While listening to some talks at the special session on categorical topology last Saturday, I realized that I had said something wrong at the end of part II: the filter spaces satisfying the property that any intersection of filters that converge to a point x must also converge to x are not the topological filter spaces, but the pretopological filter spaces. See the full post for an in-depth discussion of the matter.