Wavii delivers morsels of news, five words at a time
SAN FRANCISCO (Reuters) - Five years ago, Twitter's critics dismissed the idea that news could be transmitted in 140 characters. Now, Adrian Aoun thinks it can be done in just five or six words.
Aoun, a 29-year old entrepreneur in Seattle, unveiled on Wednesday an automated, algorithm-driven news aggregator called Wavii that strips articles down to its bare essentials -- who did what -- and presents them in a customized, Twitter-like feed.
In most news articles, Aoun said, "The 'what happened' and the analysis of what happened is coupled."
"I want to separate that a little," he said.
Coming at a time when readers increasingly consume news as morsel-sized commodities, the ambitious start-up has piqued enough interest in Silicon Valley to attract a quirky but well-known mix of angel investors, from Ron Conway to Mitch Kapor, the co-founder of Lotus Software, to Keith Rabois, Max Levchin and Scott Banister -- three entrepreneurs with deep ties to Paypal.
The app combs mainstream outlets and hunts for similar news stories on any topic selected by the user -- be it Rihanna or the Syrian uprising. It then crunches the similar stories through a "natural language processing" algorithm, spitting out a threadbare headline devoid of adjectives, context or detail. The resulting item is interactive and tagged by topic, so clicking on a story about a Justin Bieber sighting in Manhattan could, in theory, produce a list of where else Justin Bieber has been sighted recently, Aoun said. The item would also have a link back to a full-length news article.
In many ways, Aoun said he believes Wavii sits at the intersection of Google News, Twitter and Facebook. It scales the Web for news but also applies tags to people, events and actions in news developments, much like what Facebook does with friends' activities.
"We're making Facebook out of Google, and we're taking Google data and Facebook-izing it," he said.
Wavii would be more useful than Twitter in following specific topics based on a user's interests without having to follow multiple individual sources, said Banister, the investor in Wavii.
"The reason why we have journalism and news sources is because they cover everything, all sides of a story," said Banister. "I think that consumers will very much like the ability to follow real news -- the good, the bad and the ugly."
The project, which Aoun has been working on secretly for the past two years, has required significant engineering muscle.
He has amassed a team of 25 engineers, largely natural language processing academics and machine learning experts formerly with Amazon, to figure out how to parse through news stories to find relevant information to tag.
The field is also not altogether foreign to Aoun, a former software developer at Microsoft: His father, Joseph, the current president of Northeastern University, was a academic linguist who studied under Noam Chomsky at the Massachusetts Institute of Technology, he noted.
But he acknowledged the large scope of his ambition.
"We have to teach a computer to read language, which is astronomically difficult," Aoun said. "A lot of linguists have tried and failed."
(The story has been corrected in the first paragraph to change name of founder to Adrian Aoun )
(Reporting By Gerry Shih; Editing by Bernard Orr)