Summary of The Creativity Code

Looking for the book?
We have the summary! Get the key insights in just 10 minutes.

The Creativity Code book summary

Editorial Rating



Books about machine learning and AI have proliferated in recent years, creating their own genre. Many address fears of job loss and machines run amok; others celebrate AI as a mechanical savior. Marcus du Sautoy provides a hype-free assessment of AI’s current and future capacity for creativity. Not surprisingly, as an Oxford mathematician, he makes math as much the central character as creativity and machine learning. This will delight math geeks but for everyone else, Sautoy’s exploration of painting, poetry, creative writing and music will entertain and inform. Most refreshingly, he describes today’s advanced AI in terms any layperson can understand.

About the Author

Multiple award-winning mathematician, Marcus du Sautoy, teaches advanced mathematics at Oxford University in England. Sautoy has earned a reputation for being able to explain complex science and math to everyday people.


Advanced AI now learns on its own, no longer reliant on explicit coding.

At the dawn of the information age, scientists believed that computers could never do more than what was explicitly written into their code. Today’s AI, based on “reinforcement learning,” proves them wrong.

The ancient board game “Go” requires more than strategy to master. The possible combinations of moves vastly outnumber those in chess, but the game’s real challenge comes in the creativity and intuition needed to excel at it. Indeed, even as AI began to defeat the world’s best chess players, few believed it could ever master Go.

Enter Demis Hassabis, a young chess wizard brought up on video games and early generation PCs. After completing advanced degrees in computing and neuroscience at Cambridge and University College London, Hassabis founded DeepMind, a firm dedicated to creating an algorithm that could teach itself a range of skills, starting with Go. Hassabis and his team’s algorithms worked. By randomly trying things, saving what worked and discarding the rest, these algorithms were the first to use reinforcement learning, which proved a monumental breakthrough in AI.

In 2016...

Comment on this summary

More on this topic

Human Compatible
Deep Medicine
The AI Advantage
The Algorithmic Leader
The Ethical Algorithm
The AI Organization

Related Channels