Yeah, indeed. Made that post a bit too quick. I thought he was wrong, but wanted confirmation since it wasn't edited.
Should've added that after the interrupt occurence, the program counter gets placed on the stack and the ISR address gets loaded. Or just omitted that part.