Right. The real spam problems will surface further on when AI bots will make procedurally generated content. Then web-of-trust scores may become useful.