When you design a UX, you try to guide the user through your machine’s interface in the most obvious and predictable way.
Most UX/UI must make assumptions on what the user knows and what she doesn’t know. It’s a tradeoff between beginners and experts.
There are, however, different levels of preexisting knowledge you can expect within the same UX. Take a mouse. You don’t teach the user how to use a mouse, you expect her to be able. You take a collective experience with mice as granted. You must however consider that your user might not know how to, say, enter a formula in a spreadsheet.
Mobile is driving a whole new set of interaction paradigms, and UX on mobile is very specific (for many reasons and because of several constraints).
Even on mobile, however, there are things you don’t teach anymore to the user. Touch interactions are considered well-understood. We could argue there is a collective knowledge, a collective experience of touch interactions.
It is true. The problem is, while the vast majority of mobile/smartphone users can make use of touch interactions, others cannot, because they’ve never been exposed to them.
This causes as a vicious circle because many of those users will find touch interactions strange, and cumbersome, not because they are, but because your UX assumes they already know the basics, cutting away hints/suggestions. That will make adoption of touch interactions always difficult for beginners.
A perfect example is standard Android navigation patterns.
If there’s no tutorial, how is the user supposed to know what she should do to open the sidebar menu (or whatever it’s called)?
All that said, I think what is happening in the mobile field is just fine. Given the numbers involved, annoying the majority of users to benefit the minority would not be very smart.