I am not sure this is something other engines even offered at this level, but my issue is bindings support.
3.X had (3rd-party) production-ready bindings, even for niche languages.
4.X, with hopes of improving support for compiled languages, has a new bindings system meaning that all bindings need to be redone as a new effort. This happened with the language that I'm interested in, the group that made the production-ready 3.X bindings abdicated the crown and there have been splintered efforts by individuals to work on 4.X bindings.
So it (3.X vs 4.X) is language vs engine features. When/if 4.X bindings do come out, it is not known how similar they will be so (aside from non-Godot-specific code) that will likely add complication to it as well.
I don't really care about consoles (needing to jump through hoops to develop for it is one reason) so a different potential issue would web export limitations. Both for different languages and for visual quality (AA). Those were issues in the past, though I'm not actually sure where they're at now (the 4.1 docs do say you can't have C# web exports in 4.X).
I gave the title of it and I figured that would easily be found (title only because it was something I saw in not-logged-in YT recommendations, figured others may have seen it too).
I know that's probably rhetorical, but probably a similar problem to modern movies where (as described in the video Why Modern Movies Suck - They're Too Expensive) they are going after spectacle (rather than story or other elements) and due to cost they must make a 'safe' product to stay profitable, where a bland but universally palatable product will sell more tickets/copies than a stellar niche thing.
I'd also add that companies know they can usually ride the success of their own name/brand recognition. Even worse here with games because of pre-ordering, early-access as a product, and crowd-funding (which some wildly successful publishers still do--on top of unpaid self-promotion and all the other things--because people still think of them as indie).
As someone who admittedly usually has a bad experience with roguelikes, that seems about right.
My first playthrough was pretty easy until a character died quickly (it's all downhill from there), my second playthrough seemed a bit harder+more resource constrained+worse luck with research.
Also generally seems like one of those games where micromanaging is the key (even though I don't think it's really fleshed-out like games that typically do that), and some situational benefits (that may not pan out due to floor turns left, placement etc) that your characters completely lack at the beginning (also you can't see actives/passives on the character select screen). The generators seem expensive for how little they produce per-turn (at least without upgrades, and especially as their cost increases).
Then again, this is likely one of those games that'd be significantly easier with multiple players.