The reasoning systems are based on a technology called large language models, or L.L.M.s. To build reasoning systems, companies put L.L.M.s through an additional process called reinforcement learning.

Wild card Belinda Bencic levelled her head-to-head with Coco Gauff at two wins apiece after defeating the No. 3 seed in three sets to reach the BNP Paribas Open quarterfinals for the second time. Why ...