Advertisement

The incredible life of DeepMind founder Demis Hassabis, the computer whiz who sold his AI lab to Google for £400 million

Demis DeepMind
Demis DeepMind

Google

DeepMind cofounder and CEO Demis Hassabis.

Demis Hassabis, the 40-year-old cofounder of renowned artificial intelligence (AI) lab DeepMind, is recognised worldwide as one of the smartest thinkers his field.

Nicknamed the “superhero of artificial intelligence” by The Guardian, Hassabis is a former child chess prodigy with degrees in computer science and cognitive neuroscience from Cambridge and University College London respectively.

Hassabis co-created the video game “Theme Park” game when he was just 17-years-old, before going on to found his own videogames company, and eventually DeepMind in 2010.

In January 2014, Hassabis sold DeepMind to Google for £400 million in what is Google’s largest European acquisition to date. The company made history last year when its self-learning AlphaGo agent beat a world champion at the notoriously difficult Chinese board game Go. Now DeepMind is turning its attention to applying its algorithms to areas that can benefit humanity, including healthcare and climate change.

Hassabis was born in London July 27, 1976, to a Greek Cypriot father and a Chinese Singaporean mother.

Hassabis is the eldest of three siblings and his parents are teachers. According to The Guardian, his sister is a pianist and composer, while his brother is a studying creative writing.

“My parents are technophobes,” he said during an interview with The Guardian. “They don’t really like computers. They’re kind of bohemian. My sister and brother both went the artistic route, too. None of them really went in for maths or science … it’s weird, I’m not quite sure where all this came from.”

Hassabis now has two young boys of his own.

From an early age, he showed a natural aptitude for board games, particularly chess.

Hassabis first got interested in chess at the age of four while watching his father play a chess game against his uncle, according to Wired. Two weeks later, Hassabis was beating adults at the game.

By the age of five, he was competing nationally and he won the London under-eights championships at the age of six. When he was nine he was captaining England’s under-11 team.

Hassabis bought his first computer when he was eight-years-old. It was a ZX Spectrum.

Hassabis bought the machine with £200 that he won from a chess match.

“The amazing thing about computers in those days is you could just start programming them,” Hassabis told Wired. “I’d go with my dad to Foyles, and sit in the computer-programming department to learn how to give myself infinite lives in games. I intuitively understood that this was a magical device which you could unleash your creativity on.”

At the age of 13, Hassabis reached the rank of chess master. He was the second-highest-rated player in the world under 14 at the time.

He finished his GCSEs when he was 14, two years ahead of everyone else in his class. He went on to take his maths A level when he was 15, followed by A levels in further maths, physics, and chemistry when he was 16.

He applied to the University of Cambridge and got a place but Cambridge wouldn’t let him start because he would have only been 16 — so he took a gap year.

He started his career in videogames at UK studio Bullfrog Productions when he was 15 after winning a competition for a job in Amiga Power magazine. At Bullfrog, he co-designed and led programming on “Theme Park,” which challenges players to build a successful theme park.

“The most fun I had in games was early in my career in the 90s,” Hassabis told PCGamesN last July. “Especially at Bullfrog, I was lucky to be there at the most golden period it had. Maybe that has ever existed in the UK industry, if you look at the games it produced one after the other.”

During his time at the company he worked under legendary games designer Peter Molyneux, who was the founder of Bullfrog Productions.

“I think we influenced each other a lot,” Hassabis told PCGamesN. “We worked together very closely for a number of years — it’s hard to say who [influenced who] more but it was a very important part of my life.”

“Theme Park” was released in 1994 when Hassabis was 17 and it went on to sell millions of copies.

Hassabis left Bullfrog Productions in 1994 to study computer science at Cambridge.

Undergraduates at Cambridge were taught how to develop “narrow AI”, which is able to learn how to perform specific tasks, but Hassabis was always more interested with developing “general AI”, according to The Financial Times.

He graduated from Queens’ College Cambridge when he was 20 with a double first-class honours degree in 1997.

After graduating in 1997, Hassabis worked at Lionhead Studios under Molyneux once again.

At Lionhead, Hassabis worked on an early prototype version of the AI for iconic god game “Black & White”.

He left Lionhead around a year later to found his own videogames company.

In 1998, Hassabis founded Elixir Studios, which produced award-winning games for global publishers such as Vivendi Universal and Microsoft.

Elixir, which employed around 60 people at its peak, made AI simulation games such as “Republic: The Revolution” and the “Evil Genius, ” which were both BAFTA-nominated.

Acccording to The Financial Times, Hassabis sold a 5% stake in Elixir to Eidos, which created the Lara Croft “Tomb Raider” series. The stake was sold for £600,000, valuing the company at £12 million.

After a decade in videogames startups, Hassabis returned to the world of academia at University College London (UCL) in 2005, where he completed a four-year PhD in cognitive neuroscience.

During his PhD, he sought to find inspiration in the human brain for new AI algorithms.

His research on memory and imagination was listed in the top 10 scientific breakthroughs of the year by Science in 2007.

In 2009, Hassabis was then awarded a Henry Wellcome postdoctoral research fellowship to continue his research at UCL for a further year.

He also completed research stints in Boston, spending time at Harvard and…

… the Massachusetts Institute of Technology (MIT).

In 2010, Hassabis founded what would go on to be his biggest company to date: DeepMind.

DeepMind is a London-based startup that wants to “solve intelligence” and use it to “make the world a better place.”

The company is developing sophisticated self-learning algorithms that can excel at particular tasks when it is given a dataset to learn from. The algorithms are created by blending research and expertise from neuroscience and machine learning.

