The Convergence

Squash and artificial intelligence share an unlikely origin story, both born in England in the 1830s, both slow to reveal their full potential, and both arriving at 2022 at opposite ends of their respective arcs. The introduction traces squash's remarkable two-century journey from a school corridor in Harrow to a genuinely global sport, one that produces world champions on every continent and has finally earned its place at the 2028 Olympics, while quietly losing courts, participants, and cultural relevance in the decades it spent ignoring every major technological wave. In parallel, it traces AI's seventy-year path from a mathematician's dream to a technology that one hundred million people adopted in two months. The argument is simple: for the first time in history, the tools exist to do what no governing body or marketing campaign ever could for squash, and this book is the story of how both long games finally converge.

Two ideas were born in the same country, in the same decade, within twenty miles of each other. In the 1830s, Harrow School students improvising in a basement corridor accidentally invented squash — a game whose shared wall made intelligence, not power, the minimum requirement for survival. A few miles south, Charles Babbage was designing the Analytical Engine, a general-purpose computing machine that imagined the processor, the memory, and the program a century before any of them existed. Ada Lovelace, his collaborator, asked the question that would haunt AI research for the next hundred and fifty years: can a machine originate anything, or can it only do what it is told? Both ideas were genuinely significant, both were ahead of their time, and neither had yet found the world it needed.

When squash crossed the Atlantic in 1884, it did not democratize — it deepened its class associations, arriving at St. Paul's School in New Hampshire and spreading through Ivy League clubs, university athletic programs, and the social networks of American privilege. America did not open squash to everyone; it gave squash a more thoroughly elite home than it had ever had in England. Simultaneously, Herman Hollerith crossed a different kind of Atlantic, applying Babbage's punch card concept to the 1890 United States Census and demonstrating that mechanical information processing was practically superior to human labor — a commercial motivation that pure mathematics had never provided. The company Hollerith founded would eventually become IBM. Both ideas had crossed the ocean. Both had found institutional homes that would shape their trajectories for decades.

The summer of 1956 at Dartmouth College gave artificial intelligence its name and its founding ambition: that every aspect of human intelligence could be precisely described and mechanically reproduced. The optimism was genuine and, within its own frame of reference, earned — the Logic Theorist was proving mathematical theorems, and ELIZA was conducting conversations that made intelligent adults feel genuinely heard. Simultaneously, Kenneth Cooper's 1968 aerobics research found squash earning the highest cardiovascular scores of any sport, and the professional class flooded into courts across North America with champions like Barrington and Hunt giving the sport its human face. Both communities were performing brilliantly within bounded domains and beginning, without quite knowing it, to encounter the costs of those boundaries.

Both squash and artificial intelligence built their institutional infrastructures in the same era and embedded within them the same hidden flaw: institutions designed by and for the people who had created the field, optimizing for the experience those people understood and leaving everything else as an afterthought. Squash wrote its rule book through the International Squash Rackets Federation, eventually resolving the hardball-softball divide but creating a governing class whose instinct was conservation rather than adaptation. AI wrote its epistemological rule book through the physical symbol system hypothesis — the declaration that intelligence was fundamentally symbol manipulation — which was productive within its domain and quietly inadequate outside it. Jahangir Khan arrived and built his 555-match winning streak, demonstrating what the game could ask of a human being; Jansher Khan extended the Pakistani dynasty through eight World Open titles; and the sport's institutions managed the tension between who ran the sport and who was best at it without resolving it.

The Kodak story is the frame for this chapter: Steven Sasson invented digital photography inside Kodak in 1975 and watched his invention shelved while Kodak optimized for film, eventually filing for bankruptcy in 2012. Both squash and AI in the 1980s were Kodak — well-managed, optimized for what they had built, carrying the particular blindness that comes from genuine success. Squash's courts were full, its professional tour was gaining visibility through glass-backed courts, and Jansher Khan was demonstrating tactical intelligence that matched Jahangir's physical dominance. Then Jonathan Power became the first North American world number one in 1999, arriving from outside the institutional pipelines and revealing, through the sport's inability to process or capitalize on his cultural distinctiveness, that squash's institutions were optimized for the players they had always produced and had no mechanism for converting their most interesting moment into new participants. AI's second winter collapsed the expert systems boom for the same structural reason — the technology worked within its domain and failed to close the last mile outside it.

