Daniel Moore comments on the lack of progress in the Mind World Map project here. If you look at the map, you'll see that a reasonable map is created pretty quickly, but then the progress stops and it seems just random noise is added. I think there is a logical explanation. If 95% of the people click the right pixels and 5% click random pixels, you'll progress really fast towards a map of 90% accuracy, but the progress slows down and you'll never make 95%. 95% sounds good, but it means that 100 pixels or so are wrong. And 95% is probably rather high.
I have been thinking of optimizing the map, i.e. more majority voting and maybe than have the red pixel be chosen from the pixels with the greatest difference of opinion or one that borders on land and water. On the other hand, most of the 16000 clicks I got in two days after metafilter posted something about the project and ever since, progress has been slow. For now, I'm mostly curious whether the maps for different toplevel domains will start to grow apart.