researchers at Anthropic, who made up a third of Britain’s maths Olympiad team in 2017, to resign. Anthropic’s boss, Dario Amodei, was once a member of the squad, but not quite the starting line-up, of America’s physics side. OpenAI has responded by appointing Paul Christiano to its board. Mr Christiano has a long history in AI safety and a silver medal for maths. Davos Man cut his teeth at McKinsey in his 20s, but Olympiad Man was acing maths competitions before his first drink. (Olympiad Woman is far rarer.) Very clever children often become very, very expensive adults. In 2024 Google paid $2.7bn to hire Noam Shazeer, who achieved a perfect maths score in 1994. He recently decamped to OpenAI, whose head researcher has moonlighted as head coach of America’s informatics team. In the 2010s Sheryl Sandberg, the archetypal Davos Woman, was Meta’s best- paid executive; now it is probably Alexandr Wang, aged 29 and not long off America’s problem-set circuit. AI startups counting Olympiad Men among their founders include Databricks (worth $190bn), Cursor ($60bn), Cognition ($48bn) and Perplexity ($30bn). Olympiad Man is just as comfortable in financial markets, where the broad shoulders of pit trading long ago gave way to the narrow ones that build algorithms to buy and sell at the speed of light. One of the founders of Two Sigma, a big trading firm, represented America in maths. Jane Street, another trading giant and one of Olympiad Man’s favourite employers, is even besting banks in corporate-bond trading, perhaps the last vestige of the Wall Street jock, and dabbles in venture capital (presumably in the startups of fellow Olympians). Cryptocurrency has plenty of illiterate spivs, but also numerate swots. The founder of the Ethereum blockchain took bronze in informatics; that of Hyperliquid, a big exchange, gold in physics. Like their sporting namesakes, Olympiad victories are a worthy source of national pride. But the business of rearing and attracting Olympiad Men is also a source of national power and wealth. That was true when they mostly stayed in universities. It is doubly so now that many work for a living. They move in herds. The Massachusetts Institute of Technology is one magnet. Trinity College, Cambridge is another. (Both institutions are islands of excellence in towns obsessed with equality.) Deedy Das, an investor at Menlo Ventures, examined the maths, informatics and physics competitors

of the past 25 years. Google employs the most. Meta is a distant second. As well as Jane Street, Jump Trading is a popular employer in finance. There are good reasons to prefer Olympiad Man to Davos Man, the slick corporate executive who spent decades at home anywhere but nowadays feels welcome nowhere. How does one become a “Young Global Leader” at the World Economic Forum? God knows. How does one become an Olympiad Man and land a big job at an AI firm? You’re only ever two tests away, pal. The ability to audit one’s power has advantages in societies which are both contemptuous and envious of elites. It is, though, a great irony that the free market now awards its grandest prizes through a system designed by communists: the first Olympiad was held in the Soviet bloc; America started competing only in 1974. Olympiad Man’s vision of globalisation is more attractive than Davos Man’s. According to Tolga Yuret, a Turkish economist, more than half of those who participated in Olympiads in the 2000s ended up living abroad. The West, and America in particular, is an enthusiastic importer of Olympiad talent, which might start life in Romania, China or India. Studying Olympiad Man is also a corrective to bad views on immigration. Does the left not wonder whether they should spend more time attracting Olympians rather than delivery drivers? Does the right not see the outstanding achievements of immigrants in its national teams? To look only at their names, and not their star-spangled ties, it would be easy to mistake America’s maths team for the Chinese one. William F. Buckley once remarked that he would rather be governed by names drawn from a telephone directory than from a list of faculty at Harvard. What about a register of Olympiad Men? By necessity, elites are a class apart from their subjects. But Olympiad Man is truly abstracted from the world. He has few employees to speak of and even fewer friends who do not share some version of his background. His rationalist philosophy would strike ordinary folk as a mark of insanity (AI labs and hedge funds are full of effective altruists). His language is unintelligible. Not because it is complicated—AI researchers talk, often glibly, about “p doom”, the probability that their technology ends the world—but because it is not

believed. Investors think Olympiad Man a worrier. Politicians think him naive about global affairs. The populace is yet to fully comprehend Olympiad Man or his power. Until he can demonstrate virtue commensurate with his brilliance, Olympiad Man should want to keep it that way. ■ This article was downloaded by zlibrary from https://www.economist.com/business/2026/09/16/olympiad-man-has-displaced-davos-man

