We're going to have +20% or maybe +25% better odds from proficiency at max level, depending on where Larian land, but check difficulty will likely also increase. I don't think that will make the core issues go away: the pass/fail result will still be very random, many characters will have the stats and skills to reasonably achieve either outcome, and one outcome will still be significantly more interesting than the other.
I'm aware this is not going to see any massive changes, hence the "tilting at windmills" comment, but I do think the current design is a quite naïve translation of the rules as written and doesn't work very well in practice. It is probably my biggest gripe with the game as is. I am hoping that at least Larian will choose the failure results more care for future content, even if they have less care than they do with the successes.
Personally I'm very fond of 13th Age's fail forward ethos for checks, where the interpretation of a failed roll isn't necessarily that you out-and-out do not succeed at your goal but can also be that you succeed with some unexpected side-effects if that's the more narratively interesting outcome. I adore their canonical example: a suave rogue attempts to schmooze a ship's officer. He rolls 4, needing 15. The outcome is that, oh yes, the ship's officer is very interested in you. He requests your presence for dinner ... while dropping multiple hints that he's a recreational cannibal, and dinner includes the boarders you just repelled.