Exposing Floating Point Numbers

Bartosz Ciechanowski published another masterclass visual essay, this time focusing entirely on floating point numbers and how they work.

I remember covering IEEE 754 floating point representation back in college, but the exact mechanics had gotten fuzzy over the years. Whenever floating-point inaccuracies came up in discussion, I’d usually resort to hand-waving: “It’s just how floats are represented in binary, yada yada yada.”

Walking through Ciechanowski’s interactive diagrams made the underlying mechanics click again. It’s an incredible refresher on sign bits, exponents, mantissas, and why subtle precision bugs happen.