Viktoria Semaan on Instagram: "🚨 BREAKING: Text-to-SQL just hit 75%+ accuracy on the hardest benchmark. That means devs & analysts can spend less time writing SQL and more time focusing on insights. The breakthrough? Not bigger models, but smarter training. Instead of relying on human opinions (RLHF), Databricks team used Reinforcement Learning with Verifiable Rewards (RLVR) — where the reward is grounded in truth: does the query run & return the right answer? ✅ Faster than human feedback ✅ Che