And the reason for that hope has nothing to do with mechanics, and everything to do with DOS being a rubbish setting.
I mean, for me it's absolutely both.
But let me really specific about the crap mechanics, it boils down to three fixable things:
1) The weird-ass Armour/Magic system of DOS2 was, frankly, one of the worst and most counterintuitive combat designs I've ever seen in a turn-based game or a CRPG (balanced parties being an actively bad idea because of it). I believe that, if it had appeared in a different game, critics would have trashed it. DOS2 got away with it, rather than being good because of it, because shenanigans allowed you to work around it.
2) DOS2 (again specifically - DOS1 had this a bit but not as bad) made you upgrade/replace items constantly, to the point where, if you found a cool item, the odds were good it was either unusable, because it was slightly lower level than you, or rapidly became so. This was a very bizarre design choice, where there was a sort of 45 degree linear power slope on item power - so a sword from L3 would be basically a liability by L5. That kind of works in Diablo where loot constantly drops, but DOS2 didn't do that, so this focused the game on item upgrades/replacements, which wasn't good design, nor congruent with the setting/story.
3) Surfaces excessively dominated gameplay to the point where fights became exceptionally boring and repetitive unless you engaged in shenanigans outside the fight and "difficult" fights were reliant on hidden "gotcha" mechanics to change things up rather than actually engaging with their own mechanics. Again DOS2 had this much worse than DOS1.
All of that could be fixed - and fairly easily. But will it be?
I will also say, I think, at this point, significant numbers of action points that can be used very flexibly are a design cop-out/failure of design. Me from 20 years ago would be astonished to hear me make such a comment, but I've played a lot of turn-based tactical games now, and the design restrictions imposed by having more specific limitations on actions - which most modern games do, to be fair - produce tighter and vastly more tactical and interesting gameplay.
I do worry a bit about Larian's gameplay design capabilities. Writing-wise, if they actually choose to do a new IP, they're golden. There's no question they more than have the talent to create a fascinating story with memorable characters. But Larian have actually never made a game with good gameplay design. The closest they've come is Honour Mode rules in BG3 - but that's just BG3 actually doing close-to-5E rules! If we look back, DOS1 and DOS2 are completely unbalanced messes which show no particular aptitude for game design, and all their games before that had notably crap gameplay, like, not even good by Eurojank standards (and I do love Eurojank games).
I'd worry less if it wasn't for Swen's leadership and Bethesda's example. Bethesda have, in Todd Howard, a Swen-like figure. Todd Howard thinks BGS' games have good, modern gameplay. He is uncomplicatedly and objectively incorrect. Even at the time Skyrim came out, its gameplay was wildly lacking compared to similar contemporary games (Dark Souls, Dragon's Dogma), and only one game since Morrowind (so in 20 years) has seen any improvement - Fallout 4 - and it was a very minor improvement strictly to shooting gameplay. Starfield doesn't play any better than Fallout 4, despite 7+ years of development - in fact in many ways it's worse. And Swen has said he thought DOS2 had much better gameplay than BG3. He is, like Todd Howard, objectively incorrect. But given he was the director on all three games, and likely will be on future games, it tells us what he's going to see as "good gameplay".
Now, if Larian had highly experienced gameplay and systems designers, who could easily push back on Swen, that might be okay - but they don't. Most of the people designing gameplay and systems for Larian have literally no experience in the games industry prior to DOS2 or BG3. You can see that here:
The official game credits for Baldur's Gate III released on Windows in 2020. The credits include 2,943 people.
www.mobygames.com
This is a highly unusual situation. Hopefully it all turns out well, but I'd be surprised.