There is so much time, effort, and physical waste that is generated by slightly redesigning phones every year purely for the sake of making sales (as opposed to meaningful improvement upon the existing design or introduction of a new hardware feature). Think not only of people upgrading for the sake of it, but all of the cases, screen protectors, and other assorted accessories cast in plastic for previous models that are garbage now.
It would be nice if we could just space these things out to 5 years or so now, because that's probably how long it takes for anything to change enough to justify a new model.
Of course when it was new the camera opened quickly. And then Apple made their OS more heavy weight every year until my phone slowed to a crawl.
And faster phones are nice, but I think it is worth considering how valuable that really is to us as users and a society, especially if the process involves making loads and loads of ewaste and consuming tons of new resources, and all the emissions their mining and transport involves, when we could simply keep our software slim and our old devices functional.
And the big companies will never do this. Do we need to force them to allow open software to run on these devices, so that clean builds can be patched and maintained when the company over bloats them or abandons them?
The things you listed (camera and chip speed) are basically the only things left that these companies can claim is better than last year's model, but only because it's so easy to use synthetic benchmarks and numbers that mean nothing to make them sound like a dramatic improvement despite the fact that we've reached the bottom of the barrel in terms of diminishing returns on the user experience for smartphones in their current form. More megapixels don't matter anymore, CPUs are hardly a limiting factor and yearly gains on their performance are marginal at best, and we have more than enough RAM for pretty much all use cases.
My point is that if these companies insist on re-releasing the same phone every time, maybe they could space it out a little.
It makes perfect sense that five non-significant changes can add up to a significant change.
So the suggestion would be 1 or 2 releases instead, after more of the changes build up, instead of 5 releases.
Personally I think yearly is fine for manufacturers that only have a couple models. But they need to actually support things for a reasonable lifetime, and should be mocked for having frequent releases if they don't have a good support lifetime.