The Download: how AI is shaking up Go, and a cybersecurity mystery

A decade after Google DeepMind's AlphaGo AI stunned the world by defeating South Korean Go champion Lee Sedol, the technology has completely rewired professional play. The program's victory marked a pivotal moment for artificial intelligence, proving machines could master a game of profound complexity and intuition. Since then, AI has systematically overturned centuries-old strategic principles in Go, introducing entirely new moves and sequences that human grandmasters had never conceived. Today, competing at the professional level essentially requires using AI for training and analysis, as players strive to replicate the machine's optimal—yet often inscrutable—decision-making. This has democratized high-level training, contributing to a notable rise in female players climbing the professional ranks.

The transformation is a double-edged sword. While AI provides an unparalleled training partner and has elevated the overall technical level of play, it has also sparked debate about the soul of the game. Some professionals argue that the pressure to mimic AI has drained Go of its human creativity and artistic expression, reducing play to the execution of pre-calculated, 'correct' lines. Others contend there is still room for human invention and style, especially in complex, novel positions. The story of AlphaGo and Go serves as a powerful, early case study of how advanced AI can reshape not just a competitive field, but the very epistemology of a domain, forcing experts to reconcile human intuition with machine-derived truth.