It's not a question of "QR Code Readers" to me, but rather "semantic/contextual photo analysis" being applied to the default photo apps…
Basically, being able to:
- pick out text on signage,
- scan QR and barcodes,
- perform facial recognition,
- even identify specific location/orientation of the photo in real space (given enough information)
should all be included as defaults, and probably will be, eventually. We're just at the very leading edge of tricorder technology, but it's now at least possible to see how we get there.
As for timing, I'd say most of the above could be done today as third party apps, realtime processing of photos and video needs a little time to expand downmarket enough to make it "defaultable" across most consumer phones. Call it, conservatively, another 3-4 years from now.