Well, that’s the worst case scenario, it’s not representative of the average.
I deployed an app update to one of my snaps last week that ensures a minimum version of snapd that makes RAA mandatory, here’s the rollout stats.
That’s 90% of people who weren’t delayed two weeks, they were delayed enough to ensure the app wasn’t running whilst updated, which is hugely better than what would happen before (app instability).
There’ll almost certainly be some users who do end up hitting the fallback of refreshing while the app runs anyway, but it’s still a better position for most people than before.
And I totally agree that there could be improvements in the UX (e.g, triggering the refresh when the app is closed, scheduling updates during shutdown, improving the notification to have “Update Now”, etc), but from the publisher side of things, I personally think this is a much better position for the snap ecosystem to be in than without it, and hopefully more improvements will come later.