[[Benedykt Dryl]]'s [[Presentation]] on [[DD50 Summit - Where Technology Meets Business]] about `AI-Augmented Fitness Functions - scalability and maintainability sewn into PR`: [^1] > How to turn non-functional requirements into working quality "gates"? I'll show how to pigeonhole scalability and maintainability as fitness functions ([[SLO]]s, cost, architecture boundaries, [[Idempotence|idempotency]]) and use AI to detect architecture drift and predict the impact of changes on p95 and cost, while still in the PR stage. That is, how to practically use AI in the battle for shorter [[MTTR]], fewer post-deployment "surprises" and predictable costs. About [[Benedykt Dryl]]: ![[Benedykt Dryl#^bio|Bio]] [^1]: https://career.brainhub.eu/devduck/meetup50