Imagine you're building the perfect internet indexer. In an effort to generalize your model, you anticipate a scenario where a nofollow link shows up consistently on various Wikipedia pages where it persists for a long time. What are you going to do? Ignore the link because, sorry, them's the rules?
Or are you going to realize that nofollow is about as useful of a signal as <meta name="keywords">?
Granted, I never worked in SEO directly.
I'm still not sure what the truth is. You raise a good point.
Google introduced nofollow in 2005 to either let people help fill a gaping hole in Google's tech at the time or for the smart marketing move of making people fixate on their magic attribute. Both under the guise that spammers would care.
It's been twelve years.
The SEO advice industry and Matt Cutts' blog these days are probably like when I try to ponder why my cat scoots its waterdish around the room before it drinks: attempting to find backsplanations for the idiosyncrasies of a neural network that nobody quite understands.
That said, Google probably has some heuristic for considering them. It probably errs on the side of ignoring, but it would be stupid for Google to ignore all nofollow links. Even more so because spammers don't like them.
My anecdata concludes exactly the same.
A link is a link. Sure Google acknowledges that a link is nofollow but it's not going to ignore references from a giant of the web like Wikipedia just because it adds nofollow.
Not to be taken lightly in light of all the mod craziness during the Orlando shootings and related posts.
You say that like that isn't exactly what has happened with all image hosts so far.
What history has taught us from these incidents, is that when old memes and similar posts go missing, the world wide web as a whole seems to cope just fine.