It could be argued that the drivers weren't done as good as they could have been before.
...
and would indicate (if they were honest) that they did a bad job the first time leaving so much performance untapped.
Yeah..except that's not the way it worked. If that was how development worked, we all should already have 12 nm process, flawless silicon for all of our devices, have lithium batteries with five times the power density that current batteries have...etc.
Saying that when a product or driver that improves the performance over previous iterations is released means that the original was done poorly implies that somehow people can magically come to the optimal solution on the first try. That's not how it works. Back when DX 11 was released, everything was all about tesselation and compute functions. The fact that the CPU landscape was going to flatline, innovation in game coding was going to fall flat due to Console anchoring, etc. wasn't widely known.
Developers worked as best as they could with drivers that they had when games released, the driver teams learned things about how developers were doing things by looking at code and profiling performance, then incorporated fixes and alterations into further drivers that allowed for better operation in further drivers.
That's just the way that ongoing development works. When performance goes up, especially for free, the response is typically "Yay!. Games will run better!" I can't see the point in being upset or feeling like someone held back on performance. It doesn't make empirical sense from knowledge of how these things are developed and rolled out, and it doesn't make business sense for a company to do this in the first place. Sure, sometimes things get missed, and when they are pointed out, usually it would be crazy for them not to address them as quickly as possible.