That would be super-interesting to us, an issue we have been struggling with is filtering/summarizing content from sources with lots of content such as RSS for sites like The Verge.