Third-party devs led the way, first with installable drivers that munged keyboard, disk, ... and an amazing task-switcher. But those became the bad old days of Macs: OS9 was falling apart due to overload. At huge expense -- Apple almost died from not tackling the issue well -- OSX started with a clean slate. Apple finally took control of the feature set that users wanted: pre-emptive multitasking, multi users, control of various daemons, etc.
Over time, that, too, got overlaid with all sorts of cruft. My upgrade this spring to a new MBP caused me a day's worth of debugging after I migrated files: some long-forgotten and actually no-longer-used third-party system extension conflicted with another module on the new hardware. Over a day's worth of debugging and cleanup, including an hour's worth of tech support (fortunately bundled with the hardware). Cost (very roughly) a thousand dollars.
So I think it's a bit disingenuous to advocate all these "neat" iOS hacks without acknowledging the possibly of really badly busting iOS, even for expert users, and "it just works" -- already a bit tenuous -- with it. Especially stuff running in non-user space can cause issues far-removed from the source of trouble. And unexpected software can either cause bugs or expose OS bugs that were otherwise benign or devs had worked around. A program that was tested only with Apple keyboard features can suddenly fail to work, tarnishing the app developer AND the platform. And like my incident, resolution can be very difficult and expensive for all concerned.
As my history indicates, I've been a big fan of hacks. But I already have seen enough issues with "good" apps such as the NYT crashing, or calendar entries failing to synch, that if I were still a developer, I would NOT want the added burden/expense of the sort of unstable OS that hacks have caused Apple in the past. Especially given the rather modest rewards for so many developers.