> Dunning-Kruger applies to people who don't know a specific domain.

I mean, its largely a statistical artifact around which a pop science myth has accumulated, but on its own terms it applies smoothly and continuously across the entire range of ability in a domain, not in any special way just one one side of binary knowledge dividing line (the finding was basically that people across the whole range of ability to tend to rate their own relative ability closer to the 70th percentile than it actually is, but have monotonically increasing with actual relative ability.)