Apple's Siri is easily stumped. It's only pre-wired to answer certain kinds of questions. That's why Mike Tung created Diffbot, named as such because "it's a different kind of robot."
It's not a physical robot, but rather artificial intelligence that mobile app developers can use to look at a Web page and read it like a person, interpreting what's important and throwing away what's not. Diffbot cuts through the clutter to find the title, the author and the body of the story and then feeds it to a mobile app. It works on all different websites and it's not just for news.
"The origins of Diffbot started when I was a student at Stanford," Tung said. "Whenever a professor or a TA would upload new lecture slides or a new homework assignment, my phone would buzz. Diffbot would identify that content and send it to me on my phone."
Tung realized he had something and went to work on it at Stanford's StartX incubator at the AOL building in Palo Alto. Working out of the AOL building gave Diffbot a chance to connect with the people making AOL's new mobile app Editions and they ended up becoming one of Diffbot's first customers. Editions is a personalized news magazine based on what a person likes to read.
Automatically categorizing stories is one more part of the artificial intelligence that's just earned Diffbot $2 million in funding from big names, including the co-founders of EarthLink and Sun Microsystems.
Tung says he'll use that money to make Diffbot smarter and his customers happier.
"So as Diffbot grows and learns more about how to parse or how to get news from different sites, that's just another site that Editions can get news from," AOL product manager Ruby Anaya said.