Demis Hassabis: the little-known grandmaster of Google’s DeepMind AI

Demis Hassabis, boss of Google’s DeepMind AI operation, was caught on the hop by ChatGPT. He scrambled to catch up and now claims to have ushered in a new era.

Google Deepmind head Demis Hassabis speaks during a press conference ahead of the Google DeepMind Challenge Match in Seoul on March 8, 2016
(Image credit: JUNG YEON-JE/AFP via Getty Images)

Demis Hassabis has long been known as the brilliant British chess champion who became “the grandmaster of AI” (artificial intelligence) when Google bought his London-based DeepMind laboratory in 2014. 

Still, for much of the past year he’s been playing a nail-biting game of catch-up, says Wired. The astonishing popular response to OpenAI’s ChatGPT not only knocked Google off its perch as the world’s best-known AI innovator but, via its incorporation into Microsoft’s Bing search engine, posed a direct threat to Google’s future. 

“Stunned into action,” Google responded rapidly, hustling to launch its own chatbot, Bard, and promising a new general AI model to underpin it. 

Subscribe to MoneyWeek

Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE

Get 6 issues free

Sign up to Money Morning

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Sign up

Last week, it arrived. “The Gemini era is here,” tweeted Hassabis on the X social media platform. Or so he hopes.

The name of the system – a nod to the NASA project that paved the way for the Apollo moon landings – is significant. It also marks the twinning of Google’s two major AI laboratories (DeepMind in London and Brain in California) into a new unit under Hassabis’s leadership. 

When Google’s Larry Page persuaded him to sell DeepMind for £400m in 2014, a condition of the deal was that the start-up “would be shielded from pressure to make money,” says the Financial Times. All change. At 47, Hassabis has been dragged down from his ivory tower and pitched full square into the commercial battle of the decade. Those close to him claim he’s more than up to the challenge. “Demis is very pragmatic and highly, highly competitive,” says a former colleague – he “wants to be on the frontier of this.”

“Somewhere at his core,” Hassabis is a gamer, says Vox. Growing up in London, he was a child prodigy in chess from the age of four and, by eight, was writing computer games. He bought his first ZX Spectrum from his chess winnings and taught himself how to program from books. He completed his A-levels two years early before taking up a place at Queens’ College, Cambridge after a gap year spent designing video games for Bullfrog Productions. He earned enough to pay his way through university and graduated with a double first in computer science.

His interests are polymathic. Immediately after leaving Cambridge, he set up his own video games company, Elixir Studios, before returning to academia in 2005, aged 28, to study for a PhD in cognitive neuroscience at University College London. His focus was the hippocampus: the brain region crucial for navigation, memory recall, and imagining future events, says the FT

In many ways, the formation of DeepMind in 2011 reflected Hassabis’s two main strands of experience. Its propulsion into the higher echelons of Silicon Valley was down to a third, says The New York Times – chess. The story goes that Hassabis managed to land $2.5m in funding from his first major backer, Peter Thiel, by wangling an invitation to a party and seducing the legendary investor into a discussion of the unique strengths of the bishop and the knight. “I knew that he loved chess,” Hassabis later remarked. “I thought that would be my unique hook in.”

As arguably Britain’s greatest contemporary computer scientist, Hassabis, a member of the Royal Society, knows he is walking in the footsteps of Charles Babbage and Alan Turing, says the FT. The difference now is the speed of change. Back in 2015, when the aim of making “machines smart” was still viewed by some as a utopian idea, he sought to show how AI systems can learn and adapt by having them play retro arcade games such as Space Invaders or the ancient Chinese board game Go. But even he conceded that much of DeepMind’s research was still in the realms of “science-fiction stories.”

It may be time to switch game analogies. “A lot of chess players can’t handle poker,” Hassabis once observed. “In chess, if you play the right moves, you win. But life is more like poker… there are unforeseen and unknown things you can’t cater for.” Like ChatGPT, say. Fortunately for Google, Hassabis is an accomplished poker player. “Demis is very pragmatic and competitive – he wants to be on the AI frontier.”

This article was first published in MoneyWeek's magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.

Related articles

Jane writes profiles for MoneyWeek and is city editor of The Week. A former British Society of Magazine Editors editor of the year, she cut her teeth in journalism editing The Daily Telegraph’s Letters page and writing gossip for the London Evening Standard – while contributing to a kaleidoscopic range of business magazines including Personnel Today, Edge, Microscope, Computing, PC Business World, and Business & Finance.

She has edited corporate publications for accountants BDO, business psychologists YSC Consulting, and the law firm Stephenson Harwood – also enjoying a stint as a researcher for the due diligence department of a global risk advisory firm.

Her sole book to date, Stay or Go? (2016), rehearsed the arguments on both sides of the EU referendum.

She lives in north London, has a degree in modern history from Trinity College, Oxford, and is currently learning to play the drums.