LangSmith has improved its regression testing experience with features such as the Comparison View and Display options that enable users to easily select, compare, and analyze multiple experiments. This allows AI engineers to quickly evaluate different prompts, models, cognitive architectures, etc., track performance over time, and identify interesting datapoints that behave differently between runs. The platform also provides a baseline run feature that highlights changes in evaluation metrics, making it easier to drill into specific datapoints and explore data across multiple runs. These features are crucial for quick iteration and comparing individual datapoints between two or more runs, which is a unique aspect of AI testing compared to software testing.