This isn't correct either, a 32bit image with 32bit color depth would be 1 pixel in size.
Bit refers to color depth not image. For example a CMYK tiff would be 32bit color depth, not a 32bit image.
You're being a bit too literal with your interpretation of that section

So whilst you are technically correct, in common usage and language that distinction is rarely, if ever, made (by Blender users) so the general point is to simply make sure people check they're using "
RGB+A" images and not "
CMYK" (both often referred to as being "32bit images"). I've seen that mistake happen quite a few times hence it being mentioned.