Jobs answers a MacBook Pro "and an iPhone," a month before launch. Apple announced its 100 millionth iPod on April 9, 2007, after selling 10.55 million from January to March. The Zune went on sale in November 2006, and Microsoft expected its first million by the end of June. By the 2008 holiday quarter Zune revenue was down 54% on a year earlier. In one 2010 quarter Apple sold 14.1 million iPhones and 9.05 million iPods, the iPod down 11% in a year. In October 2011 Microsoft stopped making Zune players. Apple sold about 116,000 iPods a day in early 2007. The Zune's first million equals about 9 days of that. Zune revenue: about $185 million to $85 million in that holiday quarter. The case for the Zune: its interface grew into Metro, the design Microsoft carried into Windows Phone, Windows 8 and the Xbox. 100 million iPods when Gates called it a gigantic market. Four years later the Zune was gone and the phone Jobs carried that day outsold the iPod. When a rival builds a gigantic market, do you fight for a share of it, or build what replaces it?

In June SpaceX sold 555.6 million shares at $135 in the biggest IPO ever. Three days in, Gavin Baker called it one of the most important companies of all time, and the next day it hit $225. On October 30 it closed at $134.72. In the video below it opens on a toast: "A round of applause for the banker." A host calls the 20% pop flawless. Baker, an early SpaceX investor, says Goldman Sachs and Morgan Stanley "did a very good job" and that retail is probably "outperforming the overwhelming majority of professional money managers." SpaceX had fixed the price at $135 with no range, raised $75 billion and sent about 30% of the deal to retail, at least triple the norm. Retail orders topped $100 billion. Day one closed up 19%. Baker says some of that stock "may have gotten allocated to people who were flipping it," and that retail is "very sensitive to momentum." Momentum ran both ways: a $225.64 peak, under $135 by mid-July, as low as $104.83. In August a lock-up let up to 911.5 million more shares trade, more than the IPO sold. Day one: 555.6M shares x $25.95 pop = $14.4 billion. $10,000 at $135: $9,979 on October 30. $10,000 at the $225.64 peak: $5,971. Same stock, five days apart, a $4,009 gap. The case for the bankers: their price sits 28 cents from the October 30 close, and SpaceX banked the full $75 billion. $135 when the bankers priced it. $134.72 on October 30, after a round trip through $225. Was $135 the perfect IPO price, or $14 billion handed to whoever got shares and flipped them on day one?

In 2015 a CEO cut his pay from $1.1 million to $70,000, citing a Nobel study: happiness stops rising at $75,000. In 2023 the same laureate said it keeps rising to $500,000. Matthew McConaughey says he knows the exception. In the video below he tells it himself: "Hardcore capitalist, go for it. More, more, more success." Then he asks how you have "profit with your success" and names who doesn't: "unhappy billionaires." He says he loved his four years living in an Airstream, where "you only have room for one of everything." The CEO ran Gravity Payments in Seattle. About 30 of his 120 employees doubled their pay; two senior staff quit in protest. The study was Kahneman and Deaton's, built on 450,000 Gallup answers. In 2021 Penn's Matthew Killingsworth logged 1.7 million real-time check-ins and found no flat line. In 2023 he and Kahneman reworked his data together. Their answer: for most people happiness keeps rising to $500,000, faster past $100,000 for the happiest 30%. The old plateau was the unhappiest 20%, who stop gaining near $100,000. McConaughey's billionaires sit outside the data: too few people earned over $500,000 to test. $75,000 in 2010 is about $110,700 in 2025 dollars. Happiness tracks the log of income: $50,000 to $100,000 buys as much as $250,000 to $500,000. The case for the old line: Killingsworth himself says income is a modest driver of happiness and people probably overrate it. $70,000 set as a floor in 2015, just under the line. $500,000 in 2023, and still no line. Past $100,000, is the next raise worth chasing, or is that where you stop trading time for money?

