Demis Hassabis: DeepMind and the Pursuit of Artificial General Intelligence

Demis Hassabis is the co-founder and CEO of Google DeepMind. He is a computer scientist, AI researcher, and neuroscientist whose work focuses on reinforcement learning and general-purpose artificial intelligence. In 2024, he was co-awarded the Nobel Prize in Chemistry for his work on predicting protein structures using AI.

Early Career and Game Development

Born in London in 1976, Hassabis was a competitive chess player in his youth, reaching a master standard by age 13. At 15, he began working at Bullfrog Productions, contributing to the development of the simulation game Theme Park. His work involved programming dynamic behaviors for the game's non-player characters, which sparked his interest in developing systems capable of learning from experience.

Academic Background and DeepMind

After his time in the video game industry, Hassabis earned a PhD in cognitive neuroscience from University College London, researching the neural mechanisms of memory and imagination.

In 2010, he co-founded DeepMind with the objective of developing general artificial intelligence. DeepMind initially focused on deep reinforcement learning, demonstrating its capabilities by training an AI to play classic Atari games using only screen pixels and the goal of maximizing the score. The system independently developed successful strategies without pre-programmed rules.

AlphaGo

In 2016, DeepMind’s AlphaGo defeated Lee Sedol, a leading professional Go player. Go is a complex board game that relies heavily on intuition and pattern recognition. During the second game of the match, AlphaGo executed "Move 37," a highly unconventional placement that demonstrated the model's ability to identify novel strategies outside established human gameplay.

AlphaFold

Following AlphaGo, DeepMind focused on the scientific challenge of protein folding-predicting a protein's 3D structure from its amino acid sequence. In 2020, DeepMind introduced AlphaFold 2, which achieved experimental-level accuracy in its predictions.

By 2022, DeepMind released the predicted structures for nearly all known proteins, totaling around 200 million. This database has since been utilized by researchers for various applications, including drug discovery and biological research.

Current Leadership and AI Safety

Hassabis currently leads Google DeepMind, overseeing the development of models such as Gemini. He has advocated for AI safety and international regulation, emphasizing the need to align advanced AI systems with human goals.