Some materials are neatly organized, like iron or nickel, while others are a complete mess, like spin glasses. It is the latter that I propose we examine. We will define a mathematical framework that is now fairly well understood, known as the ``Sherrington-Kirkpatrick model''. Then, we will introduce some of its properties, including the celebrated Parisi formula. Finally, we will see how, surprisingly, some questions about spin glasses can be tackled using branching processes.