The "diffraction limit" used to be thought of as a fundamental barrier: you couldn't do anything with light that involved distances smaller than half a wavelength. Imagine that you're jumping rope while dancing around and using the rope's impact on the ground to sweep patterns in the dust. (Ignore your footprints; the rope is what's important here.) By just spinning the rope around yourself, you can't make patterns that are much narrower than you are.
But if you can shake the rope in intricate, carefully controlled patterns instead of just swinging it around, you can make it touch the ground in smaller and more controlled areas. Similarly, if you send the light through very carefully calculated masks, you can make the energy -- over a very short distance -- take on patterns that are quite a lot more intricate than a simple wave of light.
Another rule has been broken by this work -- this one not a "rule" of physics, but a human prediction. As the article explains, evanescent wave lithography wasn't expected to be developed for another five years. Technology seems to have a habit of doing that, these days.
So how long until we see the first positionally controlled, atomically precise diamond fabrication? Anyone want to start a betting pool?