The facial recognition is pretty good at identifying where faces are in pictures. It even found most of the
After using it for a while I've gotten better at training the facial recognition, but I think it would be even better with a few more features. The Confirm Name function is useful for confirming or rejecting its guesses, but it would be even nicer if you could correct its guesses in that interface (instead of just saying, "this is not Amy" or "this is Amy" you could say "this is not Amy, this is Rachel"). The closest thing I've found to this is going into the "Person may also be in the photos below section, opening the first picture and clicking on the Name button in the footer. That allows me to confirm the accurate guesses and correct the incorrect guesses. They keyboard support in this mode is pretty good too. You can tab from face to face. It took me a little while but eventually I figured out you can tab between the Confirm/Reject buttons on a Name guess too. If you are editing a name and tab, you will start editing the next name which saves on the keystrokes. Arrow keys can be used to switch between pictures. (I didn't ever notice you could drag select in the Confirm Name function, just heard about it online.) It'd also be nice if you could zoom into pictures while working on naming people.
The Places functionality is pretty interesting too, although it would work much better if any of my digital cameras (other than the iPhone) tagged pictures with their locations. There are some privacy concerns to consider too, although with the manual method it isn't too hard to just tag at the city or region level (which is what I did with a few of my New Zealand pictures when I wasn't really sure of the exact location).