I enjoyed reading this article that described in a clear and thorough way, that the road to developing an, at least partially, intelligent computer is long and winding.
It is too common to read articles that magnify little incremental improvements applicable only to specific context as major revolutionary milestones (does this ring a bell Google or IBM? :-))
There aren’t any real shortcuts on this matter. Even analysing billions of web pages, including wikipedia and other vast structured knowledge bases, is not enough to make a computer understand a simple phrase that even an eight year old could easily comprehend. As the author writes, it is necessary to go back to work on the core aspects of artificial intelligence, also through the semantic analysis of the text, which is the only way to understand the deeper meaning. This will require a lot of work, discipline and patience but it’s the only real way to develop computers that can be classified as, if not intelligent, at least quite clever.