With the new infill technique coming up for the next CuraEngine release, there is a new infill method that makes triangles rather than squares. Lars' idea still holds, but it's something to keep in mind because that makes it a bit more complicated.
I think the idea holds some merit, from what I know of static physics. With the same infill density, it could be stronger against certain types of force, specifically a force that is spread out throughout the surface of the object. This is because two beams stuck together will be more resistant to shear and (by extension) bending, while still having the same resistance to compression individually. Of course, the larger gaps will make the centre of these gaps weaker, so it will be weaker against sharp, piercing forces.
This sort of thing requires a lot of testing, because I predict that the thicker ribs could be visible on the outside if the shell is thin, and the bigger gaps could lead to more 'rounding errors' due to the coarser discretisation of the infill (e.g. a small part could get 0% infill because it happens to fall completely within a gap). There is also the matter of exposing too many parameters to the user, though the next Cura release will streamline that somewhat better too.