Decline rarely announces itself, and Squash's decline was precisely the kind that arrives dressed as stability — membership numbers slightly down but surely recovering, courts quieter at shoulder hours but still full on Saturday mornings. The economic conditions of the late 1980s, rising interest rates and softening real estate values, made the squash court a difficult asset to justify, and the one-way ratchet of court closures began to turn — once closed, courts were rarely replaced. The sport's governing bodies responded by presenting participation data in the most favourable possible light, relying on historical numbers rather than current surveys, breaking the feedback loop that would have allowed the leadership to see the problem clearly. Egypt went in the opposite direction, building a squash culture of extraordinary depth that would eventually produce multiple generations of world champions, proving that the sport's decline in Western markets was not inevitable but the product of specific institutional failures.

The AI winters were not interruptions in the story of artificial intelligence — they were the periods during which the most important work in the field's history was being done by people the mainstream had stopped paying attention to. Herbert Simon's 1957 prediction that computers would be world chess champions within a decade was the epitome of the field's founding optimism; the Lighthill Report of 1973 was the first honest reckoning with the gap between promise and reality, and British funding was substantially cut. Geoffrey Hinton, working with Rumelhart and Williams, published the backpropagation paper in 1986, lifting the theoretical ceiling on neural networks without having the data or hardware to demonstrate what lifting that ceiling meant. Yann LeCun was teaching neural networks to read handwritten digits at Bell Laboratories with accuracy that exceeded rule-based approaches on a real commercial task, and NVIDIA was founded in 1993 to make graphics cards for video games, inadvertently building the hardware infrastructure that would eventually make large-scale neural network training possible.

Samuel Morse's 1844 telegraph could cross forty miles in seconds and still required a messenger boy to walk the final blocks — this is the last mile problem, the persistent gap between the speed of the signal and the completion of the action the signal was meant to enable. Every communication technology in history has reproduced this gap in a new form: the telephone could connect voices but could not execute decisions; email could deliver messages instantly but could not implement the actions those messages described; social media could distribute information globally but could not convert that information into completed outcomes. Squash's version of the messenger boy is the player who must navigate booking systems, WhatsApp groups, league registration portals, and scheduling negotiations to convert the desire to play into an actual game. This friction costs the sport precisely the eighty percent of recreational players who have not learned to walk the last mile efficiently. What genuinely closing the last mile would require is a system that understands intent, maintains context, acts across multiple systems simultaneously, and completes outcomes as a direct consequence of a single natural expression of desire — a specification that no technology available to squash in the 1990s, 2000s, or early 2010s came close to meeting.

The internet was not a communication technology — it was a network effects technology, and Squash treated it as the former while missing the latter entirely. The sport put its booking sheets online, built digital brochures it called club websites, and used the most powerful connective infrastructure in human history to broadcast information to passive audiences rather than build relationships between active participants. The software fragmentation that resulted — more than 250 separate booking, club management, and event management systems serving the racquet sports market by 2024 — was the structural consequence of a market responding to fragmented demand with fragmented supply, each system creating an isolated silo that prevented the network effects the sport desperately needed. Ross Gerring, working independently without institutional support, built the most complete global database of squash venues ever assembled through the Squash Players App — later integrated into the World Squash Federation and European Squash Federation's Find a Club features — demonstrating that the sport's most important digital achievements in this era came from individuals who loved the game, alongside voices like Gerry Gibson's In Squash podcast and Conor O'Malley and Bill Buckingham's Squash Radio, who built the content and community infrastructure the institutions were not building.

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.

The iPhone arrived in 2007 as the first communication technology that was simultaneously personal, persistent, social, and always connected — eliminating the conscious act of engagement that every previous technology had required and making the network permanently present. Facebook, YouTube, Twitter, and Instagram collectively created the most powerful community-building infrastructure in human history, freely available to any sport willing to use it as community infrastructure rather than a bulletin board. Squash used it as a bulletin board — posting tournament brackets and event announcements while padel and pickleball built communities whose authentic enthusiasm spread organically through the social platforms and drove explosive participation growth. The cultural rebellion that followed — the growing, widespread recognition that general social media had delivered noise rather than connection — created precisely the appetite that a purpose-built, signal-rich squash community platform could satisfy, representing the first genuine cultural tailwind the sport had experienced since the fitness boom of the 1970s.

