Evals: The Skill That Makes AI Survive Production
Alex and Jordan go deep on AI evals, the number one skill for AI engineers: why normal software testing breaks with AI, model evals vs product evals, the three ways to score, calibrating LLM judges, Goodhart's Law and reward hacking, and the playbook for building your first eval loop. AI Generated
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.