I mean look it might become a useful tool in the future, not denying that. But what we are seeing is DLSS5 being shoved into a lot of mainstream games as a replacement for polish. We can let the big studios drive themselves into the ground chasing profits and slapping on whatever new tech they think will give them a sales boost, no complaints here.
But as it stands it's a capability that's over hyped for what it is, like you said yourself, art teams aren't using it as part of their art direction so it's clashing with the designs, and the hardware level barrier for entry isn't worth the money it costs to get there.
That's all subject to change, it could become a staple of the industry, it could become the linchpin that lets smaller games perform while still letting the game have great design. But right now there's too much opposition for too little benefit. It could be great, sure, but people playing games now don't care what it "could" be. They care what it "is", and right now for the most part what it "is" is an over hyped buzzword feature that jacks up the "cost" of a game for no real benefit to the player.
And to give a counter example, the same thing was said about Ray tracing when it initially came out. It was cool but it would practically set your PC on fire at the time, but even then, people liked it. It was overpriced and not really that different, but some people thought it was cool. DLSS5 doesn't even have that.