CNET también está disponible en español.

Ir a español

Don't show this again


Social news aggregator Tiinker uses the power of suggestion

Can machine learning for news have a human side? Find out with Tiinker, a machine learning based news aggregator.

Far separate from RSS readers lays the land of news aggregators. One of the more well-known ones is Google News, but there are also social solutions like Newsvine that let the community decide what news items rise to the top based on what's coming over the wire. A new service named Tiinker (that's somewhere in the middle) opened its doors earlier this week. The site's been in private beta since late last year, and can most easily be described as a mix between Google News and StumbleUpon. It's not meant to replace a standard news page by any means, but instead is designed to let you whittle down what types of stories you like in order to get future recommendations that are similar. The ones you mark as not liking, Tiinker will simply get rid of.

Everything on the site is handled with three basic controls, a thumbs up and down button, and a way to bookmark content that goes into a separate feed that you can share with others. The service will also keep track of which stories you've marked as liking, so you can go back and read them later. In order to abate users limiting themselves to just a few types of content, there's also a "lucky" dip section, which chooses a story that's outside of your taste.

The ultimate goal of this and services like Spotback are to help you get a news feed that's been hand-tailored to your preferences so that the only stories you'll see are ones you're interested in. Of course what makes that part critical is the source list, which is where Tiinker's a little short-sighted. There are many good sources that articles seem to pop up from, but there's no way to go in and "tinker" the complete list to see what's on it or add new, independent sources. This is mostly because of the machine learning system that's set up to balance the tastes of the entire user base, but it's a bummer when you're in the dark about where the content is coming from. Interestingly enough, up until last month the service had a built-in RSS reader that let you do such a thing, but it was pulled down for retooling.

I like the idea of Tiinker, although it's not useful to me for news or content suggestions as I'm an RSS enthusiast who reads a lot of social news sites. It's definitely a handy way to get a smattering of new feeds and story suggestions, and the machine learning concept shows promise, but until they give a little more control over where the content is coming from, advanced users will likely want to stick to a social news service or feed aggregator like Feed Each Other, and still private Streamy to get interesting news links from humans.

Tiinker pulls in news from a variety of sources. Pick the ones you like, and it will do its best to find you more that are similar. CNET Networks