I'm always extremely impressed with NYTimes' Developer Tools (<a href="http://developer.nytimes.com/" rel="nofollow">http://developer.nytimes.com/</a>) – especially when they are going through the (apparent) decline of newspapers.<p>I'm expecting a lot of interesting stuff to be built with this new API. Looks like lots and lots of search points ("An article comprises ~35 searchable fields")
What impresses me most about this API is the richness of the data. Not just the articles, but publications times and keywords, thumbnails and more. A rich API just begs for rich applications -- say, the frequency of certain phrases graphed over time to show the rise and fall of cliches like "information superhighway" and "uniter, not a divider". Cool!