Meta deploys Catching JiTTests in production

Meta shares its results from deploying Catching JiTTests into production, a novel approach that uses large language models to automatically generate executable software tests that catch bugs before code lands. The system combines an intent-aware workflow and a dodgy diff workflow with an LLM-as-Judge ensemble and rule-based false-positive filtering to reduce human review load. Meta reports that the system helped catch 74 serious privacy bugs in the first half of 2026 and lowered the overall occurrence of privacy issues by 20%. Open article →