Book review: The Infinity Machine
Though the AI race may be mid-flight, authors are already scrambling to document it. Biographers and historians usually need the distance of a few decades to piece together a narrative, but the 21st century tech boom has taught writers that when there’s a story to be told in the technology industry, you pounce on it.
This is not to say that The Infinity Machine: Demis Hassabis, DeepMind and the Quest for Superintelligence is rushed. On the contrary, the author Sebastian Mallaby began having conversations with his subject in late 2022, just at the point when generative AI was about to become the default subject of every conference, Substack essay and podcast under the sun. He had an idea that Demis Hassabis, the founder of Google-owned, London-based AI lab DeepMind, might be an important person to profile. He was right.
The book is not quite a pure biography, becoming more of a business narrative. Sometimes the focus is tightly on Hassabis, sometimes on the company, and several chapters are even dedicated to things that were happening outside of DeepMind. As other writers have discovered, it is almost impossible to tell the story of one frontier lab without also covering what was happening at the others. Parmy Olson’s 2024 book Supremacy tried to solve this by telling the stories of Hassabis and Altman in tandem, through alternating chapters. Mallaby, by necessity, gets a little sidetracked into the breakthroughs and court intrigue of OpenAI before bringing the narrative back to DeepMind.
The big success of the book though is how Mallaby guides the reader through corporate intrigue, technical details and even theoretical scientific ideas with an encouraging hand. Unsurprisingly for an author who has previously tackled subjects as vast and complex as venture capital and the global banking system, he manages to keep the explanations readable and the narrative pacey.
At the heart of all of it is a portrait of DeepMind’s founder. As the leader of one of the most important AI companies, Hassabis has kept a relatively low profile compared to peers like Elon Musk and Sam Altman. Part of that has derived from his decision to keep the business based in London, rather than joining the Silicon Valley soap opera.
Could the DeepMind story have happened anywhere other than London? Certainly, Mallaby highlights several moments where things could have gone another way. Hassabis and his colleagues are constantly flying stateside – first to study, then to secure funding from the likes of Peter Thiel, and later to negotiate terms with Google in Mountain View. If the company could only ever have been born in and remained in London, this is less of a testament to fate and more to Hassabis’s conviction. One might even say stubbornness. “I’m a weird British outlier, on this little island here, and I’ve made my own path,” he tells Mallaby.
The author evokes the spirit of London beautifully, as he describes meeting Hassabis for interviews on park benches and the upstairs rooms of pubs. He also pays close attention to the upbringings of both Hassabis and his erstwhile co-founder, Mustafa Suleyman. The two were both products of a multicultural London where one’s talents could throw you in the path of new opportunities.
This is a living biography, one about a subject whose story has not yet ended – who may indeed only be at the precipice of what he will be remembered for. The benefit of this is that Mallaby has repeat access to Hassabis. The conversations they have are the most illuminating parts of the book. Despite that access – or perhaps because of it and the conditions attached – there are some aspects of Hassabis that remain a mystery. His family life is closed off, save for a couple of mentions of playing video games with his kids. We never hear the story of how he and his wife fell in love at Cambridge. Perhaps, for someone whose identity is so synonymous with his company, it’s important that he keeps a few things separate.
Hassabis had a natural entrepreneurialism from early on, shaped by his time working under the visionary games designer Peter Molyneux. It was this training, Mallaby suggests, that helped fire that company-building instinct in a Britain that was a relative “start-up desert”. This combined with his innate talents, such as an uncanny ability to persuade, which his longtime friend and colleague David Silver dubs a “Jedi mind trick”.
While he has a brilliant technical mind himself, this power of persuasion allowed him to get the top researchers and programmers his companies needed to make breakthroughs. Then he pushes them. In one telling anecdote, he drops in on the team working on the early stages of AlphaFold, the programme for predicting protein structures that would go on to win Hassabis and his colleague John Jumper the Nobel Prize. Displeased with the team’s plan to continue their current strategy, he pressed them to aim higher. “He didn’t want to be the best in the field,” Mallaby writes. “He wanted to solve the problem.”
Mallaby does probe though, focusing on getting a better idea of how Hassabis sees himself and his role in the AI race. He finds a mixed picture of somebody who fantasises about escaping public life and work on his own ideas, but at the same time thrives in the chaos and speed of what’s happening now. “He was caught up in a terrifying capitalist contest. He relished it.”
Matt Clifford, Entrepreneurs First founder and former government adviser, has praised the book precisely because it captures the sense of Hassabis as not just a brilliant mind but a “relentless entrepreneur”. This duality in fact parallels the challenge that Britain’s AI scene now faces. Already known as a home of excellent researchers and intellectual capital, the country will also need people with ambition and a competitive spirit. As Hassabis demonstrates, they sometimes arrive in the form of the same person.
