Musk: Chess Will Eventually Be Fully Solved by Computers
Elon Musk stated that chess is difficult for humans but not for computers, and one day it will be completely solved like checkers; until then, and even after, people can still enjoy many games and sports that machines are far better at.
Checkers was mathematically proven to be a draw with perfect play by both sides in 2007 by the Chinook project team at the University of Alberta. The state space of chess is much larger than that of checkers and has not yet been exhaustively solved, but computers have already consistently surpassed human world champions in public competitions, with systems like AlphaZero replacing opening manuals with self-play reinforcement learning. Musk separates "unsolved" from "more fun": solving does not equate to the disappearance of entertainment.
Humans continue to hold world championships, street blitz games, and online rankings, with broadcasts relying on mistakes and style rather than proof trees. Similar phenomena have appeared in Go, esports, and some sports data analysis: machines serve as sparring partners and referee assistants, while tickets are still sold to those who can lose. After complete resolution, endgame tables will become public infrastructure, the space for opening innovations will narrow, but entertainment rules can still be modified by humans.
The buyers are players and audiences who want to retain the fun, while the sellers are engineers who write the computational advantage into predetermined outcomes. The event is driven by his public statements. The beneficiaries are platforms that treat the human-machine gap as a sparring product, while the pressured parties are those who narrate human champions as the ultimate measure of intelligence.
Source: Public Information
ABAB AI Insight
Musk uses the established fact of checkers to write an epitaph for chess, while immediately adding "keep playing." This reflects the engineering determinism of Tesla and SpaceX: systems that can be fully solved will eventually be solved, and entertainment is a byproduct before resolution. Solving checkers took eighteen years of computing power and proof, and while the game tree of chess is wider, the direction is written as the same path. By bringing in sports, he prevents this statement from being read as "humans are inferior."
The capital path is strong models as sparring partners, with content as the ticket. Chess platforms sell matching, broadcasting, and subscriptions, not perfect endgame tables. After complete resolution, the product will shift from "who wins" to "who loses to the machine in what style." Musk's businesses have framed autonomous driving and rocket recovery as problems that can be converged by engineering, and chess is a low-cost metaphor of the same logic.
The analogy is that human competitions still exist after Kasparov lost to Deep Blue, and there are still entries on the sprint record wall. The industry is in a coexistence phase where machines steadily outperform, and rules remain unchanged.
Structurally, this belongs to technological substitution. The mechanism is: the optimal strategy for finite perfect information games can be searched and proven to be approached; human enjoyment does not come from optimality, but from the opponent making mistakes. Solving changes the boundaries of knowledge but does not automatically cancel the ticket.
ABAB News · Cognitive Laws
- Being beaten by a computer and being fully solved are not the same day.
- After machines become stronger, tickets are still sold to those who can lose.
- Games that can be fully solved will eventually be solved, and fun still relies on rules among humans.