If a language model or some other methodology can distill information from all of these sites and proxy to them in a way that a company can still promote advertising revenue, there should be a clear path away from blog spam. One could imagine it might even disincentivize copying original works.
There are so many paths for using information once it can be transformed reliably.
If natural language understanding continues to improve, it would seem that there is a very bright future for existing search engines.