I was really excited about Swift when it was first introduced, and I hoped it could grow beyond the confines of Apple's specific needs and ecosystem requirements. Unfortunately, that never really seemed to happen, and rather than Swift actually get better in the areas I care about (coming from a background in very expressive, dynamic languages like Ruby), in some ways it got "worse" (aka lots of verbose syntax all over the place simply to placate compiler-level concerns). SwiftUI was a bizarre advancement in that it attempted to add a very dynamic DSL mindset on top of a language which was far more rigid. Needless to say, everyone I've heard from who's tried to build production-grade UIs using SwiftUI has ended up feeling pretty burned. It's a grab-bag of cool ideas yet remains a buggy nightmare in practice.
I don't really know what Apple can do per se to turn this ship around, but Swift "as a brand" has definitely lost its luster. I'd considered wading back into Apple app development when SwiftUI was first introduced (at the time I hadn't touched any of the dev tools since early-era Mac OS X Cocoa/Objective-C), but I have no interest at this point. (The App Store being a real s**show also doesn't help matters!)
Having built a few apps with pure SwiftUI, I think you are exaggerating a bit. Yes, there are many many problems with it. I wouldn’t blame those on Swift though, since most people‘s gripes are with runtime issues.
I don't really know what Apple can do per se to turn this ship around, but Swift "as a brand" has definitely lost its luster. I'd considered wading back into Apple app development when SwiftUI was first introduced (at the time I hadn't touched any of the dev tools since early-era Mac OS X Cocoa/Objective-C), but I have no interest at this point. (The App Store being a real s**show also doesn't help matters!)