1) Customers would need 3D design experience, or Amazon would need to come up with a truly novel CAD type program that the general populace could use. Even products like Google's SketchUp aren't exactly easy to use once you get past drawing cubes.
2) I think most people don't yet realize the limitations of this technology. They're going to be expecting super detailed minifigs and nice polished paint jobs, not rough plastic that doesn't have a great amount of detail in it. There's also the problem with designs that are impossible for current 3D printers to create. In both cases injection molding is still king, IMHO.
I think both of these will eventually get solved, but I don't think it will be this year. 3D printing is progressing at an awesome pace, but it still has to bake a bit before it gets as good as other creation methods. And most people are going to be expecting something better than what consumer grade 3D printing can currently offer. That means lots of returns for a site like Amazon.
You can get your own designs printed, or print other people's designs.
Or if you must insert friction for legal reasons, simply download from thingiverse and upload to Amazon. Two steps not too awful.
It actually seems like yet another uninteresting store for 3D printed tat. I don't see how it fills a niche that isn't already filled by, say, Etsy, or how it's really any different from the existing Amazon marketplace other than "Look! 3D printing: New! Trendy!"
What makes 3D printing interesting is the ability to create things that could not have been created before. If you are selling just those things (and not a design or custom production service), it's more appropriate to categorise them according to what they are, not how they were made.