I've put together a new site (NerdlyNews) that displays techie and nerdy news from sites that I like. On the back-end, I've built a news gathering and classification system that uses Bayesian Probability to publish only those articles that are likely to be important or interesting. It's working pretty well already, but I need to do some tuning of the classifications. I want to publish fewer but more interesting articles. Check it out at NerdlyNews.