we also have people like joscha bach and yoshua bengio working on alternative networks like generative flow networks that learn by sampling whatever data available unlike deep learning that needs a lot of traning dataset, almost like how humans learn.