If that is all it is, I don't see how it could be patented.
Maybe it is just software that turns the drawing of the desired final product (3D object) into some sort of negative that becomes a mold. But, a good mold is not just a solid with an inside surface that is the outside of a 3D object. It needs some thinking about object removal, flash minimization, venting for void elimination, etc. And, what about all those surface smoothness issues we have on 3D prints, already - how does this technique deal with them?
The article seemed a little strange for Science News to publish in the first place. It has no info about what was really accomplished by the author, just hype. That is why I asked here if anybody has any info.