Karl Iagnemma - Holistic Safety Arguments for Neural Networks (via searchlore.ai)