· Finance & economics

Soaring bond yields, gaping deficits and towering debts: what could go wrong? Scott Bessent should stop blaming the bond market What Africa’s biggest IPO can do for its stock markets The trickle-down economics of the AI boom The GDP-per-person slowdown threatens global living standards Ricardo Reis was too frank for the IMF

Finance & economics | Bonds-and-chain Soaring bond yields, gaping deficits and towering debts: what could go wrong? The rich world is flirting with fiscal disaster Sep 17th 2026 ACROSS THE rich world, investors are making governments pay up. Bond yields have risen in Britain, France, Germany and Japan. On September 14th the yield on ten-year American Treasuries exceeded 5%, the highest it has been in two decades. The median yield on rich countries’ ten-year government bonds is above 4%, its highest in more than 15 years and roughly five times its average in 2015-21 (see chart 1).

More worrying, borrowing costs are rising at a time when countries’ debts are higher than ever. Gross public debt as a share of GDP in advanced economies stands near 110%, up from around 70% in the early 2000s. In America it has more than doubled over that period; in Britain it has nearly tripled. There is little appetite for belt-tightening. America is expected to run a deficit of around 6% of GDP this year and France of more than 5%. Throughout much of the 2010s politicians could afford to dismiss warnings of a debt reckoning, thanks to rock-bottom interest rates. Now rates are rising. On September 16th the Federal Reserve raised its benchmark rate by a quarter of a percentage point. The European Central Bank (ECB) did so the week before. And the bond market is signalling that the era of cheap government finance is over. Big borrowers are about to discover just how painful those debt piles can be when interest rates can no longer be ignored. Governments have dealt both with high yields and with high borrowing in the past. But, at least in recent decades, not at the same time. In 2007, when the median yield on rich-world bonds was around today’s level, governments had to sell debt worth only around 11% of GDP. This year they will have to peddle more than twice as much, to finance those wide deficits and to replace debt issued when rates were far lower.

Interest payments already consume more than 3% of GDP across the OECD, a club of mostly rich countries, and nearly 5% in America (see chart 2). Much of the debt stock still carries the low rates of the old era. As that debt matures, the bill will climb. If borrowing costs stay close to where they are today, Uncle Sam’s annual interest payments could nearly triple to $2.7trn by the end of the decade, according to the Committee for a Responsible Federal Budget, a non-partisan think-tank. That is more than America spends on Medicare (to keep old people healthy) or Social Security (to pay them pensions). Yields on short-term government debt mostly reflect where investors think central-bank rates are heading. Lend for decades, however, and they are less willing to look past inflation risk and the state of the public finances. They kept paring back their holdings of long-dated bonds even after most central banks in large rich economies began cutting rates in 2024 (yields climb as bond prices fall).

The OECD reckons that by the end of 2025 the average “term premium”— the extra return investors demand for lending long rather than short—had risen by more than one percentage point from its level before the covid-19 pandemic shook markets in 2020 (see chart 3). This year it has kept climbing. In America, the term premium is now estimated to be at its highest in more than a decade. Growing doubts about America’s fiscal outlook in particular are dulling Treasuries’ special lustre. Investors used to accept lower yields on American government bonds than on close substitutes because they were exceptionally safe and liquid. No longer. Analysis by Lira Mota of the Massachusetts Institute of Technology, updated by Hanno Lustig of Stanford University, compares Treasuries with highest-rated corporate bonds after stripping out their default risk and finds that this “convenience yield”—the return investors are willing to forgo for holding Treasuries—had all but disappeared by July this year. That is an expensive privilege to lose when America must raise trillions of dollars in the bond market annually. There are reasons to think that high long-term yields are here to stay. The multitrillion-dollar investment boom in artificial-intelligence infrastructure needs capital, and is competing with governments for investors’ attention. At the same time, some previously reliable buyers are disappearing. Central