Has Techmeme met its match? Or Flipboard? New semantic news analysis company Wavii is launching its automated, personalized news "feed" app today. It's a more modern and more pretty, deconstruction of the overwhelming news barrage we're all buried under. Wavii CEO Adrian Aoun's pitch: "Facebook is good for keeping up with friends, but it's kind of annoying that you can't keep up with your world this way."
Wavii is a smart timeline of the news. Its core value is that it runs its semantic engine against several hundred news feeds and is able to determine and display what stories are about.
It works surprisingly well. In some cases it will leave a solid headline alone on a story, but in others it will present a news item under its own headline that really cuts to the heart of the matter. These headlines aren't sexy or witty, but they are clear. Wavii also tags stories in different ways. Stories get headings like "Hired:," "Business Deal:," or "Criticism:." If the story is about a topic you've flagged, like "Related to Facebook," that's presented in a different way, but equally informative. Stories based less-than-perfect information gathering are flagged with little orange "R" for "rumor." Wavii itself comes up with these flags, including the "rumor" flag, by parsing story text. It is very cool.
Wavii de-duplicates multiple stories on the same topic. As on Techmeme, you can easily see other sources for a story. But you won't see them unless you click into an item on Wavii. Wavii differentiates between two articles that cover the same topic and two articles that are identical to each other (however, in my tests on the beta, sometimes a duplicate is represented as the originating story).
You train Wavii by interacting with stories: clicking on them to go to a summary or to the source, or flagging them with emoticons (happy, angry, etc). Wavii needs a better way to train it for stories that come through that you don't want to see, though. Aoun told me that's coming soon.
You can also have a discussion with other Wavii users under a news item, or post a comment from Wavii to Facebook or Twitter. That's not a breakout feature technologically, but it does round out the Wavii reading experience.
The service has a smooth setup routine. You select popular topics you're interested in, or you can enter your own topics or publications. The library of topics that reveal themselves when you search is fairly vast. About the only things that I was able to trip it up with were obscure writers and location names. You can also subscribe to content feeds from particular sources.
Wavii is based on strong semantic technology and actually has a direct if thin lineage to linguist Noam Chomsky: Aoun's father, Joseph Aoun was "Noam Chomsky's grad student," and himself a linguistics professor (before becoming president of Northeastern University), the younger Aoun told me.
Wavii's presentation of news is compelling. Of course, it's not perfect, and unfortunately in the category of products-that-read-the-news-for-you, even a small percentage of mis-hits have a big impact on day-to-day utility. I am concerned that users could miss key stories if they're lazy with their setup or don't take the time to signal the service by clicking on stories they're interested in. However, by following key people in an industry (just as we do now with Twitter), the important stories should come through.
I have been using the Web-based version of Wavii for a few days. It's been good enough to entice me back multiple times just to see what it's highlighting for me. It's not yet giving me a complete enough picture of the news so I can drop Techmeme or my feed readers, but there is real potential here. For monitoring particular companies or issues, it could be much better than Google e-mail alerts.
The technology could also be especially valuable for mobile users, where time and screen real estate is more precious; an iPhone version should be available at launch.
The weakest part of the Wavii equation is the revenue picture. It should make money from partnerships and "very targeted ads," Aoun says. If the technology proves to scale and work over time, it's more likely it's picked up by a media company before it breaks into the mainstream. See