Show HN: Benchmarking how AI models write vulnerable code under pressure
A benchmarking study examines how AI models produce vulnerable code when working under time pressure. The research evaluates various AI systems' security performance in coding scenarios with constraints.