Lately I've been noticing that when people need to use a spherical object, instead of going Shift A -> Mesh -> Sphere, they advise adding a cube and then applying a Subsurface Modifier to it, without any real explanation as to why. This is an especially prominent technique when the object needs to be low-poly, like for a particle system or something.
So my question is, why do people do this? What advantages does this offer over using a sphere itself? Wouldn't it be more useful to just add a spere? Even if it needs to be low-poly, just adjust the resolution, right?