I introduce the Poisson process of sprinkling into a Lorentzian manifold as a mainstay of research in causal set theory, by means of which we can get our hands on causal sets that have Lorentzian manifolds as approximations. I will explain the slogan ``nothing beats Poisson'' and explain why it is provably the case in one sense but it remains a conjecture in a more physically meaningful sense. I will argue that a causal set is the only combinatorial entity that can have a Lorentzian geometry as an approximation by showing that a particular combinatorial simplicial complex cannot do the job and explaining what it is about causal sets that allows them to succeed where the combinatorial simplicial complex fails.