"Humans are not that good at chess": Carlsen's blunt AI verdict as he rewrites his preparation method

Five-time classical world champion Magnus Carlsen has said that artificial intelligence has made it clear that humans are not as excellent at chess as was long assumed. In an interview with the Spanish sports daily Marca (original source: AS), the 35-year-old Norwegian grandmaster said, "We have discovered that humans are not that good at chess, and there is still much to learn." The remark comes in his 15th consecutive year as world No. 1, and after he relinquished the right to defend the classical title in 2026. The central argument of the interview is that AI has made chess more interesting in the short term, but has also proven that creating an advantage at the top is becoming progressively harder. Carlsen's point is direct: when every rival has access to the same engine evaluation, the same database and the same analysis, the technical gap compresses and the contest narrows to understanding and mental resilience. The claim is consistent with rising draw rates in elite round-robins and the shrinking value of opening advantage. Carlsen frames it as balance; competitively, it is a compression of the skill gap. Technically, the most concrete part of the statement is his admission that he has changed his preparation method altogether. He no longer researches specific lines against a specific opponent and instead works on "more general ideas." In engine-era terms, this is a move from the memorisation-first school to the understanding-first school. The shift is practical as much as philosophical: a player who contests few classical matches through the year has little incentive to maintain a large memorised line arsenal. The second significant element is physical endurance. Carlsen conceded that his concentration capacity in five-to-six-hour classical games is no longer what it was, and that he now tries to win games a little faster. In chess, a small advantage typically converts into a real win only after move 35-40 — meaning endurance loss strikes directly at the conversion phase on which he built his reputation. This is not a sentimental admission but the most under-priced technical risk in his profile. He also claims that the average quality of his moves has improved. No independent metric accompanies that claim in the source: no average centipawn loss, no engine match rate, no game reference. Analysts argue the improvement may partly be an artefact of opponent selection, given that his classical calendar is thin and self-chosen. By way of comparison, elite classical play typically carries an average loss of roughly 15-25 centipawns per move, while engines lose only a small fraction of that. The "humans are not that good" line is therefore not false modesty but an accurate statement about the accuracy ceiling. The figure Carlsen did not supply can be attached independently through standard engine analysis. On psychology, Carlsen acknowledged that at 35, maintaining confidence and calm at the top is extremely difficult. The remark should be read as risk-reducing rather than risk-signalling: by his account there has been neither clear improvement nor clear decline — only that staying at the summit is intrinsically heavy. The most quoted line of the interview was the $20 million question. Carlsen neither accepted nor refused; the structure of the answer was plainly designed to preserve optionality while signalling the price of his return. Reading it as "Carlsen will return for $20 million" would be an over-read. Amid all this, there is one concrete item on the competitive calendar: the World All-Around Championship in Budapest from 10 to 21 November, where 24 players will compete on a combined classical, rapid and blitz score. The list includes former World Blitz champion Le Quang Liem, a specialist in the faster controls — a signal that the organisers are betting on all-format credibility rather than classical-only ranking. The event's status is double-edged. On one hand, the combined format is a genuine innovation: no existing title measures mastery of all three formats under a single crown. On the other, despite championship-grade branding, the source states nothing about FIDE sanction, prize fund or scoring rules. A 24-player invitation field means selection sits with the organiser, raising questions about both transparency and field strength. Twelve days, three formats and 24 players is a heavy workload for a 35-year-old who has himself stated that his endurance has declined. For Carlsen, the event's value is positional. Having renounced the classical title defence, the claim to being the best all-round player becomes the new basis of his supremacy — and rapid and blitz are precisely the formats where his relative edge is largest. Taken as a whole, the interview is not a game review but a methodological statement. AI has raised the quality of chess, but it has also ensured that the difference at the top will now be decided by thinking beyond the engine and by the physical capacity to endure long battles — and Carlsen is publicly defining his own position on both fronts.