Probabilistic forecasters, in contrast to point forecasters that output a single predicted value, produce a full distribution over possible outcomes, explicitly capturing uncertainty and enabling risk-aware decision-making. This approach is relevant in any domain where decisions depend on uncertain futures, such as energy systems, finance, and healthcare, where it is used to quantify variability in demand, outcomes, and risks. One way to construct such forecasts is through generative models. These models learn to sample from an unknown distribution given observations drawn from it. In forecasting settings, they can be applied autoregressively by sampling from the conditional distribution of the future given the past.