The real change is that up until now many icons got away with implying a mash-up of result, media and interaction in one clear glyph.
Even when the old-school icons were developed, the result for many things was a rectangular screen. Which is why they emphasized media or trappings, which had the useful side-effect of implying interaction. (bunny ears vs film strip vs VHS tape vs film camera)
As you correctly note, the problem now is that all media is digital. There are no distinctive trappings. There isn't anything to imply the interaction mode. (are you taking, browsing or sharing photos when you click on that picture icon? And when sharing, via what 'channel'? web service? email? mms?)
Further, we've never really had to worry about so many of these icons in a single app or on a single device before, because it was largely inconceivable that a single thing would do all of this to any acceptable degree. Even 10 years ago, what general purpose computing platform had all the features of a modern smartphone? Including access to multiple networks and services for transferring data at varying degrees of synchronicity -- think capture video vs broadcast video vs video chat. What app was commonly used by regular people to handle capture, editing, review and multiple sharing options for both photo and video?
A single glyph is simply not enough anymore. We can reasonably distinguish between text-in-a-box, still-in-a-box (cliche landscape) and video-in-a-box (stick figure with action lines). But we'll need multiple glyphs in combination to distinguish between video-to-be-streamed-down, video-to-be-broadcast, video-to-be-sent-via-email, video-to-be-captured, etc.
And this won't be too difficult. To some extent we're already doing this, with arrows and combo glyphs to distinguish between things like Save/Delete/Copy/Burn/Share.