WEBINAR: Available On-Demand
Innovate and Thrive: How to Compete in the API Economy REGISTER >
Jason Calacanis earned a nice payday by applying old-school standards to the new world of blogs, selling Weblogs Inc., his stable of professionally written sites (including the popular Engadget personal-tech blog) to AOL for $30 million in 2005. Now he wants to bring similar logic to the lucrative business of Web search with Mahalo (Ma-hahlow, Hawaiian for thank you), a company that publishes search results written by humans instead of assembled by algorithms.
Mahalo and Weblogs Inc. build on the idea that the leveling effects of technology, which have unleashed vast amounts of information and talent, don't preclude individuals with specific knowledge and skills from adding considerable value. "The pendulum has swung from experts only, with the old media companies telling you things, to the wisdom of the crowd as authority," says Calacanis, who ran the bubble-era magazine Silicon Alley Reporter. "I think it's going to swing to the middle, where you have that broader wisdom, but it is OK to trust an expert." Calacanis recently spoke with CIO Insight senior writer Edward Cone about the future of people-powered search and his new, venture-funded company.
With Google continuing to report huge numbers, aren't you swimming against the tide with Mahalo?
We're seeing an evolution. What might be called Web 1.0 put experts online; there were some new voices, but it was largely about traditional media moving to digital. Web 2.0 brought blogs, citizen media and the wisdom of crowds. We moved to everyone having a voice, but when we got to the idea that we would trust a bunch of people voting on Digg or building a Wikipedia page above The New York Times, that was getting a little hysterical. It doesn't mean these services are not good, but you can't rely on them alone. I believe we are moving into a third phase, where we will look for expertise from the most talented people to rise through the online meritocracy and then reward them with our attention.
You've called this next phase Web 3.0, offering the term with a certain irony that some commenters on the tech blogs seem to have missed.
People were incredibly charged and upset, as if there really was some sort of official definition of the term, or that I was setting it forever by floating the name. It was kind of funny. But I'm serious about Mahalo. I think it's going to work, although it may take five years and $50 million to prove my point.
What problems do you see with "wisdom of crowds" sites, and with the sort of automated searches that currently dominate the market?
There's so much average or bad information out there that it keeps people from getting to the good information.
On Digg, you see people fighting with each other over whether Playstation 3 is better than Xbox. On Wikipedia, you have people pushing their own agendas and publishing factually incorrect information.
With machine search, the algorithms essentially use voting--how many links you have coming in, and from whom--to determine results. The excellent content no longer ranks above the spam. That's why we're hand-writing search results. We take the machine-search results and look at them along with other sources, and we create pages based on a combination of the wisdom of crowds and the judgment of experts; we let the audience scrutinize pages and put in links, but we make the fi nal decision. Voting is a great device, but there's a reason we have a representative government.
What are the implications of this approach to search for businesses and technology managers?
If they have put search against assets within their organizations, they've probably found out the results were bad. It might be good for them to actually curate a bit, maybe with something like a wiki. You could actually hire people to organize the information your search engine pulls together--say, to look at human resources information, clean it up and make sure the best results are evident.
It sounds obvious, but sometimes people try to brute-force a technology solution. They don't think, if the technology is going to cost a million dollars, maybe we should hire fi ve people for $50,000 each to do this. We have a bias in the technology industry to actually prefer machine processes over human processes, and we don't always make the most objective decisions.