
Prediction Machines demystifies AI economics, showing how falling prediction costs transform business decisions. Nominated for Thinkers50's "Oscars of Management Thinking," this guide by Toronto's elite economists reveals why human judgment becomes more valuable as AI advances - not less.
Feel the book through the author's voice
Turn knowledge into engaging, example-rich insights
Capture key ideas in a flash for fast learning
Enjoy the book in a fun and engaging way
Break down key ideas from Prediction Machines into bite-sized takeaways to understand how innovative teams create, collaborate, and grow.
Distill Prediction Machines into rapid-fire memory cues that highlight Pixar’s principles of candor, teamwork, and creative resilience.

Experience Prediction Machines through vivid storytelling that turns Pixar’s innovation lessons into moments you’ll remember and apply.
Ask anything, pick the voice, and co-create insights that truly resonate with you.

From Columbia University alumni built in San Francisco

Get the Prediction Machines summary as a free PDF or EPUB. Print it or read offline anytime.
Imagine a world where machines can predict what you'll buy before you know you want it. Where your health problems are diagnosed before symptoms appear. Where traffic accidents become rare because vehicles anticipate hazards seconds before humans could notice them. This isn't science fiction-it's the economic revolution unleashed when prediction became cheap. The breakthrough moment came in 2012 at the ImageNet competition, when neural networks suddenly slashed image recognition error rates by an unprecedented margin. This triggered Google's $600 million acquisition of DeepMind and eventually prompted China's multi-billion-dollar AI investments. What happened? Prediction-filling in missing information using data you have-became radically cheaper, setting off ripples that continue transforming our economy and society. What makes this revolution so powerful yet deceptively subtle? When something fundamental becomes drastically cheaper-whether artificial light in the 1800s or prediction today-it transforms society in ways both expected and surprising. An improvement from 98% to 99.9% accuracy might seem modest, but it reduces errors twentyfold-enough to completely reshape industries. And prediction machines are now tackling problems we never previously thought of as prediction challenges: Google Translate improved by asking "what English text would a human translator produce?" rather than adding grammatical rules.