Our modern world runs on math, from social media algorithms and package delivery to weather prediction, medical imaging, and machine learning. In an interview with Ars Technica, Cornell mathematician Steven Strogatz argues that this big math is now beginning to outgrow the humans who created it. He explores the idea in a new book, "Big Math: The Hidden Codes That Run Our World," co-written with his Cornell colleague Alex Townsend, who studies numerical analysis and deep learning.
The largest AI language models today use trillions of parameters, the adjustable numbers inside a model. Yet the underlying math is often surprisingly old and simple: matrices, meaning grids of numbers, along with vectors, networks, and optimization. When problems are linear, they scale up to huge sizes and stay manageable. Non-linear problems, by contrast, resist exact treatment.
Strogatz stresses that handing math over to tools is not new. Anciant Chinese solved systems of linear equations by moving bamboo sticks on a grid. Computers started as obedient servants, then they became partners.
From tool to oracle
The book also shows big math shaping public life. One example is gerrymandering, the practice of drawing voting district maps to favor the party in power. Mathematical tests such as the efficiency gap can measure whether a map wastes some voters' ballots unfairly, but the US Supreme Court has resisted such arguments.
This story matters because the stakes keep growing as machine learning, and soon quantum computing, supercharge these tools. Big math already routes planes, predicts the weather, and powers chatbots, yet it is drifting beyond what anyone fully understands. The authors close with a plea: knowledge must serve something larger than itself. The work of this century, they write, is letting our power grow without losing our soul. Big math keeps getting bigger, so must we.