Trapit, a discovery engine for Web content from the group behind Siri, launches to the public today. The company describes itself as “Pandora (NYSE: P) for everything else,” and believes Google (NSDQ: GOOG) searches and news delivery through social networks fail to offer users content that is actually important to them.
Trapit follows on Siri’s heels as the second company to launch out of SRI International’s $200 million CALO artificial intelligence project. It has been in private beta since June. Despite Trapit’s original financial support from SRI, it is now a standalone company and is raising funding. It closed a Series A funding round last month, led by Horizon Ventures, which has also funded Facebook, Spotify and Siri.
Users begin with a search, which can be a word, phrase or URL of a piece of content they like. They can save it as a “trap” and improve its recommendations by giving feedback–clicking a “thumbs-up” or “thumbs-down” and selecting the reason they don’t like a piece of content–and the results become better and more personalized over time. In addition, Trapit chief product officer and co-founder Hank Nothhaft says, the content that Trapit turns iup is more interesting and varied than content found on Google or through social networks. “Search is dominated by SEO, and social is the same set of 200 sources that dominate the Twittersphere,” he told me. “It’s very hard for a lot of the great content published on smaller blogs by professors, hobbyists and long-tailers to be found.” Trapit tracks around 100,000 sources, each ultimately vetted by a human analyst before being included. While initial discovery, content-gathering and filtering is automated, actual people check the most “highly probable” sources to make sure they are original and not aggregators. These employees are basically checking to make sure “the content is of minimal quality,” said Nothhaft in response to my concern that, say, a pro-Obama piece could be deleted by a Republican vettor. “We are just making sure this is not content farm crap you might find on Google.”
My first thought was that Trapit could help me track book publishing-related news and perhaps decrease my reliance on Google Reader, a product I have grown ever more frustrated with since its redesign last week. Since some people might not be as interested in finding new e-reader blogs as I am, I wondered how Trapit can be used by those who don’t track certain topics for a living. “We saw two very distinct dichotomies [in terms of possible Trapit users],” Notthaft said. “One was journalists and our tech-y geeky friends,” who track news with Google Reader, keyword alerts, Twitter lists and so on and “it’s basically a part-time job.” The other group was “avid Web users” who believe the Web has become too cluttered and feel search has disintegrated into “a directory of businesses and commerce.” Search engines bad at finding new content as it’s published, Notthaft says, and “people don’t want to go out and forage. Look at the rise of services like Pinterest. People want to log in and give an indication of what their interests are” and have content delivered to them.
Trapit hopes to differentiate itself from social news sites like Flipboard and Pulse. “There is a deluge of news applications on the iPad,” Notthaft said. “Each UI speaks to a different kind of user, but they are still really news readers. We didn’t want to be lumped into that category.” So Trapit is launching as a Web site, not as an iPad app. But its design is clearly tablet-influenced, and the company expects to release an iOS app in the first quarter of 2012, with Android to follow. The homepage includes trending news, as well as feeds for broad topics like entertainment, news and sports, so you don’t necessarily miss out on the zeitgeist if you’re spending a lot of time on the site.
Coming soon are some more social services: Users will be able to publicly share their “traps” and individual articles they find, and will be able to follow each other. But the site’s overall goal is personalized content delivery, and even if someone starts following someone else’s trap, it changes as soon as he or she interacts with the content. Users will also be able to divide their traps into groups starting in December (so that I do not have to see fun animal-related content mixed in with boring work-related content if I don’t want to).
The obvious question is how it works, and the answer is well. The first topic I tracked was “book publishing.” After a few minutes of personalization and thumbs-up and downing, I began seeing content much more geared to the things I write about–lots of Kindle Owners’ Lending Library news and reaction, for example, as well as some paidContent articles written by me, which are clearly good indicators of topics I am interested in. I am still weeding out coverage of individual new books and best-of book lists, which I do not really need for work. I was particularly intrigued by the ability to trap a URL, which gives me the ability to follow a story as it evolves. I trapped the URL of this story about Apple (NSDQ: AAPL) adding an e-singles section to iTunes. Doing so brought up a host of other pieces, including an interview with Amazon’s Kindle Singles editor David Blum. (It also brought up some stupid articles, like a piece called “Kindle Books are Hot Hot Hot.”)
Trapit will not replace Google Reader for me, mainly because I want to read all the posts on a certain few sites as soon as they’re posted, and I don’t completely trust an algorithm not to miss something. (An “always show this source” feature on Trapit would be useful.) Overall, it would be easy to sit quietly at my desk clicking through the site for hours (something Trapit would surely like me to do), improving my own recommendations in a constant quest to create my own private Internet. FYI, that would be Kindle and dogs available for adoption with a dash of Trader Joe’s news, Rick Perry gaffes and cookies.