While smartphone camera sensors have come a long way, lens technology did not progress as much. You still have a fixed focal length and quite small fixed aperture.

This means that your cannot control one of your 3 parameters of the exposure triangle, which is a big limitation. Also it leads to the user not being able to control the depth of field. A smartphone camera sensor is still too weak/small to offer good low light performance and your lens isn't sharp enough to allow for radical cropping.

Those are things that amateurs/non-enthusiasts don't realize but lead to any old 35mm film camera being much more capable in the hands of a skilled photographer. That being said, research on computational photography is currently making big steps and could narrow that gap quite a bit.