The real question, IMO, is why the need to release a new model every year? I think a lot of the churn of both hardware and software had nothing to do with technological progress, and everything to do with generating sales.
If you consider one person in isolation it sure does not make sense to release a new phone every year.
If you consider all users though, they most certainly do not buy/upgrade at the same time. So let's say a new generation is released every three years, for a given user when buy-time comes they'd be buying a brand new phone that is 1.5 years old tech on average, 2.9 worst case. Releasing a new phone each year means there is the opportunity to buy the up-to-date technology (whether it's a flagship or an economic version) at any particular point in time.
Opportunity does not mean necessity. By and large once people buy a phone and are happy with it they don't change it til it's dead or slow. Most I've seen keep it even way beyond the major version support threshold, or even security update threshold (which can happen faster than expected when people buy refurbs or older models, which they do). They're appliances. There are new models of cars or TVs or washing machines every year, yet virtually nobody changes those every year. One just buys the best one within their budget x identified use case at a given point in time.
In my experience the persona that buys a new phone every year used to exist in the early smartphone days, but today is by and large an outlier figure.
Not that mobile carriers like that, instead they try to keep the churn alive against the tide, because they want customers locked in so they have you buy carrier-locked phones at a rebate under a 1y contract and dearly want you to renew at the end, which means enticing you to get a new, carrier-locked phone.
The only ones I see changing phones regularly are the ones buying phones as consumables instead of appliances - less "washing machine", more "toilet paper", the kind of phone that is cheap and underpowered, borderlining on unusable (or outright so, which I consider to be defective by design) e-waste right out of the assembly line. They don't care about the device, so they don't pay attention, so they break it every rand(0..6) mo, and then the cost to repair is barely lower than the price new plus "it was so slow anyway", so they buy a new cheap one and the circle starts again; which doesn't quite work out financially of course, they'd be better off buying a good phone and paying attention to it (which they would simply because it's not cheap). But back to the point, this kind of persona doesn't care at all about the device "newness".
Well, it does obviously have to do with generating sales. But to play devil's advocate, I never have enough battery life, even with a fresh battery. There are incremental improvements, year-over-year, in terms of processor efficiency. I take a lot of photos, and I definitely notice having a 5x optical zoom now. There are details that some people would say are worth upgrading. But it is not the case that all people fit this template.
> Well, it does obviously have to do with generating sales.
I don't think it's that obvious, and I don't think manufacturers count on it ("it" being short length updates)
Instead I think it's about handling competition: say a manufacturer updates their product line every 3 years, should a user that kept their phone for a while need to buy in the middle to last bout, they'll look at the market and see "oh that other manufacturer has released an update to their product line a month ago, so I'm going to buy that instead of a 2yo product from my current manufacturer". The comparison may be done on any kind of metric but it ultimately comes down to what's state of the art at a given point in time.
It doesn't mean the user hasn't kept their previous phone a long time. But to be competitive across the time continuum the product line needs to be fresh enough.
In that sense it generates sale, or rather, helps with retention, not by having one update every year or other year, but by creating the opportunity to stay within one's current brand.
I think leading manufacturers have realised a long time ago that satisfied customers are what generates reliable sales and growth long term. Even Samsung, king of the shiny "innovative"-but-useless bullet-point feature list have somewhat gotten their act together.
I don't think it's unreasonable to release new hardware every year. Apple is iterating on CPUs on an annual basis, and the new CPUs make for a better phone. It probably doesn't often make sense for a user to upgrade a single step (anymore), but if you're buying, the newer one has a better cpu, so it's worth considering.
That'a separate from software churn though. Desktop OSes tend to support hardware for much longer than mobile OSes. Apple does better than Android here, but they both should do better. My perception is that most android apps support much older versions of Android than iOS apps support of iOS, but I don't know how that ends up looking for age of device supported.
I'd say that is correct. In a sense it's like part availability for cars or washing machines, and is similarly quite an outrageous situation too: around here law mandates 10y of parts availability; trouble is, reliability has progressively improved to the point that this becomes ludicrously insufficient.
An example: I bought a car five years ago, it's well tended and shows no sign of being anywhere close to the scrape yard five years from now, yet parts will become hard to come by; case in point my other car is turning 16 (sixteen!) and barring the need for a full body paint job because the varnish is gradually peeling away (which is kinda expected), is equally in such good order as to most probably be in similar condition for another 10-15 years with proper maintenance; but will I be able to? Parts are getting rarer as they've been in a stock-remainder only basis for five years.
Another example: my washing machine had a part fail at the 7 year mark. Fine, parts are supposed to be available, so just repair? Well, the part moved out of production to stock-basis because they evaluated statistically that they have enough stock to cover the 10y mark. Problem: the stock is a sort of archival with glacial operation speeds, the time quoted for part shipping was shy of 6 months, and I couldn't find a replacement. So I bought a new machine out of sheer necessity.
10y is way too short for parts. Software, notably OS and its compatibility primitives for third party software, should be counted as parts. 7y for software is better than before but still way too short.
> Desktop OSes tend to support hardware for much longer than mobile OSes.
I feel like this is increasingly less true, at least to the point that the difference becomes immaterial to the problem.
Windows 11 needs an SSE4.2, which limits to Intel Nehalem and newer (2008+), although I don't know what revision Atoms are needed, and AMD Bulldozer and newer (2011+).
You'll most likely need to bypass the cpu model check, and the tpm check for an older machine, but otherwise should work. Windows 10 support's more hardware, of course. I forget where Linux cuts off now, I think it's somewhere between a 486 and a Pentium Pro/II. Although I'd expect you do need a 64-bit CPU to run modern software.
> I don't think it's unreasonable to release new hardware every year.
It wasn't ten years ago, but now they seem to be struggling to give you an actual reason for the new models to exist. The argument that not everyone upgrades at the same pace is not unreasonable, but still, it makes absolutely no difference if you get the iPhone 13, 14, 15 or 16. The number of people who needs the 16 for something in particular is so small that it wouldn't be a market in it's own right.
You say this like it’s a bad thing, and maybe it is, but it’s hardly a surprising thing.
Literally everything Apple does is entirely about generating sales. They’re not running a charity for the benefit of their fans. The only reason to come up with designs that people like is to sell them.
