Cuil Needs To Fix Its Technology Before It Can Get Hot

Search engine expert Stephen Arnold believes the surge of interest in Cuil shows there's a huge pent-up demand for an alternative to Google.

W. David Gardner, Contributor

July 31, 2008

4 Min Read

With the dust -- and the first wave of searches -- settling on the new Cuil search engine, most of the reviews were negative when results were compared with Google.

But search engine expert Stephen Arnold believes Cuil's debut has some impressive features -- and potential problems -- that will show up later.

What stands out in Arnold's analysis is that after spending a few million dollars, Cuil has been able to search an index the start-up claims is three times larger than that of Google, which has spent billions developing its technology. "It's a normal start-up," said Arnold. "It's a miracle it works at all."

Noting that Cuil's first days of public usage stacked up poorly against Google, Arnold, in an interview Thursday said: "They'll fix it. Search is an iterative process. But what the (surge of) interest in Cuil showed is that there's a huge pent-up demand for an alternative to Google."

Arnold, who has tracked search patents for years and written about them in several books and reports including his recent "Beyond Search" for the Gilbane Group, noted that Cuil co-founder Anna Patterson left Google in 2006 after authoring several patents related to search.

"Dr Patterson joined Google in 2004 and filed six patent applications," Arnold said. "Now, as an ex Googler or Xoogler, she has focused on some of the weak points in the Google approach to search -- namely, privacy and comprehensive indexing."

Google, which officially began in 1998 after some earlier work by the company's early hires, took time to get wheeling, said Arnold, who added that Cuil probably needs additional funding. Patterson told The New York Times that the search engine was overwhelmed by 50 million queries in its first day of operation Monday and that number was way more than the startup had expected.

"We're engineers, and we are reading the press for good feedback and critical feedback," she said. "If we didn't have improvements to make, we wouldn't have a job here."

So who will want to use Cuil? Assuming Cuil improves its search results, Arnold believes the next three or four months will be crucial as searchers experiment with the service.

Users worried about privacy issues could be attracted to Cuil, Arnold said, noting that many searchers believe Google doesn't do enough to protect their identity.

"Will Cuil get traction," Arnold asked rhetorically? "The answer is yes. My hypothesis is that the folks who flock to Cuil will be Google users, but the real impact of Cuil may well be taking orphaned or disaffected users from Ask.com, Live.com, and Yahoo.com search."

Arnold points to some Cuil "bells and whistles" that can appeal to users -- features like insets for suggested searches, tabs for slicing results and "snazzier results displays." Arnold has never been a fan of Microsoft's and Yahoo's search technologies, which he notes have been cobbled together from different acquisitions while Google has had a focused approach to its search technology from its early days.

"Cuil hit Google with its larger index of 120 million Web pages processed to Google's 30 to 40 million pages," said Arnold. "Keep in mind that size doesn't matter, but it is a public relations hook." Google, which doesn't reveal its index numbers, won't concede the point to Cuil, but Cuil's capability to search so much with so little hardware -- the firm is said to use just 120 servers -- is impressive. Google has hundreds of thousands of servers in its massive data centers.

Arnold traces much of Cuil's advances -- and the advances of other search engines -- to simple hardware designs that were carried out years ago by Alta Vista, the pioneering search engine build around the Alpha 64-bit processor at Digital Equipment Corporation. Cuil's husband-and-wife founders team Patterson and her husband, Tom Costello, who researched and developed search technology at IBM, have taken advantage of those hardware advances. Arnold expects Costello to leverage his background in text analytics at IBM to improve Cuil, possibly with data mining elements.

The presence at Cuil of Louis Monier, ex-Alta Vista technologist, is important, Arnold believes because the Alpha experience with multicore processors helps Cuil take advantage of "cheap data centers" for search.

In Arnold's view, the current leader using multicore processors is France's rapidly-growing Exalead, also founded by a former Alta Vista technologist, Francois Bourdoncle. "Exalead and Cuil are newer (hardware) technology than Google's," said Arnold. "And that can give them an edge over Google. Cuil can use fewer machines and get more from Web sites."

Arnold said Cuil has some roll out bugs to squash. The Cuil system often pulls an image from one Web site and places that image in a description of another Web site. "Cuil also has to address bandwidth issues," Arnold said. "The surge in users Monday caused slowdowns and brown outs. These are normal glitches. Google had them. Cuil has them."

Although Arnold sees plenty of opportunities for search engines competing with Google, he believes Google's dominance is secure. "These smaller search engine companies can become hundreds of millions companies," he said. "Google can be a hundreds of billions company."

Read more about:

2008
Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like


More Insights