The simple answer is that they do stick together, and in the case of nanoscopic crystals they may permanently bind. But for larger crystals they stick so weakly that even the gentlest breeze can pull them apart. So the question is: why do large crystals bind so weakly?
Let's start at the beginning.
The shape (or morphology) of a crystal that naturally grows can be characterised in two ways. One is in terms of the hypothetical ideal (so-called thermodynamic) morphology which is the structure with the lowest possible free energy. In the case of sodium chloride this is a cube with perfectly flat faces (facets).
In the real world, however, the rapid fluctuations that occur at the surface of a growing crystal can produce a radically different (so-called kinetic) morphology.
Sodium chloride does grow to form crystals that, overall, look pretty close to a cube. However, the surfaces are not perfectly faceted, they are extremely rough at the microscopic and nanoscopic scales - covered in terraces, steps and kinks (see the TLK model of crystal growth). So when two crystals come together, there are very few points of contact (and therefore binding):
It is also worth noting that ionic bonding is relatively weak (in comparison to covalent bonding).
In fact, here is an image of table salt under a microscope:
There are two important observations to make:
Notice how rough the surfaces are? Well they get even rougher as you zoom in!
In this particular photo there are two grains of salt (each about 0.1 mm in size) which are actually stuck together. How did this happen? The most likely mechanism (although there are others) is that when these two crystals were much smaller (nanoscopic in size) they bumped into each other. And because they were so small, they would have had relatively flat faces that could stick together. And then they grew, forever wed. If you're curious why they're aligned at a perfect 45 degrees to one another, see crystal twinning.