My colours were nailed to the mast on this topic from the get-go. I honestly think Spell Versatility is bad design. It is to address a perceived problem with the inflexibility inherent in the "spells known" classes.
Well, it's not just a perceived problem, it is a very real difference in versatility, and thus effectiveness - in 3.5, it delineated the difference between Tier 1 and Tier 2 casters. It still does, in 5e, it's just that the gap is wider, because...
It is a fundamental change to the way in which those classes "know" spells. It is, quite simply, making "spells known" classes into "spells prepared" classes. It breaks down the distinction between the two different mechanics.
TBF, 5e already broke down that distinction.
Probably un-needed hiistory lesson: 3.0 introduced 'spontaneous casting,' with the sorcerer. The 3.x spontaneous casters had fixed known spells but flexibility in how they cast them with slots, from round to round (and, not for nothing, moar slots, more hps, and - probably for nothing, a few more weapons and a few less skill points), while the prepped casters used the more traditional Vancian memorization mechanic, able to change their spells every day, but needing to decide not only which spells they could cast each day, but how many times they'd be able to cast each of them. So the Tier 2 spontaneous casters had 'tactical flexibility' and the extra slots to 'spam' a spell turned out to be particularly good at them moment, while the Tier 1s had 'strategic flexibility' to bring the best spell for the job to any situation (they could forsee).
5e has both removed the extra slots from spontaneous casters, and given their crown-jewel, spontaneous casting mechanic, to prepped casters, so now they're 'known spell' casters, their casting strictly inferior to the unprecedented-even-in-3.x versatility of the neo-Vancian set.
And that is, apparently, a fire that needs more gasoline.