If WotC decides to resume designing games where magic isn't ludicrously better than nonmagic at solving any given problem, then I'll believe the conversation can become much more nuanced and intricate (and, thus, much more interesting).
But either magic needs to be nerfed, which will never happen because the caster fanatics will riot, or non-magic needs to be buffed, which will never happen because the caster fanatics will riot. Unless and until at least one of those two things is true, we're going to be sucked right back into this quicksand trap every time.
In 5e, revised or not, there are enough always-excellent spells such that any given spellcaster can have at least a very good response to nearly any problem the party might face, with some room left over for more tricksy choices at higher levels. Doubly so if the caster picks up some of the better rituals, since those can be cast without a slot, this freeing slots for more damage without losing utility. Classes with zero access to magic simply cannot replicate anything like that.
If you need to be a blaster all day and have no need for utility (whether exploration, investigation, or socialization), every spellcaster can do that quite well with only 2-4 spells, plus a good cantrip. If you need to be pure utility, many spellcasters can pull that off with just 3-5 spells, many of which have multiple uses (e.g. fly is also a good combat buff). If you need a mix...by level 5 or 6, that's not hard to do.
So...yeah. Versatility is king, because "I'm very good at a couple skills" just can't compete with the breadth that even a Sorcerer or Warlock can bring, to say nothing of the Wizard.
Particularly because 5.5e has now actually made Batman Wizard real. Wizards can now change out one prepared spell per short rest, meaning every spell a Wizard knows is accessible with just an hour's nap.
