> One of the biggest challenges modern OS developers face (Apple and Microsoft namely) is how to make the new version of the OS appealing enough for users to move to.
More fundamentally, how do they convince users to continue spending money?
> A lot of people don’t care about the carrots though. Their computer does everything they need it to and does so reliably. In a lot of ways the OS is a solved problem, but the companies who make them still need to pay the bills for all the security work that will always need to be done.
Exactly. Consumer desktop/laptop operating systems (Linux/MacOS/Windows) have been largely stagnant for years, at least in terms of offering meaningfully better ways to get things done[1]. That's a problem for the commercial offerings: why would anyone buy the new version if it doesn't do anything new/better than its predecessor?
Microsoft's path is pretty clear: the OS is evolving into an internet terminal. Consumers don't pay for the OS: they pay a subscription to do anything, with features increasingly delivered online and, crucially, all their data "managed" (aka controlled by) by Microsoft too. Device sales don't matter first order in that model: it's a dumb terminal. But Microsoft has to keep the device manufacturers sweet: it's imperative consumers use Microsoft's dumb terminal, because it's connected to Microsoft's services. It would be a disaster if HP, Dell et al decided to use some other front end. Microsoft will do whatever it takes to remain the dominant desktop.
Apple has more of a dilemma. It's always been about the integrated experience, and hardware sales are a first-order driver of revenue and profitability. So it needs to keep consumers buying hardware as well as trying to up its rental offerings (iCloud etc). Moving to its own silicon has powered that for the last few years (more supply chain control, better performance, and good marketing on why it's better).
>It makes me wonder if Apple (for example) put macOS into a no-new-features mode and just improved security from here-on-out what the long term effects would be for users and the industry.
That's arguably what MacOS has been for the last few years: security updates plus support for their own silicon.
Beyond the hype, LLMs possibly offer an opportunity for a notable change in desktop interaction (note 'change' not necessarily 'better' - I'm withholding judgement on that). An OS that could respond to users' queries/instructions directly could be better than having to click about semi-randomly to find the right menu option.
[1] which doesn't detract from e.g. security upgrades, different peripheral connection standards, hardware improvements, etc.