<>
In this talk, I will try to paraphrase the goals of the nucleus centre (as I perceived them from the introductory talk) based on my own complementary experience of research on semantic search. I will try to sketch the design of a system that can enable the type of semantic search that was put forward in the funding pitch. I will then try to give an impression of why such a system does not already exist: for each component of the system, I will cover the main challenges, what is the state of the art (including work I've already been involved in), and what are the research questions to tackle going forward. Though I doubt we can make Google bankrupt by 2017, I do hope that the talk can help to identify important research questions in the area ... questions that hopefully match up with our joint expertise.