TL;DR: No one is talking about Apple announcing it's _finally_ going to bring Web Push to Safari.
If you ask me, the biggest thing to come out of WWDC 2022 was the announcement of Web Push for Safari. MacOS is going to get it first and iOS will receive it in a later iOS 16.x update, sometime in 2023.
iOS. Is. Getting. Web Push. 💃
Apple announcing support for a web feature may not seem like a big deal, especially during WWDC. The conference is all about Apple's platforms, after all, not the web, right? Many would think that this is just a thing that Apple is finally enabling web apps to do, which is nice for the web, but how does this trump any other announcement made at WWDC? They announced macOS Whatever, iOS and iPadOS 16, new Macbooks with new M2 processors!
How can adding support for a single web feature beat anything they had to say?
Short answer: Web Push on iOS will change the "we need to build a native app" decision.
The lack of the web being able to do push notifications on iOS is probably the biggest reason why web developers see a potential project end up being built as a native app instead of a web app: Native apps get to send those sweet, sweet notifications! They are just so darn useful!
(And, I will admit, also a way to boost engagement for the sake of engagement if you're willing to choose The Path Of The Dark Side)
Until iOS gets Web Push support, every single project that (thinks it) needs notifications - a very basic requirement in this day and age - will end up being built as a native app.
For iOS. Probably also Android. And then a web app for desktop.
Push notifications are definitely not the sole reason to go native, but in my experience, it's one of the first things clients ask for. They may very well be the thing that pushes your client over the edge and forces them, you and the entire project to accept the logic of the app store model.
Some projects have requirements that can't be pulled off on the web - which is fine! Native apps can be great! It's just that there are a lot of cases where a web app could do the work and offer significant benefits compared to native.
Up until now developers haven't had the option to deliver a light touch approach to apps, because they've been forced to build a native app anyway. Now they maybe won't have to. Does anybody think that the world's developers can't build a better user experience when they don't need to make you open an App Store at all? Anybody who thinks that forcing an app installation in the middle of your user experience must make the best user experience, must have a pretty low view of designers' skills.
I remember this great talk Luke Wroblewski (formerly Apple, currently head of All Things iOS at Google) gave. He used bike sharing apps as an example of how you can completely mess up on-boarding: Endless tutorial screens, complicated and/or difficult-to-use sign up forms, etcetera. The thing that stuck out the most for me: The process a user has to go through before he even ends up there. You come up to a bike, you want to use it, but you have to install an app first. Even if you only use that bike sharing service once, you need to go to some App Store, find the app, install it, open it, and then the (in-app) on-boarding starts.
You could cut these steps down to (basically) one: scan a QR-code, it opens a web app. No app icon on your home screen, no updates. Just: Welcome to on-boarding.
This cuts down the (pre-) on-boarding by several steps.
To give a little perspective on how much of a barrier you would be taking down by not having a native app: According to research, people install an average of zero apps per month. What chance does your app have ending up on your potential users' home screens? Pretty much zero.
But how many websites do people visit? A lot - and I would bet most of those websites only get a single visit per person, ever.
So now that Web Push is coming to iOS, the question becomes: What if you could have that lower barrier-to-entry and notifications?
What if you wouldn't have to go through the App Store review process? (That would be pretty sweet, wouldn't it?)
What if that would require building a single (web) app instead of three?
What if... you could build your (client's) project as a web app?