Saturday, December 11, 2004

Turning noise to knowledge

A few days ago, Findory started personalizing the recent articles on our source pages.

For example, when I go to Wired magazine on Findory, because of my reading history, two articles are marked as personalized, "Troops stay in touch on the internet" and "Yahoo searches desktops too". When I go to Scobleizer on Findory, four articles are highlighted for me.

The problem with current web feed readers is that they don't solve the information overload problem. Sure, I can pick and choose which RSS feeds I subscribe to. But, once you have tens of subscribed feeds, reading them becomes this cumbersome process. Click on a feed, skim the articles. Anything interesting in that one? No. Click, skim. Click, skim. Click, skim. Ugh.

With Findory, the important news bubbles to the top. On the home page, interesting articles are selected just for you, pulled from thousands of news and blogs. On a search, relevant articles are highlighted. When you read a blog on Findory, important posts are highlighted.

Current RSS readers merely reformat XML for display. That isn't enough. They need to filter and prioritize. Show me what matters. Help me find what I need. Next-generation RSS readers will be personalized.

This is about more than just reading news. This is about information. Where before there was an undifferentiated glut of information, now there is focus. Where before there was noise, now there is knowledge.

What will this future look like? Findory has taken the first steps. Come and take a look.


Seun Osewa said...

I really do believe the ranking algorithms still need a LOT(!) of fundamental work before they come close to achieving the promise that Findory implicitly makes. Really. This sort of personalization is hard, I guess, but since you're promising it we the users need to hold you up to it!

Greg Linden said...

You're right, Seun. This is hard AI. Truly solving the information overload problem will take many, many years.

Findory has taken some first steps toward the goal. I hope you agree that those first steps show promise.

Seun Osewa said...

Of course it shows promise, but it doesn't appear to have become more useful than it was when it started. Other features you've added, such as source pages, are nice, but the core feature - selection AI - really doesn't appear to be improving. The 'users who read this article also read this...' approach is probably not very meaningful for news. I think you need to focus more on analysis of the articles' content.

More, more, more! I really love the concept but it needs to be made to really work!

Greg Linden said...

Interesting feedback, Seun. We have been modifying the core algorithms, but the changes are subtle and difficult to see from the outside, mostly adapting to rapidly increasing data as traffic floods in. As we experiment and learn, there will be more changes.

Our pace of innovation is rapid. We continue to refine our core algorithms, and we are applying our personalization algorithms to new areas such as personalized web, news, and blog search.