1924: Benoît Mandelbrot is born
On 20 November 1924, Benoît B. Mandelbrot was born in Warsaw. The mathematician who later spent thirty-five years at IBM’s Thomas J. Watson Research Center would give science a new language for roughness: fractal geometry.
Where classical mathematics preferred smooth curves, Mandelbrot studied coastlines, clouds, lung bronchi and commodity prices. He showed that many natural and social forms are self-similar—statistically alike at every magnification—and can be generated by simple recursive rules. Computers were essential to the discovery. In the 1970s and 1980s he used early graphical displays to render the set that now bears his name, turning pure iteration into images of unexpected beauty and infinite detail. His 1982 book The Fractal Geometry of Nature made these ideas accessible far beyond mathematics.
The practical consequences arrived quickly. Fractal algorithms supplied compact ways to synthesise realistic terrain, textures and antenna shapes. They also furnished quantitative measures of complexity—fractal dimension—that let researchers compare the irregularity of a coastline with that of a stock chart or a biological network.
Those same notions now circulate inside artificial intelligence. Engineers examine the fractal dimension of loss landscapes to understand why deep networks train as they do. Generative models that produce endless variations of images, audio or 3-D scenes frequently rely on multi-scale, self-similar processes first clarified by Mandelbrot. Even the analysis of attention patterns and the design of sparse neural architectures draw on scale-free statistics he helped introduce.
Mandelbrot never set out to build thinking machines. Yet by insisting that irregularity is information rather than noise, and by harnessing computers to see it, he supplied both tools and a sensibility that continue to shape how AI systems model, compress and recreate the world.