Nitpick: in the daytime, rods do not contribute significantly to vision, as the rods are completely inhibited by the brightness. (Somewhat un-intuitively, retinal receptors are inhibited by light, they fire less and less, the brighter the light.) Furthermore, more receptors does not necessarily translate to greater visual acuity. In the case of rods, the greater number of cells allows larger “sensor sizes”, with the output of several rods being summed before leaving the retina. Overall, rods have much lower spatial acuity than cones. Even with cones, the signal that goes out of the eye goes through spatial filters that involve several neighbouring cells.
That’s the common view, but there is also a hypothesis, championed by Mark Changizi, that the purpose of the third cone is to allow primates to better read each others’ skin. The argument goes that dichromats that depend on fruit don’t have that much problem finding them so the evolutionary pressure isn’t huge there. Trichromat primates all have naked skin visible on their faces, whereas dichromats do not. The peak spectral sensitivity of the L (red) and M (green) cone perfectly match a spectral peak and trough of skin spectral reflectance that are due to oxygenated hemoglobin. In other words, primates who have L and M cones are able to tell the level of oxygenation of others’ capillaries. With a naked face, we’re able to tell when others blush, are angry, or sick. For social animals that need to intensively care for their newborns, this is a great advantage. That it also allows you to better judge the ripeness of fruit is a secondary advantage. Or so goes the argument. (Changizi has a bad tendency to overstate his case in his popular writing.)