Title: What can geometry tell us about theoretical computer science?

Abstract: A fundamental problem in theoretical computer science has recently (thanks to AI)  become a practical one: how to multiply large unstructured matrices efficiently? I will explain how to approach this question as a problem in geometry, more precisely algebraic geometry and representation theory. The geometric objects in question are tensors, i.e., multi-linear maps. I'll begin with an overview of tensors and their uses in numerous areas of mathematics (and other areas such as engineering and computer science) and then focus on the geometry of the matrix multiplication problem.