It's really a zombie/cockroach meme (to use Paul Krugman's terminology) - it's dead, but keeps coming back, just like the Eskimos and their 100 words for snow etc.
But isn't color names one of the last refuges of Sapir-Whorf, one of the places where the weakest form of Sapir-Whorf might be having some sort of effect, because color names are one tool (out of many) that we can use to consciously distinguish colors?
As an example, these Munsell tiles are supposed to be perceptually uniform, so just based on the mechanisms of vision, each tile would be as easy to distinguish from its neighbors as all the others. If you showed me tile A5, then later showed me C5 and asked if it's the same color, I would know it isn't. One is red and the other is orange.
But I'd have a much harder time with K3 and M3, or J6 and L6 -- and these are colors that the article points out are harder to distinguish in words.
http://www1.icsi.berkeley.edu/wcs/images/jrus-20100531/wcs-c...
You can’t get a complete idea of what the original color chips looked like, as quite a few of them are outside the sRGB gamut (so what you get in this picture is a less colorful representation of the same lightness/hue). But it should give you a reasonable impression.
Here’s a more continuous picture of the full sRGB gamut in Munsell space, along with some points of interest:
https://upload.wikimedia.org/wikipedia/en/e/ed/Color-map-2.p...
(For example you can see that the ISCC–NBS labels for sRGB primaries are “reddish orange”, “greenish yellow”, and “purplish blue”, respectively)
If you actually look at a chart where a bunch of Munsell color chips of different hues are all presented at the same lightness (value) and chroma, and you have standard color vision, then they should look pretty well equally spaced. (Or the same story for holding any pair of attributes constant and varying the third.)
One significant problem with trying to glance at and interpret these kinds of more limited hue–value charts (i.e. not the full set of Munsell chips) is that the chroma varies from one part of the chart to another, and chroma differences are often more salient than hue differences.
But some named categories are certainly larger and fuzzier than others.
Yes, if I knew I were being tested on this, I could invent a description to try to remember what K3 looks like. But I don't have to invent anything to recognize the difference between red and orange.
But maybe jacobolus is right that this is just an effect of RGB monitors and not of language. These tiles are in the cyan range, which RGB screens don't display faithfully, so maybe the actual difference between the tiles is more obvious when you're looking at them.