In his 1967 letter to Andre Weil, Robert Langlands outlined, among other things, what came to be known as the Langlands $L$-group. It is a semi-direct product of a Galois piece and a complex reductive group and it turns out to play a crucial role in the Langlands conjectures, both locally and globally and may other places. I will attempt to introduce this construct and illustrate its importance in the representation theory of p-adic groups (of characteristic zero—for this talk), and perhaps also some global aspects of its role if there is time. This will be a survey talk.