Particle.news
Download on the App Store

Musk and Chess.com Clash Over Whether AI Can Solve Chess

The online spat highlights a definitional gap on counts of games versus legal positions that shapes claims about whether future AI could ever fully map perfect play.

Overview

  • Elon Musk wrote that "chess will be fully solved one day" after tweeting that only a small fraction of chess moves are sensible, and Chess.com replied with a short retort of "skill issue."
  • Chess.com tweeted that the number of possible full-game sequences is about 40 orders of magnitude larger than the atoms in the observable universe, citing the Shannon-style estimate for possible games.
  • Musk countered that Chess.com confused full-game sequences with legal board positions and said legal positions are roughly 40 orders of magnitude fewer, adding that a future artificial superintelligence could compress or solve the game.
  • Reporters and analysts noted both sides are correct in their contexts—Chess.com meant possible game sequences (roughly 10^120) while Musk referred to estimated legal positions (commonly cited near 10^44)—and no technical breakthrough toward solving chess has been announced.
  • The exchange drew wide attention from chess figures and commentators and underscored how social-media shorthand and numbers can obscure the real technical question: powerful engines beat humans but chess’s game-theoretic outcome with perfect play remains unknown.