[ Field Note · EMBA ]
Mar 27, 2026
The Link Between Economics and Machine Learning
We've been telling the AI story wrong. Not completely wrong — but incomplete in a way that matters.
For years, the default analogy for Machine Learning has been biological: neural networks modelled on brain cells, artificial neurons firing like synapses, deep learning as a rough sketch of how human minds work. It's a useful story. But after completing the Game Theory module of my EMBA, I realised it obscures something more fundamental.
Under the hood, modern AI — particularly Reinforcement Learning — isn't primarily mimicking biology. It is executing pure mathematics. An RL agent doesn't "think" like a brain. It calculates payoffs, searches for stable equilibria, and optimises utility functions across a strategic environment. It is acting as a rational player in a formal game.
The intellectual origin of this leads back to someone I feel a deep cultural connection to. To the rest of the world, he is John von Neumann. To us Hungarians, he is Neumann János. His 1944 work, Theory of Games and Economic Behaviour, established the exact mathematical framework for rational decision-making under strategic uncertainty — payoff matrices, minimax strategies, the foundations of utility theory. The conceptual architecture of modern Reinforcement Learning is sitting right there on the page, eight decades early.
The "magic" of AI isn't magic. It is math that a Hungarian genius largely wrote before the first commercial computer even existed.