In April 2024 California's auditor found the state had not consistently tracked what nearly $24 billion on homelessness bought. Five months later Gavin Newsom vetoed a bill to track it. Joe Rogan asked how you spend $24 billion. In the video below he tells it himself: "He opened my eyes to the homelessness thing... there's a whole business behind it." Crediting Colion Noir, he says the people running it make "hundreds of thousands of dollars a year" and "no one's being punished for it." Tucker Carlson's take: "if you pay for something you get more of it." The $24 billion went through at least 30 programs in nine agencies; four of the five budgets were Newsom's. The auditor graded five and found two "likely cost-effective." From 2019 to 2023 the count rose from about 151,000 to 181,000. Five weeks before the audit, voters narrowly approved a $6.38 billion bond for treatment beds and housing. The vetoed bill would have made each program report costs and outcomes yearly. It passed unopposed. Newsom wrote that "similar measures are already in place." $24 billion over five years: about $4.8 billion a year. Three reviewed programs got over $9.4 billion since 2020 and couldn't be graded for lack of data. Homekey, one the auditor could grade: about $144,000 a unit, against $380,000 to $570,000 to build new. The case for the spending: California's count fell 2.8% in 2025, per HUD, and Newsom's office now puts unsheltered homelessness down about 13% since 2024. $24 billion by 2023, results not consistently tracked. A bill to track them, vetoed in 2024. Should a state stop paying for homelessness programs it cannot measure, or keep paying now that the count is falling?

In 2020 Adidas paid Kanye West $191 million in Yeezy royalties. Two years later it cut him, kept the designs and sold the leftover pairs for about $1.5 billion. Kanye West says he was told Adidas would never put him on its board. In the video below he tells it himself: "I'm 43 years old. I was 53 million dollars in debt... four years ago, and now it's proven that I'm the new Michael Jordan of products." He says a man helping him build factories told him flatly, in Malibu: "Adidas will never put you on the board." His reaction: "this wall has to come down." Forbes put his cut at 15% of wholesale, roughly triple Michael Jordan's at Nike. By 2020 Yeezy sold nearly $1.7 billion a year. He tells Rogan he is worth "five billion dollars" (Forbes said $1.3 billion) and still has to "ask him for something." In October 2022 he said on a podcast Adidas could not drop him. On October 25 it did, citing antisemitism and calling his comments and actions "unacceptable, hateful and dangerous." Worked out. His side: $2 billion to $400 million in a day, per Forbes. Adidas: about $1.5 billion of leftover pairs sold over two years, roughly $540 million in operating profit. First drop: orders for about 4 million pairs worth $563 million, about $140 a pair. The case for Adidas: a seat would put the royalty collector on both sides of his own contract. $191 million to him in 2020. Nothing either way in the 2024 settlement. When one partner brings in up to 15% of a company's profit, should he get a board seat, or is that the risk a board exists to keep out?

Kalshi CEO Tarek Mansour says when customers lose, they don't lose it to him. He takes 0.8% of every trade either way. “Why isn't Kalshi a casino in your pocket?” “On Kalshi, when customers lose, they don't lose it to me. They lose it to other customers.” “New York Stock Exchange takes a fee on every trade, right? We take 0.8%, they take something similar.” By InGame's count, that 0.8% came to $430 million during the World Cup. “I have a financial interest in having more users, yes.” Co-founder Luana Lopes Lara, on the 75% who lose money: “...it's never quoted next to how many people lose money day trading.” The CFTC says Trump's teleprompter operator made $107,539 on contracts tied to his speeches: “Manipulation and insider trading exists in all financial markets, the same way that fraud and cheating exists in any system.” Donald Trump Jr., a paid advisor, got a $300,000 Kalshi stake in 2025. Asked what he advises on: “We don't talk about regulatory or policy.” Bloomberg reports Trump Jr. told Republican attorneys general behind closed doors to back off.

On Rogan in 2020, Musk called Buffett's job "kind of a boring job." $10,000 in each man's stock that day ended up far apart. In the video below Musk says it himself: "Let me say Warren Buffett, for example..." He said Buffett reads "a lot of annual reports," that picking whether "Coke or Pepsi deserve more capital" is "kind of a boring job if you ask me," and that it is still "important to figure out." Buffett had answered the Coke question long before. In 1994 Berkshire finished buying 400 million Coca-Cola shares for $1.3 billion. That year Coke paid it $75 million in dividends. In 2022 it paid $704 million. Musk's way paid faster. Tesla's stock rose 743% in 2020, and by January 2021 he was the richest person alive. Success. $10,000 on May 7, 2020, the day the episode aired: Tesla: about $75,300. Berkshire: about $31,600. On the way, Tesla fell 74% between November 2021 and January 2023. Berkshire's worst fall was 27%, in 2022. Coke now pays Berkshire more than half of what the whole stake cost, every year. The case for boring: Berkshire never fell more than 27% in those six years, and Coke's dividend arrives every quarter whatever the stock does. Would you rather own the 7.5x that fell 74% on the way, or the "boring" 3.2x?

