The word "state" has become deeply entrenched in the lexicon of systems neuroscience. There are brain states and behavioral states, internal states and external states, discrete states and continuous states, circuit states and cell states... The taxonomy can be a bit overwhelming. Despite their popularity, there seems to be substantial confusion as to what constitutes a neural or behavioral state, and little agreement about how to determine them. This talk aims to do three things. First, I will give a rigorous definition of what constitutes a state by reviewing probabilistic state space models (SSMs) and showcasing some of my group's recent work on this class of models. Second, I'll try to convince you why SSMs are key to modern neuroscience, especially as we scale to brain-wide recordings and complex behavioral datasets, through example applications in a variety of model organisms. Third, I'll show you how to infer these latent states using novel algorithms for Bayesian inference and a software package that my lab is developing, which is aptly called SSM.