1821: Arthur Cayley is born

On 16 August 1821, Arthur Cayley was born in Richmond, Surrey. Over a long career he produced nearly a thousand mathematical papers and helped establish the modern British school of pure mathematics. Among his many creations, one would prove unexpectedly central to the machines we now call intelligent: the abstract theory of matrices.

In the 1850s Cayley treated rectangular arrays of numbers as mathematical objects in their own right. He defined how they could be added, multiplied and inverted, and he proved the celebrated Cayley–Hamilton theorem. At the time the work looked like pure abstraction, remote from any practical calculation. Matrices were simply a new language for capturing linear relationships.

That language quietly became indispensable. When electronic computers appeared a century later, matrix arithmetic turned out to be the natural medium for solving large systems of equations, transforming geometric data and analysing observations. By the 1980s and 1990s, researchers rediscovering neural networks realised that both the forward computation and the backward propagation of errors could be expressed cleanly as sequences of matrix multiplications. The same compact notation Cayley had introduced on paper now ran at enormous scale on specialised hardware.

Today every major deep-learning framework is built around highly optimised matrix engines. Whether a model is translating a sentence, recognising a face or generating an image, the dominant operation inside its silicon is the linear-algebraic step that Cayley first made rigorous. His Victorian abstractions travel, almost unnoticed, through every layer of contemporary artificial intelligence.

Cayley’s career is a reminder that tools we now regard as essential to AI often began as pure intellectual play, far removed from any dream of thinking machines. The path from a nineteenth-century notebook to a GPU cluster is long, yet the mathematical thread remains unbroken.