Would you mind elaborating on how Pinterest derives value from your "aggregate information about links on the web"? What types of data do you/they glean through Instapaper?
free supervision of extractors for web page content (people complain if the scraper doesn't work)
some notion of link quality