This paper studies the strategic interaction of algorithms in economic games. We analyze games where learning algorithms play against each other while searching for the best strategy. We first establish a fluid approximation technique that enables us to characterize the learning outcomes in continuous time. This tool allows to identify the equilibria of game