I have had to configure some relatively large forms in an application using JSON. Its not a particularly easy to read format, and when the nesting is deep its easy to loose a brace, or not notice a trailing comma in a list.
(I don't know if YAML solves all of those problems, but it certainly looks easier to read).
Just look at the YAML spec - http://www.yaml.org/spec/1.2/spec.html
compared to JSON - http://json.org/
and JSON is already everywhere (as a built-in library, a known mime type, etc). I have to go out of my way in order to use YAML.
I guess the other reason is that I never trust projects with names like "Yet Another..." because I feel like it conveys a lack of craftsmanship. Looking at that YAML spec again today, it looks like every feature was thrown into it with little to no discrimination.
Anyway, I didn't say that I don't use it because of the length of the spec. I said I don't use it because it's orders of magnitude more complex than JSON and I used the spec as an example of that. A bigger spec means a bigger parser and a higher probability that other people will not limit themselves to the parts that "you generally use".
To that, I added that JSON is everywhere already and YAML isn't. My point about the name seems a little silly, but my intuition has served me well over the years.