The founder of TSMC got no equity in the company. He was 56 years old. He's worth ~$10B today from buying TSMC stock with his own money. Ben Thompson called him "a one of one" who belongs on the Mount Rushmore of the greatest tech executives of all time. This is his story. Morris Chang was born in Ningbo, China in 1931 and spent much of his childhood fleeing war. At 18, he arrived at Harvard as the only Chinese student in a freshman class of over 1,000. In 1955 he had two job offers, one from Ford and one from a transistor maker that paid $1 more per month. Ford refused to match, so Chang took the other job. He spent the next 25 years at Texas Instruments (TI) and eventually rose to run its worldwide semiconductor business. Then his ascent stalled. TI moved him out of semiconductors and into a job he later described as being "put out to pasture." Two years later, he quit. He landed next at General Instrument in New York, living on the 53rd floor of Trump Tower, where Donald Trump was his neighbor. The job lasted about a year. It was there, in 1984, that Gordon Campbell asked him for $50 million to start a chip company. A few weeks later, Campbell called back and said he only needed $5 million. Why? "I'm not going to build a fab." Taiwan had been trying to recruit Chang for years, determined to move beyond cheap plastics and running shoes and build a high-tech economy. The government had asked him outright to run its industrial research institute, ITRI, back in 1982, while he was still at TI, and he'd said no. After General Instrument let him go, it asked again. In 1985, he said yes. Donald Lee-Brown, who runs research at Positive Sum, likes to say the very best entrepreneurs can divine where the puck is going. Chang saw that chip design and chip manufacturing were going to split, and someone would need to do the manufacturing for everyone else. On his second day at ITRI, he inherited requests from three chip companies, each asking the institute to help fund its own fab. Instead of building three factories, Chang proposed building one they could all share. That idea became TSMC. Taiwan put in $70 million for 48% of the company. Philips invested $40 million, and private investors contributed another $35 million. Chang, despite founding and running it, got none of the equity. Intel, TI, Motorola, AMD, Panasonic, and Sony all passed on funding it. TSMC grew revenue 49% annually for its first decade. By 2000, most of its revenue came from fabless companies like Qualcomm, Broadcom, and Nvidia. In 1997 Chang got a letter from the CEO of a four-year-old company that was facing bankruptcy and whose calls TSMC's sales office had been ignoring. Chang called the number printed on the stationery. Jensen Huang picked up amid a bunch of shouting, then said to his team, "Quiet, Morris Chang is calling me." Chang stepped down as TSMC's CEO in 2005. Four years later, at 77, he came back. One of the first problems waiting for him was Nvidia. A manufacturing failure had damaged the relationship, and TSMC's management had refused Jensen’s request for compensation. Chang emailed Jensen and told him he would be at his house at six. Salad and pizza at 6:30. Business at eight sharp. In Jensen's study he offered more than $100 million, good for 48 hours, no negotiating. Jensen accepted it in two days. Back in the CEO seat, Chang nearly tripled capital spending in a single year to own the 28-nanometer node. In his memoir he quotes Shakespeare on the decision, "there is a tide in the affairs of men which, taken at the flood, leads on to fortune." The tide turned out to be the smartphone. And then Apple literally showed up at his front door. Chang's wife, Sophie, is the second cousin of Foxconn founder Terry Gou. One evening in 2010, Gou invited himself to dinner and brought Apple’s COO, Jeff Williams. Eventually, over lunch at Apple, Tim Cook told Chang: "Intel just does not know how to be a foundry." Today, TSMC manufactures the A- and M-series processors at the heart of Apple's iPhones, iPads, and Macs. When Chang founded TSMC, the global semiconductor market was about $26 billion. Last year it reached $795 billion. Roughly 30x since 1987. Philips did even better. Chang estimated that by the time Philips sold its final TSMC shares in 2008, it had earned 135x its original investment.

Gavin Baker from December 2025: "It is interesting that Tesla, SpaceX and xAI are all converging. So xAI will be the intelligence module for Optimus made by Tesla with Tesla Vision as its perception system. And then SpaceX will have the data centers in space that will power a lot of the AI, presumably for xAI and Tesla and the Optimuses and a lot of other companies. And it's just interesting the way that they're converging. And each one is kind of creating competitive advantage for the other."

Anthropic's draft S-1, per the WSJ: - H1 2026 revenue: $16.3B (vs. $1.1B in H1 2025) - H1 net loss: $80.7B ($72B of it non-cash) - H1 compute spend: nearly $9B (vs. $7.3B in all of 2025) - Future infrastructure commitments: $600B+ - Cloud deals: $111B Google, $110B AWS, $31B Microsoft - Two customers: ~20% of H1 revenue - The Amodeis pledged 80% of their equity to philanthropy

Christina Cacioppo on what makes them worth 10x: "High agency, get shit done. That sounds so milquetoast, and like that's worth millions of dollars? But yes, actually it is. A high-context person who understands the company and the customers and the domain, making generally reasonable decisions and sometimes bad ones, and then iterating on their bad ones and just moving is so valuable."

Christina Cacioppo's personal site has a list of 30+ products she built that failed before starting Vanta (worth $4B+). "The function of the overwhelming majority of your artwork is simply to teach you how to make the small fraction of your artwork that soars."

Walter Russell Mead (@wrmead) shared a dispatch from Ukraine when we recorded with him in September: "I was back there a couple of months ago, and...I talked to people who said that they don't treat bullet wounds anymore because it's all drones. The Ukraine war is no longer a battalion over here shooting at a battalion over there. With these drones, you have this gray zone, where at most you'd have a couple of soldiers together. One of the hardest problems in warfare now is, how do you evacuate someone who's wounded? They're having to build a new kind of ground drone that can go retrieve a person, cover them with a metal shield, and wheel them out. Even to do this, you need a whole team of drones working with this thing because they're looking for other drones that would try to attack this. They tell me they had one soldier that it took 63 days to get evacuated after a wound. This is so unlike what war was even six years ago. If the tank and the rifle are becoming as obsolete as the cavalry charge and the sword, what kind of forces do you need? We're actually the most vulnerable to technological change, because we've got so many tanks. If the tank becomes useless, that's more of a problem for us than it is for somebody who's figured out the tank killer."

David Vélez on how Nubank, a company with over 10,000 employees, responds to AI in real time: "Opus 4.6 came in in December of 2025, and that was the moment that a lot of people suddenly realized, 'This actually works.' And we didn't catch that moment. We only caught it 60 days after. 60 days made a huge amount of difference between us and best-in-class companies that were using AI. So we went back to Brazil, and went code red around the entire organization. We mapped out 120 different workflows across the entire organization. Every part of the organization, 120 workflows were reinvented natively using AI tools."

Berkshire Hathaway just added another $330 million to its investment in Bending Spoons, bringing its total commitment to roughly $930 million for a 4% stake in the company. Bending Spoons is up nearly 40% today.