Frank Jons: chromatic explosions!
He could have ended up as a partner in an auditing firm, but a job interview convinced him to follow his true calling: painting. Interview.
Demis Hassabis (DeepMind): A Nobel Prize of Chemistry goes to a Geek!
Not many people have turned a passion for gaming into a Nobel Prize. But for Demis Hassabis, Co-Founder and CEO of DeepMind, a lifelong enthusiasm for the logic of games and programming led to the most prestigious prize in science.
A Lifelong Passion
Born in London in 1976, Hassabis developed an interest in chess at the age of four and in computers at eight. He taught himself programming from books, first coding a version of the game Othello, and became a chess master at 13. After completing school two years early, he turned these passions into a profession by coding for Bullfrog Productions, where he was lead developer on the influential game Theme Park. He went on to study computer science at the University of Cambridge, and then became lead AI programmer at Lionhead Studios, where he worked on the game Black & White. In 1998, he founded Elixir Studios, where he produced award-winning games including Republic: The Revolution and Evil Genius. A common thread ran through all this work – the use of AI.
Building AI
From 2005, Hassabis took his enthusiasm for AI to academia, with a PhD in cognitive neuroscience at the University College London followed by postdoctoral work at Harvard and MIT. His papers on autobiographical memory and amnesia were published in some of the world’s most prestigious journals. In 2011, he co-founded AI startup DeepMind, which developed and trained AI models using video games, with the goal of creating an artificial general intelligence (AGI) that could solve the world’s problems. DeepMind became the basis for AlphaGo, which defeated the top go player four games to one in 2016, and a more general neural network that mastered chess and shogi. Having trained on games, it was time for DeepMind to tackle real world problems.
“Software built on a passion for games earned him the greatest prize in science.”
From Games to the Building Blocks of Life
Proteins are fundamental to life, controlling its chemical functions, so mapping them has the potential to transform human health. Scientists have been working towards this goal for years, and Hassabis turned his AI onto the problem. He and John Jumper developed an AI called AlphaFold2, trained on existing databases to find likely protein structures. By 2020, it reached an accuracy of 90 percent, predicting the shape of proteins down to the nearest atom and solving one of the greatest challenges in chemistry. In 2021, it calculated the structure of almost all 50,000 human proteins, before moving on to the 200 million known proteins that make up all life. For this transformational work, Hassabis was one of three recipients of the Nobel Prize for Chemistry in 2024. Software built on a passion for games had earned him the greatest prize in science.