Getting the News — Evan Williams
blog.news.me
blog.news.me
When recommenders are applied to news you'll get more of the stuff you like, but you're losing out because you are not seeing the stuff you didn't know that you would like.
To avoid the problem of not being able to surface new topics because the user model has converged somewhere, you can use a bandit approach(seriously, who comes up with these names?). By that we mean that you do a bit of exploration by showing news topics drawn at random from some distribution and then adapt your model to what you learned. The naive implementation is "Every random(10,20) news items suggested by the recommender, display 1 drawn at random from the entire corpus. Adapt_Model(user_clicked)"
This is so true (for me at least). Especially when you reach that point that you actually checked some site only an hour or so earlier and know nothing new of interest with be there.
I can't believe the responsive web design trend hasn't done more to take advantage of gigantic screens. Everyone is pushing for smaller screens on mobile devices, while computer screens are way bigger than the 800-1100px width most sites are built for. Fullscreen apps on OSX are getting people used to the gigantic screen, but news in a 600px column doesn't really take advantage.
Pinterest does a good job with this, where they add columns as the browser gets bigger. Some may find it a little more cluttered, but at least it's using the space efficiently.
Nassim Taleb wrote about this too in Fooled By Randomness, worth a read:
http://www.curatedalpha.com/2011/nassim-taleb-minimal-exposu...