He looks at both the holy grail of classic A.I. applications, such as natural language understanding and what he calls the "genie in a bottle", and then proceeds to look at more realistic apps, such as semantic knowledge databases, semantic search, social graphs, and shortcuts.
On area that he misses is in the enterprise integration space and is sometimes called semantic mapping, ontology mapping, or semantic integration.
Other useful discussions of possible application areas and interesting adoptions are here:
- Semantic Web Adoption, W3C
- Semantic Web Application Areas, Division of Mathematics @ Free University Amsterdam and Oracle Corp
- Semantic Web Applications, MIT
So it seems that we still need to wait a bit for the killer app of the Semantic Web. One thing is clear, though: whoever creates that killer app will probably be using Altova SemanticWorks for their RDF and OWL editing, and ontology creation in the process. While the killer app is not here yet, the perfect developer tool to create that killer app already exists... :)