NYU professor Andrew Gordon Wilson critiques deep learning's focus on toy models, urging researchers to analyze real systems

Gael Varoquaux defended reductionism as necessary for scientific generalization.