Reinforcement learning used to be seen as the path to AGI, but now it's more of a step in deep learning frameworks. Deep neural networks seem to be doing most of the work. Reinforcement learning is better for learning novel skills through interaction with an environment. Still, there will be cases where it's needed.