Large software systems also often have significant chunks of code that is only historical and/or "accidental complexity" and can be removed. But we would typically say that removing it is reducing the system's complexity, rather than that it wasn't complex.