Carlsen: 'AI shows humans are not that good'

In the last three matches I rewatched on tape, there was one detail that had nothing to do with any move: Magnus Carlsen placed his hand on his forehead, held that posture for four seconds, then lowered his hand and played a move that the computer evaluated as absolutely precise. Four seconds, in a game where each player has over a hundred minutes, is nothing. But if you place it next to his latest interview with AS, you will see it is a full stop for an entire era.

On October 21, Carlsen gave an interview and said that artificial intelligence makes chess more interesting, while simultaneously making it increasingly difficult to create an advantage over opponents. He also said a sentence that international media has quoted thousands of times: AI shows humans are not that good. And he admitted he no longer plays such long games as before, so he tries to win faster.

Those three sentences, read separately, are three news items. Read together, they are a self-narrative about the limits of humanity in a sport where machines surpassed us in 2026.

I have followed professional chess since 2026, when I was still competing and organizing tournaments. I watched Carlsen win the world championship in 2026, 2026, 2026, 2026, and 2026. I watched him defend his title against Karjakin, Caruana, and Nepomniachtchi. And I watched him announce he would not defend the title in 2026, when Ding Liren won the world championship without having to defeat the strongest player.

That is the necessary context to understand this interview. Carlsen is no longer world champion. But he is still the world number one by Elo rating, having held that position for 15 consecutive years. This is a record unprecedented in the modern era. And it is the hardest fact in the interview, while most of the remaining statements are philosophy without data.

The core of the story is not how strong AI is, but how much AI has leveled the playing field.

Before the computer era, the advantage of an elite player came from three sources: opening memory, calculation ability, and a team of seconds preparing before the match. Today, all three sources have been leveled by computers. A 2500 Elo player can open Stockfish on a phone and see exactly the move Carlsen took four seconds to find. A 2700 Elo player can hire three seconds and get an opening repertoire equivalent to a world champion's.

When everyone has the same tool, the marginal advantage of deep preparation falls. That is why Carlsen says he shifted from studying specific opening variations to more general ideas. This is not a decline in capability. It is a rational cost-benefit adjustment in a saturated preparation market.

But there is a deeper layer the interview does not state. If opening preparation is leveled, where does the decisive factor migrate? The answer is: over-the-board calculation, time management, and psychological resilience. These three factors cannot be copied from a computer. And that is precisely where Carlsen is having problems.

He admits his physical fitness has declined. He says he no longer plays those 5-6 hour games as before. This is a very credible signal, because it comes from the athlete himself rather than from outside inference. In classical chess, peak performance typically falls in the late twenties to early thirties. Carlsen is 35 this year. The age curve is going exactly the way it always goes for every player.

And this is where I want to pause a little longer.

There is a fact rarely mentioned in analyses of Carlsen: he has not participated in the world championship cycle since 2026. He says he is interested in many other aspects of chess more than in a championship match. He also says he does not oppose returning, if the prize is around 20 million dollars.

That 20 million dollar figure has become the headline of hundreds of articles. But if you read carefully, you will see it placed next to another figure: the classical world championship currently pays the winner just over 1 million dollars. The gap between the two figures is 20 times.

This is not an offer. This is a benchmark. Carlsen is telling organizers: if you want me back, pay me what the strongest person in the world deserves. And he knows that at the current level, that will not happen.

In sports, when the strongest person refuses to contest the championship, that title loses part of its legitimacy. This is what happened to heavyweight boxing in the 1990s, and to women's tennis in the 2000s. Chess is going exactly down that path.

But there is one difference. In boxing and tennis, people argue about who is stronger than whom. In chess, we have an objective measure: Elo. And by that measure, Carlsen is still number one. This is a strange configuration with no precedent at this scale: the championship crown and the number one crown have separated.

And that is why this interview matters. Not because it talks about AI. But because it is the first time the strongest person in the world publicly says the current system no longer deserves him.

I rewatched the game Carlsen won against Nepomniachtchi in game 6 of 2026. It was a 136-move game lasting nearly 8 hours. Carlsen won by patiently converting a small advantage into a winning endgame. That is the kind of win he is famous for: not flashy, not breakthrough, just endurance and precision in every move.

