Yeah this what I immediately think too any time I see an article like this. Adjustments like contrast and saturation are plausible to show before/after, but before any sort of tone curve makes no sense unless you have some magic extreme HDR linear display technology (we don't). Putting linear data into 0-255 pixels which are interpreted as SRGB makes no sense whatsoever. You are basically viewing junk. It's not like that's what the camera actually "sees". The camera sees a similar scene to what we see with our eyes, although it natively stores and interprets it differently to how our brain does (i.e. linear vs perceptual).