So far, the algorithms have been used to defeat the best human player of Chinese board game Go and to help Google slash its enormous electricity bill. DeepMind is also applying its algorithms to a number of NHS projects.

Hassabis cofounded DeepMind with childhood friend Mustafa Suleyman.

Suleyman is head of applied AI at DeepMind and head of the DeepMind Health division, which is working with the NHS on a number of projects.

Suleyman is also bright, having gained a place to study philosophy and theology at the University of Oxford. However, he dropped out in his second year when he was 19 and went on to launch the Muslim Youth Helpline. Suleyman went on to work as a policy officer for then Mayor of London, Ken Livingstone. Following that, he founded “Change Labs,” which is a consultancy aimed at navigating complex problems.

And New Zealander Shane Legg, who was another postdoc at UCL’s Gatsby Computational Neuroscience Unit.

Legg is chief scientist at DeepMind. He obtained his PhD from IDSIA in Switzerland, where he was supervised by Prof. Marcus Hutter, an expert on theoretical models of super intelligent machines.

The 43-year-old works alongside Hassabis to lead DeepMind’s research. Much of Legg’s time is dedicated to hiring and deciding where DeepMind should focus its efforts next. Arguably more importantly, he also leads DeepMind’s work on AI safety, which recently included developing a “big red button” to turn off machines when they start behaving in ways that humans don’t want them to.

Legg tends to stay out of the limelight and gives significantly fewer talks and far less quotes to journalists than his other cofounders. With the exception of this rare Bloomberg interview, you’ll be hard pushed to find many stories about DeepMind that contain quotes from Legg.

Hassabis’s early investors include the likes of Tesla billionaire Elon Musk and Skype cofounder Jaan Tallinn.

Musk explained his DeepMind investment to Vanity Fair earlier this year.

“It gave me more visibility into the rate at which things were improving, and I think they’re really improving at an accelerating rate, far faster than people realise,” said Musk. “Mostly because in everyday life you don’t see robots walking around. Maybe your Roomba or something. But Roombas aren’t going to take over the world.”

Not all of DeepMind’s projects are well-known. Before DeepMind was acquired by Google, the company had an AI-powered fashion website called KITSEE.

KITSEE used AI to recommend clothes to people that they could then go on and buy. The website also featured a range of articles about fashion that were produced by a team of DeepMind writers.

KITSEE appears to have been abandoned around the time DeepMind was acquired by Google, suggesting the search giant may not have been interested in it.

In December 2013, DeepMind revealed that it had made a breakthrough by training a piece of software to play Atari games at a superhuman level by only using the raw pixels on the screen as inputs.

DeepMind was acquired by Google in 2014 for a reported £400 million when it had around 50 employees.

Today, DeepMind sits under Google parent company Alphabet and employs around 400 staff in King’s Cross, London. It also employs a small team at Google’s headquarters in Mountain View, California, that are working on applying DeepMind’s technology to Google products. However, it remains an independent organisation.

As part of the Google acquisition, Hassabis and his cofounders made Google set up an AI ethics board. Who sits on that board has never been made public.

In 2015, DeepMind made the front cover of Nature — something that many scientists dream of achieving — for its paper on how it created algorithms that could learn to master Atari arcade games.

After cracking the Atari games, Hassabis and his rapidly-expanding team turned their attention to the ancient Chinese board game Go — a game that has been widely regarded as the holy grail in the AI community.

Dating back more than 3,000 years, Go is a two-player board game that appears to be relatively simple on the surface — each player takes it in turns to lay a stone, with the objective being to surround the other player’s pieces.

However, the sheer number of potential moves on any given turn means that Go is in fact one of the most complex games humans have developed and AI scientists have been unable to master it for decades.

AlphaGo learns by playing thousands of games against itself and gradually learning from its mistakes.

In March 2016, DeepMind pitched its Go-playing algorithm “AlphaGo” against world Go champion Lee SeDol.

It was a five-game match that took place at the five star Four Seasons hotel in Seoul, South Korea.

DeepMind won convincingly, taking four of the five games.

Google cofounder and Alphabet president Sergey Brin came along to spectate.

As did former Google CEO and Alphabet executive chairman Eric Schmidt.

Lee Sedol was left stunned but he hasn’t lost a game since. Ever since the defeat, he’s been improving his game by practising on the AlphaGo algorithm.

The main programmer on the AlphaGo algorithm was a man named David Silver, who was a fellow undergraduate with Hassabis at Cambridge.

Silver and Hassabis also worked together at Elixir before DeepMind existed.

“Dave and I have got a long history together,” Hassabis told The Guardian in February 2016. “We used to dream about doing this in our lifetimes, so our 19-year-old selves would probably have been very relieved that we got here.”

DeepMind made the front cover of Nature for a second time in January 2016 for its work on the AlphaGo algorithm.

AlphaGo is now on its way to China to take on some of the best players in the country where the game was invented.

A number of AI experts, including Oxford professor Nick Bostrom, believe that DeepMind is currently winning the AI race, ahead of companies such as Facebook and Amazon.

Hassabis has won a number of honours awards during his lifetime, including this one from The Asian Awards in May.

In 2014, he won the Mullard Award from the Royal Society, and in 2016, he was ranked in Nature’s “10 people that matter this year”.

In 2017, Time nominated him as one of the 100 most influential people in the world.

The post The incredible life of DeepMind founder Demis Hassabis, the computer whiz who sold his AI lab to Google for £400 million appeared first on Business Insider.