How much has really changed with AI?

The talks I heard today reminded me that AI hasn't rewritten the rules of quality thinking, it's just raised the cost of ignoring them. Shift-left and shift-right, risk-based prioritisation, separating behaviour from implementation - these are old, well-worn ideas, and they're still doing the heavy lifting, and are just as important as ever.  None of that makes experienced quality judgement less relevant. It makes it the thing that decides whether an AI-powered system actually works.