Well there are semantic web concepts, and there are concepts like a searchable information graph. I read a really good article on this site a few days ago laying out a pretty good architecture for searching a distributed knowledge graph.

The core problem is a disconnect between the knowledge available on the internet and search tools. Indexing is a stop gap to enable the web to be usable without solving this problem, but it is becoming untenable and was always known to be non viable, just treated as a problem for another day. The sheer amount of information online and its disorganization, as well as the misaligned incentives of search tools are beginning to make this once far away manageable problem a day to day reality for all of us, the solution requires an architectural change to the way information is stored and accessed and updated online.

Do you have any idea of what that solution might look like? The existence of a problem doesn't mean that there necessarily exists a better alternative.