But 0.004% is common in evolutionary terms - adds up to about 240,000 folks worldwide, by my poor maths - sufficient, surely, to have become fixed at some stage if it were “a better design”. In the case of many adaptive traits, we’re assuming mutations that occurred only once - and whales are renowned for their juggling an entire body plan in just a few million years to colonize a new niche, in comparison to which shifting the course of a nerve for better efficiency all round seems small beer.
I spoke from a medical background (small print in anatomy texts - the recurrent laryngeal nerve gets taught because of Phil’s reason of complications of Ca bronchus, but also because idiot surgeons might otherwise slice through it and turn people into mutes), but there seems no reason why the same, or a similar, mutation would not occur in other vertebrates, including giraffes.
But it happens, 4 in 100,000 times in humans - and it’s survivable, or it would not be in the anatomy texts as a variant. Whether it’s linked to other ill effects genetically I don’t know - polydactyly is, but not enough to prevent whole tribes in S America inheriting it. But once again, entire bauplans need to change for many adaptations - wings for flight, for example. So whilst one can plausibly invoke fitness landscapes to say why one set of things won’t happen, it’s absurd to invoke the same explanation when equally big things do happen.