Lots of people don't believe it. Then again, lots of people haven't spent much time blending the types of mixes where it's most commonly seen.
Still, seeing is believeing. It's easy to show that a mix with a large volume of, say, oxygen, topped with a small volume of another gas/mix (like air) added very slowly does not mix by itself instantly. It's also easy to compare the results obtained by letting the mix sit for different periods of time to the results you get when you immediately agitate the tank and analyze. Admitedly you aren't likely to see this when blending mixes like 32% or 36%. Try blending some 80% as I described (add the air at about 60 psi/ minute or so) and let me know what you see...and no I don't dive 80% these days and there's no blending issues with streight O2. LOL
It's a simple enough experiment. Try it...oh, just use something common like an AL 80. There's no need to confuse things with tricky geometries.