Another attempt to help all those who think less MP means better IQ.
Sorry that I'm not a teacher at all, and for not writing idiomatic english.
Photons are discrete quantities of light, each photon with a specific wavelength or if you like, single colors from the rainbow spectrum. The ideal sensor would not use pixels at all. It would record the exact position of arrival and colour for each photon, within the exposure time.
View the perfect image on a large enough monitor, and you will only see discrete color dots distributed on a black surface. A clean, white surface won't even look white. It will look as a chaos of intense color dots with different colors.
Move away to the correct viewing distance, and the image on the monitor will appear to the eye just as the natural scene does. The chaos of color dots will blend into a perfectly clean, white surface.
The smaller the pixels are, the better is the approximation to the ideal sensor.