It was March 2016, and at the Four Seasons Hotel in Seoul, the world was gathered to watch the culmination of a battle 2,500 years in the making. On one side was the South Korean Lee Se-dol, the second-highest ranking Go player in the world. On the other was AlphaGo \u2013 a computer program developed by London-based artificial intelligence research company DeepMind.<\/p>\n
\u201cChess is the greatest game mankind has invented,\u201d game designer Alex Randolph once said. \u201cGo is the greatest game mankind has discovered.\u201d Something about the ancient Chinese duel, where players place stones on a grid, trying to capture territory, feels fundamental \u2013 inevitable, even. Chess had fallen to the robots nearly 20 years earlier, when DeepBlue beat Kasparov, but Go, with its vast decision space (there are far more legal board positions than atoms in the observable universe) remained a plucky holdout.<\/p>\n
Over five matches, watched by more than 200 million people globally, DeepMind won four games to one. After his third loss, Lee Se-dol apologised: \u201cI, Lee Se-dol, lost, but mankind did not.\u201d The real winner was DeepMind and its CEO, Demis Hassabis.<\/p>\n
The Infinity Machine is a detailed account of Hassabis\u2019s journey from chess prodigy to the driving force behind a global leader in AI technology. Parts of this journey are rather remarkable. His Chinese Singaporean mother, we are told, \u201cgrew up in absolute poverty\u201d, spending part of her childhood as an orphan on the streets of Singapore, while his Greek Cypriot father had dreams of music stardom and \u201csold toys out of the back of a beaten-up red Volkswagen van\u201d.<\/p>\n
Hassabis was unusually bright from an early age. He started playing chess \u2013 and beating adults \u2013 at four. By five, he was competing in tournaments, sitting on a phone book on top of two stacked chairs so he could see the table. By nine, he was captain of the England under-11 team. By 13, he had achieved the rank of chess master and was the second-strongest player in his age group globally.<\/p>\n
It was an intensely pressured environment \u2013 wooden boards were placed under the tables \u201cto prevent players from kicking each other\u201d. If Hassabis lost, his dad frequently \u201cwent mental\u201d, screaming at him, and when he counselled him to \u201cdo his best\u201d, Hassabis took it literally; the only way he\u2019d know he\u2019d succeeded was if he exerted himself to near-collapse \u2013 \u201cbasically if I pushed myself to the point just before death\u201d.<\/p>\n
After a stint at Bullfrog games, working on the wildly successful Theme Park under designer Peter Molyneux, Hassabis studied at Cambridge, founded his own game studio, then returned to academia to complete a PhD in neuroscience. In 2010, along with his friend Mustafa Suleyman and Shane Legg, whom he met as a postdoc, he founded DeepMind.<\/p>\n
DeepMind\u2019s precise mission shifted depending on who they were pitching to. Soon, they got an audience with Peter Thiel, cofounder of PayPal and the governmental and military surveillance company Palantir, who hooked them up with investment. Thiel went on to become a Maga eminence grise who oversaw JD Vance\u2019s ascent to power and now tours the podcast circuit explaining that big tech is the only way to combat the rise of the antichrist.<\/p>\n
Thiel\u2019s views look positively pedestrian in the weird, feverish bubble of tech finance, especially among those\u00a0lured by the carrot Hassabis was dangling \u2013 AGI, or Artificial General Intelligence, an ability to meet or beat human performance across every major cognitive domain. \u201cIt\u2019s really finding God\u2019s algorithm,\u201d said one\u00a0investor.<\/p>\n
This world of hype and speculation is both fascinating and horrifying, but unfortunately, Mallaby appears unwilling or unable to engage critically with his interviewees\u2019 claims. \u201cArtificial intelligence,\u201d he writes, \u201cheralds a transformation more profound than anything since Homo sapiens acquired the capacity for abstract thought.\u201d No it doesn\u2019t, Sebastian. More profound than agriculture? More profound than language?<\/p>\n
Certainly, people such as OpenAI CEO Sam Altman \u2013 who appears largely in this book as a cynical and cavalier wrecker, stealing DeepMind\u2019s thunder by releasing ChatGPT before anyone else was ready \u2013 would like us to believe we\u2019re on the verge of something unimaginably consequential. How else can they source investment to build more loss-making, drought-causing, energy-hungry datacentres? The Singularity is just over the next hill.Mallaby\u2019s prose is tortuous and intrusive throughout. No one is permitted to have merely \u201csaid\u201d something \u2013 rather, they \u201cconfessed\u201d, \u201cadded\u201d, \u201cdeclared\u201d, \u201cconcluded\u201d, \u201creflected\u201d, \u201cfretted\u201d, \u201cvowed earnestly\u201d. This cavalcade of purple dialogue tags may reflect an unconscious attempt to jazz up various interviewees\u2019 contributions, most of which, including those from the (perfectly nice-seeming) Hassabis, are rather dull.<\/p>\n
\u201cDoing science is, sort of, like reading the mind of God,\u201d Hassabis tells a credulous Mallaby. \u201cUnderstanding the deep mystery of the universe is my religion, kind of.\u201d Sort of. Kind of. These hedging phrases all but give the game away, the sly hitching of the actual to the notionally adjacent but thematically sexy. It\u2019s instructive that Hassabis began his life in computing as the protege of Molyneux, a man notorious for making big promises about the depth and innovation of upcoming titles such as Black & White (which would go on to earn the dubious accolade of \u201cmost overrated game of all time\u201d).<\/p>\n
Perhaps anticipating that \u2013 given the reputation of men such as Thiel and Elon Musk \u2013 the public\u2019s thirst for tech CEO hagiography is on the wane, Mallaby reassures readers that \u201cFor every Donald Trump supporter in the tech industry \u2026 there is also a Bill Gates\u201d \u2013 a comparison that, since the release of Gates\u2019s email exchanges with Jeffrey Epstein, reveals rather more than Mallaby intended.<\/p>\n
Hassabis comes across as far less objectionable than these people, but remains an underwhelming protagonist. Mallaby\u2019s focus on interminable dinners and wrangles over management structure, while faithful to the reality of running a tech startup, chokes out Hassabis\u2019s impressive achievements, such as receiving the Nobel prize for chemistry with John Jumper for their work adapting DeepMind to predict protein structures. This is precisely the kind of research \u2013 miles from tech-bro hype about building God or founding quasi-libertarian AI-run fiefdoms \u2013 that could actually change lives.<\/p>\n
Sadly, Mallaby mistakes Hassabis\u2019s intelligence in one field \u2013 computing \u2013 for general brilliance across all domains, treating his half-formed pub takes on the nature of reality and aspirations to build a Large Hadron Collider as if they were revelatory dispatches. \u201cI am really a practical philosopher,\u201d Hassabis tells him. \u201cI\u2019m not just sitting there thinking \u2026 I\u2019m also doing experiments. Isn\u2019t that wonderful?\u201d<\/p>\n
Sort of. Kind of.<\/p>\n