Carlsen at 35 in the AI era: the king chooses to win faster

At 5:40 in the morning in Binh Duong, I opened my laptop before the first truck crossed the highway. The report I was reading was a translation from Marca, built on Magnus Carlsen's interview with AS dated October 21. Among the many lines about preparation, formats and prize money, one sentence made me put my coffee down: “AI shows that humans are not that good.” I sat with that line for a while. Twenty-one years around the board — as a player, an organiser, a commentator — I have heard every version of the machine-surpasses-human story. But this time the speaker was the world number one of fifteen consecutive years, and he spoke like a man who had accepted a limit rather than a man who had lost an argument. What drew me more than the content was the way he placed two clauses side by side. On one hand, AI makes chess “more interesting”, for fans and professionals alike. On the other, the same technology makes it “increasingly difficult to create an advantage over opponents”. Two sentences, one subject, and a gap in the middle that the original article leaves untouched. Carlsen is 35. He has held the world number one ranking for fifteen straight years and won five classical world titles. In 2026 he declined to defend the crown, letting a championship match take place without the strongest player in it. Asked about a return, he said he is interested in many other aspects of chess more than in a title match, though he does not oppose the idea — provided the prize fund sits around 20 million USD. The winner's share of the current world final, according to the article, is “just over 1 million USD”. A twentyfold gap. I read that number three times, not because it shocked me, but because it was so specific. A player holding the highest position in the rating system was pricing his sport's crown with a simple division. There is a detail worth noting for Vietnamese fans in November. From 10 to 21 November, Budapest hosts a “World All-Around Championship” gathering 24 of the strongest players, including Le Quang Liem. The format is mixed — classical, rapid and blitz combined into a single all-round champion. The event sits outside FIDE's world championship cycle. Three data points — AI flattening preparation, a body slowing down, and a new event placed beside an old prize — fit together into a fairly clear technical story. First, the flattening. When every elite player uses the same strongest engines, the marginal return on deep home preparation falls. What once created separation — hundreds of hours grinding opening lines with a team of seconds — becomes high cost, low differentiation. Carlsen says he has moved to preparation built on “more general ideas” rather than specific lines for specific opponents. That is a rational cost-benefit adjustment, not a decline in ability. But it is also a structural consequence: the heavyweight second-team model only made sense when attached to a championship cycle he left in 2026. Second, winning faster. This is technically the most interesting shift. If his edge in five-to-six-hour games has eroded with age, compressing the game into earlier resolution converts what remains — opening knowledge, practical decision quality — into points before stamina becomes the binding constraint. I remember watching him in classical events at his peak: he did not win with a single brilliant blow, he won by letting opponents exhaust themselves around move sixty. That model needs a body that does not tire. At 35, the body no longer answers on schedule. Third, the format. A 24-player, three-time-control, eleven-day event rewards all-rounders and punishes specialists. It is also a direct answer to the familiar complaint that the classical world championship is slow, draw-heavy and hard to sell to broadcasters. For a player with a balanced classical-rapid-blitz profile, it is a natural home. But one detail matters more than the rest: the two crowns have separated. Carlsen holds the rating crown; the reigning world champion holds the title without rating supremacy. A dual-track configuration lasting this long has almost no precedent in the modern game. And when the highest-rated player publicly declines to contest the title, the title's claim to identify “the best” erodes from within — eroded by the strongest player's own words. At the governance layer, the article leaves a large question open: whether an event carrying the word “World” outside the FIDE cycle has FIDE sanction, and what its prize fund is. Without those two facts, any assessment of its stature is speculation. I will not assign it a tier before I see the published regulations. On industry transmission, the clearest point sits in the coaching layer. AI lowers the cost of high-quality analysis, eroding the advantage of wealthy federations and expensive coaching stables. The beneficiaries are large, well-organised, engine-native youth pipelines. The losers are traditional apprenticeship models. Which brings me to what the article leaves out. One: the thesis that “AI balances chess” sounds pleasant, but its mechanism points somewhere less pleasant. If preparation equalises and advantage becomes hard to create, draw rates rise and drama — the sport's core product — thins out. Carlsen himself concedes AI makes advantage harder to manufacture. The article quotes rather than analyses, leaving the darker half on the table. The three-format event, read this way, is a structural defence: rapid and blitz compress the incentive to draw. Two: “the average quality of my moves has improved” is the most attractive and least evidenced claim in the piece. Said seriously, it requires average centipawn loss or engine-match-rate data. The article offers no figure. I file it as a hypothesis: it likely reflects his own engine-assisted post-game review, and if the sample skews toward shorter games, the comparison is favourably biased. Three: the 20 million USD figure should be read as a negotiating anchor, not a reservation price. His phrasing is hedged at both ends — he prefers other aspects of chess, and he does not oppose the idea. A twentyfold gap from the current prize is a deliberately distant opening position, not a closing one. Reading it as a return signal is over-reading, and it is the single biggest headline risk in the story. Four: fifteen years at number one is a longevity record, not a peak-performance record. And part of that longevity is protected by scheduling: fewer classical games means fewer chances to lose rating points. A crown defended by absence is beautiful but thin — one unfavourable elite round-robin is enough to shake it. I am not saying he avoids anyone. I am saying scheduling is part of the story, and the article offers no scheduling data. Five, and perhaps the blind spot of collective memory: people read “humans are not that good” as a surrender. Placed beside the whole interview, it is a statement of epistemic humility — human understanding of chess remains far from the engines. The two readings produce two very different media products. One is tragedy. One is philosophy. You already know which one sells more clicks. Sixty days of watching recordings, and I imagined the roar of empty seats. I learned to write about chess without chess, and I learned this: when a great player talks about the limits of humans, he is usually talking about his own limits first. He no longer runs fast, but every step is an untold story. Age does not extinguish love for the board, it only changes its colour — from the colour of six-hour games to the colour of three-hour games, from the colour of a team of seconds to the colour of general ideas. And one small detail I kept. In that morning's report, beside the Budapest line, was the name Le Quang Liem. A Vietnamese player inside a field of 24, in an event the world number one is preparing to enter. Twenty years ago, while I was commentating at VTC, that was a hypothesis nobody dared put on paper. The game ends, but the applause still echoes in my memory. The question I carried out of that morning was not whether Carlsen will return to the title. It was this: when a sport has taught humans to be humble before machines, does it still have the courage to sell uncertainty — the only thing no machine can produce on our behalf?