It's simply about physics: tiny pixels = less light per pixel, less DR because of their small capacity, lower signal-to-noise ratio, and added to this you unfortunately need faster shutter speeds to freeze motion on the pixel level. You also lose more actively photon collecting area in total on the sensor, because each pixel adds some dead zones due to its electronics etc. Another fact based on physics is diffraction blur, so the optimum aperture is smaller with smaller pixels if you want to convert their full potential into an image. You lose then depth of field, which e.g. can be limiting in classic landscape/cityscape photography. This is no argument against high MP FF cameras: they CAN make sense, if you shoot with a lot of light and can live with a smaller DoF, then you have an additional advantage that you can crop massively if you need. In fact, I use a 7D2 for wildlife which equals a 45 MP FF sensor, but this crop camera really needs much light (ok, it is old technology, too).
As a science journalist I am currently dealing with smartphone camera technology in depth and therefore I am extremely aware about the physical limits of small pixels. In smartphones, a lot of algorithms have to turn a huge loss of real image information into heavily post-processed images. One of the latest trends is adding artificial image content from data bases to sort of fill-in losses. Such tweaked images are not my personal idea about photography, I have to say.