Demis Hassabis
Architect of Intelligent Machines
Demis Hassabis is the co-founder and CEO of DeepMind, a groundbreaking AI research organization acquired by Google. With a background in neuroscience and game development, he has led remarkable advances in machine learning, most notably the creation of AlphaGo—an AI that defeated a world champion in the ancient game of Go. Driven by a vision to push the frontiers of technology and benefit humanity, Hassabis’s work continues to redefine the boundaries of artificial intelligence and inspire cutting-edge scientific discovery.
- 15 min listen
- Nobel Laureate
- DeepMind
- Artificial intelligence
Demis Hassabis’s full 15 min audio biography, Architect of Intelligent Machines, is available on Lyfestori for iOS. Listen to a free preview below.
A life in brief
The shape of a life.
He was beating his uncles at chess before he turned five, and grew into the second-ranked player in the world for his age group — behind only Judit Polgar. He spent his gap year winning a job at a video game studio and shipping a cult classic. He is a five-time Pentamind champion and has taken six figures off poker tables. Demis Hassabis treats all of it as one question: how does a mind decide? The answer he built, DeepMind's AlphaFold, won him the 2024 Nobel Prize in Chemistry.
- 1976
Born in London
Born 27 July into a supportive family that fed his curiosity. Before he is five he is routinely beating his uncles at chess.
- 1994
Theme Park
Having graduated a year early, he wins a place at Bullfrog Productions through their 'Win a job at Bullfrog' contest and works as project leader and lead programmer on Theme Park — a management sim that sells millions and helps establish the sandbox genre.
- 1998
Elixir Studios
Founds his own games company after Cambridge. Its first release, Republic: The Revolution, uses AI to model the consequences of a player's political decisions across a fictional country. Critics praise its originality if not its sales.
- 2003
Pentamind champion
Wins the Pentamind at the Mind Sports Olympiad for the fifth time, taking five of the six years from 1998. The event crowns the best all-round games player in the world across five different games.
- 2004
Diplomacy world champion
Part of the winning team at the Diplomacy World Team Championships. Through his twenties and early thirties he also plays poker around the world, including the World Series of Poker, amassing over a hundred thousand dollars.
- 2010
DeepMind
Co-founds the AI research lab with Shane Legg and Mustafa Suleyman, with a stated aim: solve intelligence, then use intelligence to solve everything else.
- 2014
Google, and a harder game
Google acquires DeepMind. The team starts building AlphaGo — aiming not at chess, where a correct move exists, but at Go, which has more possible board states than there are atoms in the known universe and was widely thought unsolvable.
- 2016
Seoul
AlphaGo beats Go grandmaster Lee Sedol four games to one at the Four Seasons in Seoul. Sedol had predicted a landslide in his own favour. In game two, move 37 breaks with human strategy so completely that observers cannot tell whether it is a blunder.
- 2019
Sedol retires
The world's strongest Go player leaves the game in his prime, describing AI as an entity that could not be defeated.
- 2024
Nobel Prize in Chemistry
Awarded for AlphaFold, which predicted the three-dimensional structure of every protein in existence — hundreds of millions of them — and published the results in a public database.
- 2024
The Thinking Game
A documentary on his life and his pursuit of AI as the ultimate tool for humanity is released. He estimates human-level machine intelligence is only a few years away.
Major achievements
What Demis Hassabis gave the world.
- 01
AlphaFold and the 2024 Nobel Prize
Proteins are the building blocks of life, and their shape is how they work. AlphaFold predicted the structure of every known protein and gave the database away — the work that won him the Nobel Prize in Chemistry.
- 02
AlphaGo
Beat the best Go player alive four to one in 2016, years earlier than the field expected, on a game whose complexity was supposed to put it out of reach of machines for decades.
- 03
DeepMind
Co-founded in 2010 to teach computers to solve abstract problems where information is distorted, delayed or incomplete. Google acquired it in 2014.
- 04
Theme Park
Built as a teenager killing time before university. Its depth — neglect the signposts and nobody finds your gift shop; overprice the candy floss and visitors go hungry — made it a cult classic and a genre.