In 2013 the CIA picked Amazon for a $600 million secret cloud, though IBM's price was $54 million lower. Edward Snowden says he pitched the CIA a cloud for Dell. The goal, he says: "how much money can we get out of the government." In the video below he tells it himself: "I went to the CIA again, now I was working as a private employee for Dell." He says he sat across from chiefs of "enormous CIA divisions," paired with a salesman, and "Dell ended up getting beat out by Amazon." IBM protested. On June 6, 2013, the day after Snowden's first leak ran, the GAO sided with IBM. In October a judge restored Amazon's deal, writing that IBM "lacked any chance." In 2020 the CIA's next cloud contract, worth "tens of billions," went to five companies, Amazon among them. In 2021 the NSA gave Amazon a secret cloud deal worth up to $10 billion. Snowden says a top-secret clearance plus tech skills paid "a ridiculous amount of money for doing very little work." Booz Allen, which fired him in 2013, put his salary at $122,000. The Guardian had reported $200,000. Worked out. Amazon's evaluated price: $148 million. IBM's: $94 million. The CIA chose the price 57% higher. $600 million in 2013, up to $10 billion in 2021: about 17 times. The case for paying more: the CIA rated IBM's proposal high risk and Amazon's low, and the judge found IBM lost on quality, not price. $54 million cheaper, turned down in 2013. Up to $10 billion in 2021. When a spy agency pays 57% more for the bid it trusts, is it buying safety or getting sold?

In 2023 the FCC denied Elon Musk's Starlink $885 million, partly citing Starship's failed launches. Ten months later the tower caught Starship's 233-foot booster. Donald Trump says he forgot an important caller to phone Musk. In the video below Trump tells it himself: "I'm talking to this very important guy, I said wait a minute, I'm looking at something." He says he put the man on hold, watched "the two arms grab it," forgot him and called Musk to ask if that was him. "That was me." Starlink had won the $885 million in a 2020 FCC auction, to connect 642,925 rural homes and businesses. The FCC later called it a "still developing technology" that needed a $600 dish. Congress set aside $42.45 billion for rural internet in 2021. By October 2024 it had not connected a single home. Then Hurricane Helene cut off western North Carolina. Trump says a local asked if he knew Musk, he called, and Musk "got it for him so fast." FEMA says Trump had no part in its decision to send Starlink. Worked out. June 2025: Trump's Commerce Department rewrote the rules so the cheapest bid wins. Starlink won about $734 million for 473,000 addresses, roughly $1,550 each. In Texas the top fiber winner got about $12,700 per address. Starlink: $1,700. The case for fiber: a cable lasts decades and carries far more data, while a satellite's capacity is shared by everyone under it. $885 million denied in 2023, partly over a rocket that kept failing. $734 million awarded in 2025, after the tower caught it. Should public money pay for cables that last decades, or for dishes that work next month?

In 1987 Ken Griffin started trading in his Harvard dorm room with $265,000, some of it his grandmother's. Citadel has since made $90 billion for its investors, more than any hedge fund ever. On Tuesday he said there are "almost certainly bubbles" in AI. Then he named his regret from the last one: "I just wish I'd kept my Amazon stock." - Amazon, December 1999 to September 2001: down 94% - It didn't close above its 1999 high again until October 2009 - $1,000 bought at the 1999 top and never sold: about $47,000 today - Citadel in 2008: its two biggest funds lost 55%, and Griffin blocked withdrawals for 10 months - Citadel in 2022: $16 billion for investors, the largest annual gain in hedge fund history His frame: "any radical transformation of our economy is generally intertwined with a bubble." The companies that make it through lead their markets for decades. Keeping Amazon meant watching it fall 94% first.

