The rolling/incremental release is something that Apple does, but it's not like Apple is the only company doing it. Several Linux distros follow the same release model, and other software (not just OSes) does it too - MATLAB switched to this style model years ago, and it is nice at work to have something that is never more than 6 months out of date.
The only downside is that you have to really plan out deprecating features. It's easy to drop features when going from major release to major release, but in a rolling update, you need to give developers a heads up that things are about to change and give them time to fix things in their code.
We dropped MATLAB because the communal version refuses to work without internet connection (standard in the field) and we cannot afford a licence for all field laptops, where they will be used very little (but still to do things that need to be done).
Python, mainly. Some people still use older R scripts. It's kind of hard to make people stick to the same language when there is no real common code base (we're scientists).
The only downside is that you have to really plan out deprecating features. It's easy to drop features when going from major release to major release, but in a rolling update, you need to give developers a heads up that things are about to change and give them time to fix things in their code.