The study of linear spaces of matrices arises naturally (and independently) in many different areas of mathematics and computer science.  In this survey talk, I will describe some of these motivations, and state and prove some (old and new) structural and algorithmic results.  Plenty of open problems remain.