In 1987 Howard Schultz had 60 days to raise $3.8 million and buy Starbucks. A month in, one of his own investors bid $4 million behind his back. So a friend took him to a 6-foot-7 lawyer named Bill Gates Sr. Schultz kept this story quiet for years. Below, he repeats what Gates Sr. said across the rival's desk. In 1981 a small Seattle shop ordered so many of the Swedish coffee cones Schultz sold in New York that he went to see it. Starbucks hired him, then refused his espresso bars. In 1985 he left to open his own cafes. In 1987 founder Jerry Baldwin put the six Starbucks stores up for sale. Schultz had no money, only an exclusive. He had half when Baldwin told him an early investor in Schultz's own cafes had gone around him: $4 million, no due diligence. Schultz calls the man a Seattle titan. Worked out. The rival offered $200,000 more, about 5%. Starbucks is worth more than $110 billion today. About 29,000 times the 1987 price, roughly 30% a year for 39 years. Schultz has never named the rival. He says the family still lives in Seattle. The case for the rival: the exclusive was Baldwin's promise, not his. He offered more, with no conditions. This is Schultz's account. Outside, Gates Sr. told him: "my son and I are going to help you." He never told anyone, Schultz says. His son, Bill Gates, heard the story 30 years later. $4 million lost to $3.8 million and a walk across the street. Was Gates Sr. protecting a kid's dream, or using his clout to kill a better offer?

January 27, 2025: DeepSeek's low-cost model wiped $593 billion off Nvidia in one day. Cheap AI was supposed to kill demand for the chips. Gavin Baker, who took a Fidelity tech fund up 291%, explained in Berlin this week why it went the other way: "The price of tokens has gone down... the cost to rent an hour of GPU has gone vertical." - B300 rent: up 38% since token prices peaked on May 31 - DeepSeek, Oct 20: $38 of revenue per GPU hour, at its off-peak price - Baker's math on that: about $100B a year per gigawatt, double at peak - SpaceX, same day: compute priced above $50 a watt, paying for itself in under a year - Musk on that call: old H100s now rent for more than they did new $38 an hour is $332,880 a year from one chip. Baker's frame: squeeze more useful tokens out of a GPU hour and the hour is worth more. The cheaper the token, the more a GPU earns.

In September 2000 Elon Musk left X.com for a fundraising trip that doubled as his honeymoon. While he was away, his own team got the board to replace him with Peter Thiel. The company became PayPal and sold for $1.5 billion. In the video below, filmed in 2008, Musk tells it himself. Listen to what he says about leaving the office. Musk started X.com in March 1999, at 27. A year later it merged with Confinity, the startup of Thiel and Max Levchin, which owned a payments product called PayPal. Musk ran the merged company and wanted X.com on everything. His colleagues had secretly tested both brands and found PayPal stronger. He didn't fight the vote. Worked out. eBay paid $1.5 billion in 2002. Musk, still the largest shareholder, got about $176 million for his 11.7%. He put $100 million of it into SpaceX. PayPal today: about $50 billion. SpaceX: $2.14 trillion, 43 times more. The case for the board. The dot-com crash had started and the CEO left for two weeks. Even Musk says in the clip it was the wrong moment to go. His side, as he later told his biographer: "Of course, if I had stayed, PayPal would be a trillion-dollar company." In August 2008, the week SpaceX lost its third rocket in a row, Thiel's Founders Fund put $20 million into it. They took the name X off his company in 2001. In 2023 he put it on the app you're reading this on. Was the board right to replace a CEO who left in the middle of a crash, or did they fire the man who would build a company worth 43 PayPals?

Michael Burry made his investors more than $700 million on a trade against American mortgages. Goldman Sachs took the other side. Two years later, he says, Goldman wanted a bigger piece of the same trade. The Big Short made $133 million at the box office telling that story. Burry told it himself in 2011, free, in 31 minutes at Vanderbilt, where he earned his medical degree. Christian Bale got an Oscar nomination for playing him. The real talk has about 300,000 views in 15 years. Burry is a medical doctor. He told the room he was "definitely 50th percentile here at Vanderbilt." He learned the mortgage market from the filings Wall Street was required to publish: "At times I felt I was the only one reading these things." Strip out the history and five rules are left. Read the filings. Subprime loans sold to investors came with mandatory public filings. By summer 2005 they showed interest-only loans at over 40% of the subprime loans passing through Wall Street, up from 10% a year earlier. Watch the payment, not the rate. With rates at 40-year lows, lenders started selling the monthly payment instead: interest-only loans, then loans where the borrower could "pay next to nothing each month." He timed the crash to 2007, when most of those teaser periods would reset. Easy credit marks the top. "The point at which the provision of credit was most lax, in my mind, would mark the point of maximal price in the asset." Growth with falling quality is a warning. In 2004 Countrywide's subprime lending rose 158% while its total lending fell 24%. The top job sees it last. At nearly every failed bank and every agency, he says, someone saw it as clearly as he did. "None, zero, zero were in the top job." Hank Paulson came to Treasury from the top job at Goldman Sachs and ran the rescues of AIG, Fannie Mae and Freddie Mac. AIG alone got up to $182 billion of public money. Burry closed the talk with one piece of advice: "I might recommend opening a bank account in Canada." The talk is free on Vanderbilt's YouTube channel. The filings he read were public the whole time. His edge was reading the pages everyone skipped. Before you sign your next loan, read the page that says what the payment becomes, not what it starts at.

