Tired of Google’s stagnant, antiquated approach to search, a group of ex-Googlers and other techies have opened up their own shop: Cuil.com.

Pronounced “cool”, the Cuil search engine has an index spanning 121 billion web pages, and operates by comparing your keyword to the rest of a page’s contents instead of relying on “superficial popularity metrics”. This mechanism calls to mind the semantic web, but Cuil’s About page is too ambiguous to tell for sure how Cuil works.

What is sure is that Cuil’s team sees their engine as the future of search. Anna Patterson, who moved to Google after building the search engine used by Internet Archive, has low hopes for Google. From CNN Money:

Patterson enjoyed her time at Google, but became disenchanted with the company’s approach to search. “Google has looked pretty much the same for 10 years now,” she said, “and I can guarantee it will look the same a year from now.”

With these bold words and the talent behind them, as well as Cuil’s claim to fame of indexing a hypothesized three times as many web pages as Google, it is hard dismiss this Google contender along with all the others. But even a cursory glance at the engine’s functionality shows a less-than-polished product without much bite.

A lot of people have been finding massive differences in the number of search results between Cuil versus Google. An interesting example is found in “banana” and its French translation “banane”. “Banana” returns 47 million results in Cuil versus 97 million in Google, while “banane” closes the lead a little bit by returning 5 million results versus 11 million (though still roughly the same percentage-wise). A clear numbers win for Google, but dig a little deeper and you see Cuil vastly improving within only 24 hours since its launch. PontifexPrimus originally found only 15 million results for “banana”, and none at all for “banane” (unless you, curiously, turned Safe Search off).

Of course, the raw number of search results is much less important than the number of useful results. Unfortunately, Cuil falters here as well. Searching for “Spencer Sugarman” on Google returns maybe every individual website I am involved with, all on the front page. Cuil, on the other hand, returns 10 links to Undress Me Robot, and 1 link to an article about Burt Sugarman. The “banana” search mentioned above is much better, providing many relevant and unique results. But what Cuil fails to do on any search I have tried is surpass the quality of Google’s results. Instead, all they manage so far is to land somewhere along the spectrum of Much Worse to As Good.

This is likely to change. It is difficult to tell whether Cuil is incorporating changes as its engine is stress-tested, or if the learning is built-in, but either way the engine seems to be improving. If Cuil expands its hardware to ward against the downtime problems they were experiencing yesterday, and continues improving its search results, the search engine could soon stand strong against Google. But as CNN points out, Google’s immense power as a brand means winning will take a lot more than just marginally better search results or a slicker design.

You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>