Now he says he no longer plays such long games. He says he tries to win faster. This is a tactical change with a clear mechanism: if the endurance advantage is gone, shift to formats where endurance is not the decisive factor. That is rapid and blitz.

But this is where I want to pose a question the interview does not answer. If Carlsen shifts to fast play, he is trading certainty for luck. In rapid chess, variance is higher. A 2700 Elo player can beat Carlsen in a 3-minute blitz game. That is nearly impossible in a 6-hour classical game.

So is the win-faster strategy an adaptation, or a surrender dressed up in tactical language?

I think the answer lies in another detail of the interview. Carlsen says the average quality of his moves has improved. This is the most technically interesting and least evidenced claim in the piece. He provides no ACPL figure, no engine match rate, no metric whatsoever. But if true, it means he is playing more precisely, despite weaker fitness.

This is explainable. In the computer era, post-game analysis tools raise the floor of decision quality. You do not need superhuman memory to play correctly if you have seen thousands of similar games on a screen. But the ceiling of quality is still limited by endurance, especially in long games and complex endgames.

So Carlsen's claim may be true in a sample of short games, but not in long games. This is a hypothesis, not a finding.

And this is where I want to talk about an event the interview mentions only once.

From November 10 to 21, Budapest will host a tournament called the World All-Around Championship, gathering the 24 strongest players in the world, including Le Quang Liem of Vietnam. The tournament combines three formats: classical, rapid, and blitz. The winner will be called the world all-around champion.

This is not a tournament within FIDE's world championship cycle. It is an independent commercial product, designed to compete with the classical world championship.

And Carlsen is participating.

This is a very clear signal about where he thinks the competitive frontier of chess is shifting. He is no longer interested in contesting the classical championship. He is interested in becoming the all-around champion, the best at all three formats.

There is a rarely mentioned fact: Carlsen is one of the few players with a balanced profile across classical, rapid, and blitz. He once held the number one spot in all three formats simultaneously. This makes him a natural candidate for an all-around tournament. But it also raises a question: can an all-around tournament replace a classical championship in determining who is the strongest?

I do not think so. Classical chess is the only format where endurance and strategic depth are fully tested. An all-around tournament is a better entertainment product, but it is not a more accurate measure of chess strength.

However, this is a fact FIDE must face: if a commercial tournament can attract the 24 strongest players in the world, then the official world championship is losing its monopoly position.

And this is where I want to talk about something the interview does not state.

For 15 years, Carlsen has held the world number one position. This is a great record. But it is also a record protected by selective tournament choice. He does not play as many classical events as other players. He chooses events that favor him. This is not wrong strategically, but it means his number one position is partly maintained by absence.

If he played more classical events, his rating might have dropped. This is a hypothesis, not an assertion. But it is a hypothesis worth considering when evaluating his 15-year record.

And this is where I want to talk about something I believe is most important in the interview.

Carlsen says AI shows humans are not that good. This is a philosophical statement, not a technical finding. The article provides no data on the gap between humans and computers. No ACPL, no match rate, no metric.

But that statement is still true on another level. When computers can beat every human player, we are forced to admit that chess is not a sport of perfect intellect. It is a sport of human intellect, with all its limits.

And that is what makes chess interesting. Not because humans play perfectly, but because humans play imperfectly.

When the stadium is empty, I hear the ball speak for thousands of people for the first time.

In chess, the quietest moment is when a player sits before the board, hand on a pawn, and does not know which move to play. That is the moment a computer never experiences. And that is the moment we, the viewers, are waiting for.

Modric's tears do not belong to Croatia; they belong to those who have lost in order to understand.

Carlsen's statement does not belong to chess; it belongs to everyone who has ever realized they are not the best at anything.

Sport does not promise victory; it only promises a heartbeat brave enough to continue.

And Carlsen, at 35, with declining fitness, with a tournament system he no longer trusts, is still continuing. He shifts to fast play. He joins new tournaments. He speaks of AI with curiosity rather than fear.

That is not a surrender. That is an adaptation.

And in a sport where machines have surpassed humans, adaptation is the only way to survive.

The question I want to leave: if the strongest person in the world says humans are not that good, what are we watching chess for? A perfect result, or a human moment?