In May 1998 Bill Gates, the richest man alive, sat next to Warren Buffett and told 350 students that tech stocks, his own included, should be valued below Coca-Cola. The Nasdaq co-sponsored the broadcast. Over 21 months it climbed from under 2,000 to 5,048, then lost 78%. Thirty-one months after the peak it sat lower than on the day Gates said it. It needed until April 2015 to see 5,048 again. PBS filmed that afternoon at the University of Washington to air during pledge drives. The upload has about 6,000 views. Fortune put Gates at about $48 billion that day and Buffett at $36 billion. Buffett walked on in a red Nebraska shirt Gates had handed him backstage. "Every internet stock has gone crazy here in recent months, but I don't know how to look at one or the other and tell the winners." That is Buffett, answering a student. Gates went further: "I think the multiples of technology stocks should be quite a bit lower than the multiples of stocks like Coke and Gillette, because we are subject to complete change in the rules." The rest of the hour comes down to six rules. Horsepower. Buffett told the room that everyone in it had more IQ than his job needs. Plenty of people own a 400-horsepower motor and get 100 out of it. A 200-horsepower motor that delivers all 200 wins. The tiptoe. When technology makes every company more efficient, the gain gets competed away. One person at a parade stands on tiptoe, and thirty seconds later everyone does. Sore legs, same view. People. Buffett has turned down decent deals because he disliked the people. That, he said, is like marrying for money: a bad idea anyway, and absolutely crazy if you're already rich. A year of payroll. Gates kept enough cash to pay every employee for a full year with no money coming in. In 1998 that was about $10 billion. Other people's money. Buffett said losing everything he owned wouldn't bother him much. Taking other people down with him would destroy him. Predictable beats exciting. Buffett wants businesses he can picture in 20 years. The internet won't change how people chew gum, drink Coke or shave. Gates expected Microsoft to face at least three crises in ten years. Microsoft's share price didn't set a new high until October 2016, seventeen years after its 1999 peak. Fortune ran the full transcript in its July 1998 issue. Gates priced three crises a decade into his own company. Price at least one into the next tech stock you buy.

In 2012 the man running the world's largest hedge fund took apart the "balanced" portfolio advisors charge 1 percent a year to build. The video has fewer than 300,000 views in 14 years. The talk is free. Bridgewater's SEC filings set its All Weather minimum fee at $500,000. Bridgewater's founder is Ray Dalio. He started the firm in his two-bedroom New York apartment in 1975. By the end of 2011 its Pure Alpha fund had made clients $35.8 billion after fees, more than any hedge fund in history. In 2008, the year the S&P 500 lost 37 percent, it finished up about 9 percent. The full 67 minutes are Dalio at the Council on Foreign Relations on September 12, 2012, with Maria Bartiromo, then at CNBC, asking the questions. His whole portfolio logic runs on two numbers: growth and inflation. Assume you don't know the future. Pick the mix before you have an opinion about next year. Four seasons. Growth can come in above or below what markets expect, and so can inflation. Own something that holds up in each. One ruinous asset class per generation, and nobody knows in advance which one. Japanese stocks needed 34 years to win back their 1989 peak. Gold is a currency. Hold some next to your cash. He floated about 10 percent. Balance risk, not dollars. Stocks swing more than twice as hard as bonds, so a portfolio split by dollars ends up riding on one asset. State Street, which runs Bridgewater's All Weather ETF, puts the stock share of a 60/40 portfolio's risk at more than 90 percent. Bartiromo asked how much of a portfolio belongs in gold versus stocks. Dalio answered that most people should not play his game at all: "They're going to lose at the poker table." That is the man whose fund had just made clients $13.8 billion in one year, describing his own table. Pension funds and central banks pay Bridgewater to spread risk across all four seasons. The textbook advisor portfolio, at 1 percent a year, still splits by dollars: 60/40. The full talk is free on the Council's YouTube channel, and cfr.org still hosts the complete transcript. Before you sign your next "balanced" allocation, ask what share of its risk sits in stocks. At 90 percent, it is a stock portfolio with a nicer name.

