Dmitri Dolgov of Waymo Discusses Data-centric AI

Dmitri Dolgov of Waymo discusses the principles of data-centric AI, an approach that prioritizes improving data quality over solely iterating on model code. This methodology is crucial for complex applications like autonomous driving, where the performance and safety of the system depend heavily on the data it's trained on. The conversation covers how Waymo leverages vast amounts of real-world and simulated data to build more robust self-driving technology. This involves creating a powerful data engine to systematically collect, label, and curate datasets to address edge cases and improve model reliability.