Somehow flat images allows the modellers brain to see mistakes/omissions that we don’t see when we have the 3D-object in front of us.
Maybe the 2D-photo somehow prevents us from seeing our mental image of the object. I heard about some graphics artist who checks his/her work by viewing them upside down (standing the images on their head sort of), The artist said that the upside down view allows the brain to look for blemishes instead of interpreting the coloured surfaces as a face/car/train/whatever object it shows.
Edit: Wow, this is embarrassing.
I thought I responded to a reply to a post I made in another topic:
The comments above about photos and their usefulness when checking for assembly or finishing mistakes isn’t totally off the mark here but they make a lot more sense in the other topic