Interesting. I like the idea of currently/today's/yesterday's news. The scores, though, are rather buggy. On the home page right now a Forbes article about App

Interesting. I like the idea of currently/today's/yesterday's news. The scores, though, are rather buggy. On the home page right now a Forbes article about Apple's headphone jack is listed twice in the top right with two different scores (8,046 and 7,908). On the CNBC subpage, Marc Andreessen's interview in which he called Snowden a traitor is listed three times, with three different scores. All are probably very low -- the article was well-discussed online, has 300+ comments, and the discussion topped the HN home page yesterday for a while: http://newsscale.com/site/cnbc.com http://newsscale.com/site/cnbc.com I'd probably narrow categories considerably. Two of the top three articles in "entertainment" deal with random human interest stories (a sick 4-year old and a puppy video). That's what I'd call "cute" or something, but not what most people would think of in an entertainment category. Similarly, in business the top articles include student debt (should be in personal finance), landlords and dogs (should be in real estate or pets), an Obama story (should be in politics), a musician obituary (should be in music or entertainment), and guns (politics or law enforcement). The problem is that solving those problems and taking Newscale to the next level requires rather more work than merely crawling RSS feeds and making some calls to Facebook's graph API to get sharing counts. :) I've looked into similar problems while working on creating a recommendation engine and iOS/Android app for personalized news (I quit CBS to found http://recent.io http://recent.io this year) and it gets a bit more difficult from here. Happy to chat offline if you like. I think my email address is in my profile.