Yes. I should have noted for clarity that vdom diff is a special case, not a general one like that of subj. But their similarity reminded me of the experience.
It's a good point about edge cases in diffing. Dynamic value diffing in dynamically-typed languages is an interesting problem to think about.