Come for the jokes. Stay for the performance skills. What happens when Gen AI tries to tune your database? You often get a mix of odd ideas, confident but incorrect suggestions, and a great opportunity to learn how real performance tuning works. In this lively one hour session, we explore the worst practices Gen AI might confidently recommend and turn each one into a useful teaching moment. You will see questionable indexing advice, unusual query tuning logic, confusing join guidance, and chaotic locking suggestions. We first laugh at them, then fix them with clean and practical tuning principles. We start by looking at foundational tuning areas that often go wrong when AI tries to guess its way through performance problems. From there, we walk through key parts of the query pipeline that tend to suffer when recommendations are not rooted in real workload behavior. We also look at how AI can misread performance signals and how that leads to poor tuning decisions related to resource usage and concurrency. Each section highlights why the advice fails and what the real approach should be. By the end of this session, you will know how to • Avoid common indexing mistakes • Fix bad query patterns the right way • Prevent join related performance issues • Handle blocking safely • Read waits without guessing • Keep parallelism under control This session is fun, fast, and packed with real world tuning lessons.