Definitely agree here, want to add a link to this paper which shows how far we still have to go (and questions whether the current models will ever replicate human vision): http://arxiv.org/abs/1412.1897