Would those optimizations be possible if this was tested out in the standard library preview package?