Stanford, 2014. A founder turned down by 60 VC firms teaches a class how to raise money. Next to him sit Marc Andreessen and Ron Conway. Both already had money in his company. Seven months later investors gave it $500 million at a $4.5 billion valuation. Then regulators learned how his salespeople were getting their licenses. listen from 2:25, the whole playbook in one sentence "if you guys were the Twitter guys... put whatever up here... we would invest in you" "If you can be like the Twitter guys, like nothing else matters" His name is Parker Conrad. The company was Zenefits. California makes anyone selling health insurance complete a 52-hour course first. The SEC found that a script Conrad wrote let his staff spend less time on it than the law required, and that investors were never told. He resigned in February 2016 and the valuation was cut to $2 billion. That same year he founded Rippling. In 2017 he settled with the SEC for $533,692, without admitting or denying the findings. Rippling's last funding round valued it at $16.8 billion. A fine worth a hundredth of a percent of the prize stops nobody. It goes in the budget.

In 1994, Warren Buffett was asked about Nike. He said he understood his own shoe companies better. One was Dexter. Months earlier he had paid $433 million for it, in Berkshire stock. Dexter went to zero. Those 25,203 Berkshire shares are worth $19.9 billion today. Nike, the company he couldn't read, went up more than 100x from that meeting to its 2021 peak. Buffett's exact words that day: "I don't understand their competitive position and the likelihood of permanence of their competitive position over a 10 or 20 year period as well as I think I understand the position of Brown and Dexter." Dexter's permanence lasted a few years. He later called it the worst deal he ever made. In 1998 he was asked about Nike again. He called Phil Knight "a terrific operator" and kept his view of the stock to himself. In 2014 he was asked about sports companies. His answer: "Sports equipment has generally not been a very good business." Nike, he added, had done incredibly well. Nike closed at $36.49 the day before that meeting. Today it trades around $35, 80% below its 2021 peak. And the stock he paid for Dexter with? Since the 1994 meeting, Berkshire is up 51x. Nike, dividends included, about 29x. The shoe company he understood went to zero. The one he couldn't read went up 100x, then gave back 80%. Was Buffett right to stay out of what he couldn't understand, or did he just understand the wrong shoe company?

In late 2008, with the banks still falling, an MIT professor told his class that making money hits the same brain circuit as cocaine. It was his last class of the semester, weeks after Congress approved a $700 billion bailout. Almost nobody with a brokerage app on their phone has sat through it. His name is Andrew Lo. He teaches finance at MIT Sloan and founded the investment firm AlphaSimplex in 1999. In 2012 Time named him one of the 100 most influential people in the world, for the theory he lays out in this lecture. The video is that full 54-minute class. Most of it is about why people can't stop taking risk while they are winning. The end runs the whole course past in about ten minutes, which he calls "a near-death experience." Take away the 2008 headlines and it comes down to six rules. Profit is an anesthetic. The longer you win, the fewer hard questions you ask. Pain protects. A loss is what makes you pull back. Numb it with a winning streak and nothing does. Decide before you are winning. A limit you plan to set during a hot streak never gets set. Pay for the sprinkler. Nobody would pay extra for fire sprinklers by choice, because everyone prices the fire at zero. That is why it has to be a rule. Ask who is left. A market is only as sharp as the people still trading in it. In the late 1990s that was the librarian and the pharmacist on E-Trade, and the market got measurably less efficient. Survive first. In his framework, survival is the only thing that matters. Lose 50 percent and you need a 100 percent gain just to get back to even. "This is not an analogy or a metaphor. It's a physiological fact." That is Lo on the cocaine comparison, in front of his class in 2008. He also tells them about a friend's family worth around $300 million. Their 95-year-old grandmother still takes the bus, because she remembers standing in line during the Depression, not knowing if there would be dinner left by her turn. Bad years rewire people for life. His point is that good years do too, in the other direction. The whole course is free on MIT OpenCourseWare. Write your sell rule before your next position turns green. After that, the dopamine is writing it.