The argument in the comic is that mostly-colorblind octopuses know what camouflage color to use based on light polarity.
For myself, as an outspoken blowhard somewhere on the ignorant end of the Dunning-Kruger curve, I of course have a different opinion: Octopuses know what camouflage color to use primarily by taste. Evidence: They can taste with the suckers on their legs, so they know what surface they're trying to emulate, but their eyes are on top of their heads, so they can't always 'see' it. Ergo: they are tasting the surface to know what colors to use.
I'm not suggesting that their suckers can actually see colors, but that they know what, say, a barnacle tastes like, and through evolution they can now change their skin to 'barnacle color'. So it doesn't matter if they can see various wavelengths or not.
This might be easily tested by dyeing a rock some weird color and seeing if an octopus changes to that color, or changes to the natural color of the rock.
- Comic about octopus camouflage, by Audra McNamee
