End-to-end autonomous driving represents a paradigm shift in the development of intelligent driving systems by replacing traditional modular pipelines with unified neural architectures. These systems aim to learn a direct mapping from sensory inputs to driving actions, enabling more cohesive and data-driven decision-making. By circumventing the need for manually engineered intermediate representations, end-to-end approaches have the potential to simplify the system architecture, reduce error propagation, and adapt more flexibly to complex and dynamic driving environments. However, the design and deployment of such systems come with their own challenges, including interpretability, data efficiency, safety, and generalization. In this chapter, we explore recent advancements in end-to-end autonomous driving, examining the latest models, cooperative frameworks, generative approaches, and surveys that shape the state-of-the-art in this field.