In an answer to a previous question I asked, where the Kahler differentials of a variety over a finite field were discussed, it was stated that:

You can certainly define de Rham cohomology using Kähler differentials, but over a field of characteristic p>0 . . . it is somewhat pathological: the Poincaré lemma can fail etc

(1) What does this "etc" mean here? That is to say, what else goes wrong?

(2) Is the de Rham complex for Kähler differentials useful for anything? Or is it just a curiosity?