Advertisement

Firefox experiment suggests articles based on your web history

A dash of machine learning could help you discover new stories.

You can already save articles for later viewing in Firefox thanks to Pocket integration. That's not going to help you find articles, however, and it's all too tempting to stick to the same old sites for your reading habits. Mozilla wants you to be a little more adventurous. It's introducing a new Test Pilot experiment, Advance, that recommends stories based on both your current page and recent web history. The extension uses machine learning technology from Laserlike to identify common traits in the trusted sites you visit and offer new pieces in a sidebar. If you've been reading many football articles, for instance, Firefox may suggest a piece about a recent player trade.

You can flag and block articles you don't want, so you'll ideally refine your results over time to get only the content you'd actually care to see. And Mozilla, the privacy advocate that it is, knows that the recommendation system may make some people uncomfortable. On top of the warning that the extension will send browsing activity to Laserlike, it's promising controls to both halt the experiment as well as review the collected data and delete it on Laserlike's side. Mozilla said it would try out "different methods" of supplying recommendations if there's enough interest, so you might see alternatives if this approach makes you uneasy.

You can install Advance today. This probably won't pull you away from Chrome or another browser if you're fiercely loyal, but it does show a slight shift in strategy for Mozilla. Where past Test Pilots have tended to focus on raw browser functionality, this is more about keeping the actual content fresh. It might work. Many browsers already have good-enough performance and features for typical browsing -- few of them invite you to explore.