AlexNet's 15.3 percent error rate at the 2012 ImageNet competition, against the second-place team's 26.2 percent, was not an incremental improvement — it was the demonstration that the neural network approach was so fundamentally superior to the symbolic paradigm that the field's center of gravity shifted permanently in a single afternoon. The GPU, developed by NVIDIA for video games since 1993, turned out to be the perfect hardware for training neural networks — a massive parallel processing architecture suited to both rendering 3D graphics and performing the matrix multiplications that deep learning requires — and it had been advancing continuously for two decades before anyone in AI realized it was the infrastructure they needed. The Transformer architecture, published in 2017, solved the sequence problem that had kept neural networks from handling language reliably, processing entire sequences simultaneously through an attention mechanism that allowed any word to directly consider its relationship to every other word, regardless of distance. RLHF — Reinforcement Learning from Human Feedback — was the final ingredient, aligning model outputs with human preferences through systematic human judgment and producing systems that were not merely capable but genuinely useful, setting the stage for the moment five years later when the world woke up to artificial intelligence.

GPT-3 arrives with 175 billion parameters and changes everything — not because of what it can do, but because of what it reveals. Capability is no longer a research problem. It is an engineering problem. The scaling laws make the trajectory of AI predictable for the first time in the field's history.

Between 2020 and 2022, AI speciated. DALL-E, Codex, AlphaFold 2 — the Transformer template replicates across every domain that can be represented as data. The AI-native company category forms. By the end of 2022, every component of a unifying squash platform would exist. The ignition is eleven weeks away.

On November 30, 2022, OpenAI released ChatGPT. One million users in five days. One hundred million in two months. No consumer technology in history moves this fast. Underneath the hype is a genuine discontinuity — and for squash, the moment the last mile finally closes.

The gap between capability and utility — the last mile — is the hidden thread connecting squash's fragmented infrastructure and AI's long journey toward agency. Instruction following, tool use, and a decade of quietly built API infrastructure converge simultaneously. The messenger boy goes home.

Rob Eberhard picked up a squash racquet at twelve years old, handed to him by his father. He never put it down — building clubs, teaching players, selling racquets out of the back of his car between tournament rounds, and eventually spending eleven years and five pivots trying to build the infrastructure the sport was missing. His daughter Sage left competitive squash not because she stopped loving it but because the social thread connecting her to its community was never strong enough to hold her. The corporate league experiment in Whistler tripled participation in two months and proved the thesis: this was never a passion problem. The rideshare shifts that kept the lights on became a pitch laboratory. And when Sage came back to squash at Queen's University — earning the Jack Fairs Trophy from the coaches of every opposing team and the MVP award from her own — the book's central question answered itself in the most human way possible. The infrastructure that holds people in the sport is not an abstraction. It is the difference between a daughter who leaves and a daughter who comes back.

How did you first start playing squash? Someone invited you. Network effects explained through Metcalfe's Law, Facebook's social graph, and the three-thousand-telephone-companies catastrophe of 1904. The difference between a tool and a network: tools solve problems, networks compound. What would a universal connection mean for a sport whose players have never truly been connected to each other?

K greater than one. The participation in mathematics that explains why squash has been declining despite genuine effort. The initiative trap — programs that produce participants but not advocates. The specific digital architecture that converts participants into advocates and replaces individual effort, which resets, with infrastructure, which compounds. The public facility is the sport's most important and most neglected opportunity.

Bell's patents expired in 1894, and six thousand telephone companies fragment a network that only works when it is one network. Visa, Stripe, and AWS followed the pattern that follows: fragmentation, neutral connector, interoperability, compounding value. The most defensible position in the sports technology stack is not the most powerful one. It is the most trusted one. Trust cannot be purchased or replicated at any speed.

The mission and the commercial model are not aligned. They are the same thing. Everything on the platform needs to lead to the growth of the sport, or it does not get added. A platform that grows only when the ecosystem grows cannot extract value from a declining community — it can only create value by growing one. The structural constraint that makes the right behaviour the commercially necessary behaviour. The network is forming. The rest of the story is being written now

Squash's challenge has never been passion — it has been infrastructure. ClubHub is building the compounding digital layer the sport has always needed, and three forces have converged to make this moment uniquely actionable: AI is production-ready, culture is hungry for purposeful community over algorithmic noise, and the 2028 Los Angeles Olympics will deliver the largest audience squash has ever had. The pipeline to receive that audience must be built before it arrives. This is an invitation to governing bodies not to evaluate a product, but to help build one — and the associations that engage now will shape what the sport's digital future becomes.