- 05
Five-time Pentamind champion
Winner five times in six years at the Mind Sports Olympiad, a competition built to mirror the Olympics but for mental acumen, scored across five different games.
- 06
A 2300 chess rating
A FIDE peak that made him the second-best player in the world for his age group. He also won a Diplomacy world team title and over a hundred thousand dollars at poker.
Turning points
The moments that decided it.
- Aged eight
A ZX Spectrum, bought with chess winnings
He spent prize money from a youth chess tournament on an 8-bit home computer and used it to write chess programs. To make a machine play chess he had to work out how he himself thought about chess — and the loop between the two became the whole of his career.
- Gap year
Win a job at Bullfrog
A studio ran an open contest to redesign Space Invaders. He submitted Chess Invaders, replacing the aliens with descending chess pieces — his two obsessions welded together — and won a job instead of a prize.
- 2010
Solve intelligence first
He had gone from managing a virtual theme park to simulating an entire fictional republic. DeepMind inverted the ambition: rather than build software to solve a problem, build the thing that solves problems, and point it at everything afterwards.
- 2016
Move 37
A machine played a move no human would have played, in a game humans had refined for thousands of years, and it was not a mistake. It was the moment his thesis stopped being a thesis: AI was not just calculating faster, it was finding things nobody knew were there.
I had this feeling that this was a kind of magical device. Looking back on it now, just as if you think of a car as a machine that amplifies human capabilities physically, the computer to me felt like that for the mind. I think I intuitively understood that as a child.
Demis Hassabis, on the ZX Spectrum he bought aged eight
What you’ll take away
Lessons that outlast the headlines.
- 01
The hobbies were the work
Chess, poker, Diplomacy, Go, game design. They look like the scattered enthusiasms of a restless mind. They are one question asked repeatedly: how does a mind decide when the information is incomplete?
- 02
Build what amplifies you
He saw a computer at eight the way others see a car — a machine that extends what a person can do, except pointed at the mind instead of the legs. Everything since has been that same instinct at a larger scale.
- 03
Aim at the impossible one
Chess had a right answer, so DeepMind ignored it and took Go, which the field called unsolvable. Then protein folding. The problems people have given up on are the ones where a breakthrough still counts.
Did you know
Fun facts.
- 01
Did you know that Google DeepMind founder Demis Hassabis won the 2024 Nobel Prize in Chemistry for his work using AI to map the 3D structure of hundreds of millions of proteins?
- 02
Did you know that before his groundbreaking work in AI, Demis Hassabis was a five-time champion of the Pentamind, an elite competition designed to crown the best all-around game player in the world?
- 03
Did you know that as a teenager taking a gap year before Cambridge, Demis Hassabis won a job designing games and helped create the revolutionary sandbox game *Theme Park*?
Questions
Good to know.
The 2024 Nobel Prize in Chemistry, for using AI to map protein structures. DeepMind's AlphaFold predicted the three-dimensional shape of every protein in existence — hundreds of millions — and published them in a public database, opening the door to work on disease and famine.
The AI research company he co-founded in 2010 with Shane Legg and Mustafa Suleyman, acquired by Google in 2014. Its aim, in his words, is to solve intelligence and then use intelligence to solve everything else — teaching machines to handle problems where information is distorted, delayed or incomplete.
Go has more possible board states than there are atoms in the known universe, which is why it was considered unsolvable by machines. AlphaGo won four to one in 2016, years ahead of expectations. Sedol, who had predicted a landslide win, retired from the game in 2019, calling AI an entity that could not be defeated.
Yes. He won a job at Bullfrog Productions during his gap year by entering a contest with Chess Invaders, then worked as project leader and lead programmer on the 1994 cult classic Theme Park. In 1998 he founded Elixir Studios, which made the AI-driven political simulation Republic: The Revolution.
He estimates it is only a few years away, and sees AI research on the cusp of major discovery. In his Nobel lecture he went further, saying he believes AI could eventually deliver a grand unified theory of everything.
15 min · one sitting