Chapter 10 - Data becomes the new Oil

Data was accumulating through the 1990s and early 2000s as a byproduct of ordinary human activity on the internet — not gold, which requires intentional extraction, but oil seeping from the ground, valuable only when the technology to refine it arrived. Google's PageRank demonstrated that learning from the statistical patterns in human behaviour at scale produced better results than encoding rules, establishing the proof of concept for a philosophy that the neural network researchers had been developing for a decade. Amazon's behavioural data and Facebook's social graph added behavioural signal and relationship structure to the accumulating corpus, making the internet the largest and most diverse dataset ever assembled — one that nobody had built for AI but that AI would eventually require. While this infrastructure was assembling itself invisibly, squash's structural challenges were compounding in the wrong direction, each missed digital opportunity making the eventual solution more complex, each court closure making the infrastructure base slightly smaller, each year of institutional conservatism raising the height of the wall between the sport and the generation it needed to recruit.