The Bay Table: analysis and commentary on the business of technology

The world this week

Briefs from the issue of 17 October 2026

• Student founders. Student debt makes skipping university look shrewd, but the best evidence says successful founders are closer to 45 than 19, so the case for dropping out is narrower than its fans claim. Read more on Tune in, drop out ›
• WeWork. The collapse of WeWork's flotation is not a failure of public markets but their finest hour: investors read the prospectus and declined. Read more on Out of office ›
• Campuses close. The class of 2020 is learning that a degree bundles a lecture with a place and a crowd, and universities should now say which parts of the bill they are still delivering. Read more on Zoom, zoom, zoom ›
• Foreign students and America. America nearly told 1.1m international students to go home over a technicality, and the founders it would have lost are worth far more than the rule's tidiness. Read more on Leave to remain ›
• Quibi. Quibi's failure is not a verdict on short video but on a business that spent $1.75bn before asking anyone what they wanted. Read more on Short, sharp, shut ›
• Retail trading. Robinhood's curbs on GameStop were forced by plumbing, not conspiracy, but a market whose settlement lags its apps has no business promising instant access. Read more on Short changed ›
• NFTs and ownership. The $69.3m paid for Beeple's collage buys a verifiable record of who owns a picture the whole world can still look at, which is a real novelty but not remotely worth the price, and buyers should insist on being told what the receipt actually covers Read more on Right-click, save ›
• Archegos and the banks. The collapse of Archegos shows that banks, not regulators, were the only ones who could see a family office's leveraged swaps, and the law should make them say what they see Read more on Margin for error ›
• SPACs and the SEC. The SEC staff's warrant ruling is a footnote with teeth, and the SPAC boom should be allowed to cool under it. Read more on Warrant officers ›
• Coinbase's listing. Coinbase has shown that a company can go public without paying a bank to guess its price, but the first day showed how much guessing the market itself does. Read more on Going direct ›
• Venture capital. Startups have raised more in nine months than in any full year before, and the boom rests on a price of money that is about to change. Read more on The roaring 2021s ›
• Rivian's listing. Investors who pay a Detroit-sized price for a firm that has delivered 156 pickups are buying a story, and should demand that the story be audited like a factory. Read more on If you build it ›
• Algorithmic stablecoins. Congress should stop letting tokens that promise a dollar without holding one call themselves stable, and write rules by what a coin promises rather than by who issues it. Read more on Dollar days ›
• FTX. FTX's collapse needs no new crypto law, only the enforcement of old rules about whose money is whose, and a little more suspicion from investors. Read more on Token gesture ›
• ChatGPT in schools. New York's ban on ChatGPT protects nothing it claims to protect, and schools would do better to teach pupils to catch the bot's mistakes than to pretend it is not there. Read more on Cheat sheet ›
• Skipping university. Americans are leaving college and starting firms at record rates, and cheap AI will make that trade more tempting, but a degree remains the better bet for most would-be founders. Read more on Class dismissed ›
• Silicon Valley Bank. Washington was right to protect SVB's depositors, but startups should learn the lesson that rescue cannot teach them. Read more on A run for their money ›
• Student loans and the courts. The Supreme Court was right to stop the president cancelling $400bn of student debt by himself, but Congress should now fix the lending that created the debt instead of sending taxpayers the bill. Read more on Debt of honour ›
• Instacart's IPO. Instacart's listing proves that the window for technology flotations is open again, but only for firms willing to accept what the market will pay. Read more on Cart blanche ›
• Bitcoin ETFs. The SEC was right to let spot bitcoin ETFs trade, but it should stop pretending a court did not force its hand. Read more on Coin of the realm ›
• Reddit's IPO. Reddit was right to let its users buy in, but a flotation that leaves control with insiders is a poll in which only some votes count. Read more on Upvoted ›
• California's school phones. California is right to make every district write a phone policy, and right to stop there; the next generation of founders will need its attention more than its apps. Read more on Calling time ›
• Australia's social-media ban. Australia's under-16 ban answers a real collective-action problem, but it fines platforms for an age check nobody has yet shown can work, and should have waited for the trial. Read more on Sweet sixteen ›
• Thiel Fellowship. Peter Thiel has doubled the price of dropping out just as AI makes small teams mighty, yet students are drifting back to campus, and on the evidence they are right to. Read more on Double or quits ›
• Initial public offerings. Figma's dazzling debut shows that public investors want good startups, and that bankers still struggle to price them. Read more on Pop idol ›
• AI and young workers. A Stanford study finds that young workers in AI-exposed jobs are being hired less, which is worth taking seriously, but the evidence is too thin to justify rules against the technology. Read more on Canary in the cubicle ›
• Klarna and credit. Klarna's listing at a third of its private-market peak shows that buy-now-pay-later is a lender's business, and its customers and regulators should treat it as one Read more on Owe, owe, owe ›
• Teens and chatbots. More than half of American teenagers now use chatbots for schoolwork, and schools that treat this as a cheating epidemic are measuring the wrong thing. Read more on Dog ate my homework ›
• Anysphere and Cursor. Four MIT students built one of the fastest-growing software firms on record, then sold it to a buyer with its own models, which shows how thin the moat was. Read more on Cursor on the quad ›
• Prediction markets. Raising the minimum age on Kalshi to 21 would be tidy and mostly beside the point, because the product is a sportsbook with a federal licence and its youngest customers are its smallest problem. Read more on Swap meet ›
• AI side hustles. The surest earner in the AI-business boom is the person selling the course, and young would-be founders should price the pitch by the regulators' record, not by the screenshots. Read more on Selling the shovels ›
• Teen fintech. Banking apps for teenagers teach money management about as well as financial education ever has, which is to say barely, and their real product is a customer who stays until 18 and beyond. Read more on Cards on the table ›

All the week's briefs ›