Note: during live blogging, see my comments on Twitter
Steven Marder, Eurekster
Distributed social search & discovery. Eurekster focus: help publishers create specialized search engines (with video, image, text, etc; prioritize certain domains) which learn from consumers’ usage. Value proposition: focus on editorial control. Custom search engine, the Swicki, also available as a widget. It uses publishers and users, algorithms and humans. Now, 100,000 engines by 23,000+ publishers.
Jimmy Wales, Wikia Search 0.2
Big differentiator for them: all of it’s open-source. "Search has been a black box for too long. We have only the vaguest idea of why and how Google ranks things." [There are some people in the room who might disagree.] He notes Wikia Search’s results "are not very good right now." Only crawling small subset of the web. No stealth mode startup phase. Anybody can add a URL to the displayed results. Anyone can add, delete comments, or spotlight an entry to help make results easier to find if they’r enot yet high enough. WS0.2 includes related searches, also open to debate.
Big challenge: spam will emerge. Wikipedia: link spam not as much an issue. In the search space "there are a lot of economically motivated bad actors." Lots of tools to manage deletion, undeletion. Bad guys can be blocked. There’s social pressure where people’s reputation among their connections matters. Some features are on the site now, others are in the works.
My main question: even with all of this crowdsourcing, will it be better than Google? That’s the acid test and all that matters.
Jason Calacanis, Mahalo
Human-powered ssearch engine – "ungameable." The search engines themselves have issues – eg for searches on "paris hotels," Google has affiliate spam high up, and on Ask you can’t tell what’s an ad or natural. Mahalo’s more organized. Refers to "the second click service." For Paris hotels, Mahalo cross-references what all its main travel sites say about each major hotel listed, mimicking user behavior.
Trust score built for each individual submitting links. "Google essentially created the SEO industry… if you spend your time on SEO and not building a great site, it’s basically wasted energy." It’s a ‘duh’ comment from Jason, a kind of straw man, except that there are some SEOs who need to hear it.
Mahalo Social: when seeing people’s recommendations on the sidebar for searches, two factors matter: how close they are (degrees away, a la LinkedIn) and their trust score.
Calacanis’s vision for Mahalo: it’ll take $50 million to build it out, $100 million a year to keep it going, with an editorial staff of thousands.
Refers to "curated search" as alternative to human powered.
For SEO, he says a huge problem is that a lot of the clients have "sucky" sites that don’t deserve to be ranked. The job of a good SEO is to fix that. And a good SEO also won’t try to get sites to rank for queries that the page has no business appearing for. He compares it to great companies not hiring PR firms. He’s wrong about PR too though. Today I’ve received emails from five PR people who represent five different companies that can all claim to be great in their own way. And I probably would have no interaction with those five companies right now if the PR people didn’t send me a heads up. This one hit a nerve; I love PR folks, and while there many who don’t do a good job (there are bad teachers, firemen, and sanitation workers too), most I’ve met are pretty sharp if not brilliant. And I’ll note that PR folks are definitely overrepresented among my personal friends – even more so than search engine optimizers, who are also alright in my book.