Cool Discussion of Cuil
July 28, 2008
Xooglers Anna Patterson and Louth man Tom Costello (husband and wife brains behind Xift which sold to AltaVista.com and Recall), Louis Monier (AltaVista.com top wizard), and Russell Power (worked on TeraGoogle) teamed up to create a next-generation Google. Michael Liedtke’s New Search Engine Claims Three Times the Grunt of Google is worth reading. You can find one instance of the write up here.
TechCrunch wrote about Cuil in 2007. You can read that essay here. The key points in the TechCrunch write up were that Cuil can index Web content faster and more economically than Google. Venture funding was $33 million, which is a healthy chunk for search technology.
Mr. Liedtke pulls together some useful information. For me, the most interesting points in the write up were:
- The Cuil index contains 120 billion Web pages.
- Cuil is derived from an Irish name.
- The search results will appear in a “magazine like format”, not a laundry list of results.
- Google has looked the same for the last 10 years and will look the same in the next 10 years.
Although Dr. Patterson left Google in 2006, she authored several patent documents related to search. I profiled these documents in Google Version 2.0, and these provide some insight into how Dr. Patterson thinks about extracting meaning from content. The patent documents are available from the USPTO, and she is listed as the sole inventor on the patent applications.
If Cuil’s index contains 120 billion Web pages, it would be three times larger than Google’s Web page index of 40 billion Web pages and six times larger than Live.com 20 billion page index. Google has indexed structured data which makes the index far larger, but Google does not reveal the total number of items in its index. The “my fish was this big” approach to search is essentially meaningless without context.
The AltaVista.com connection via Louis Monier is important. A number of AltaVista.com engineers have not joined Google. One company–Exalead–has plumbing that meets or exceeds Google’s infrastructure. My thought is that Cuil will include innovations that Google cannot easily retrofit. Therefore, if Exalead has a killer infrastructure, it is likely that Cuil will have one too. As Mr. Liedtke’s article points out, Google has not changed search in a decade. This observation comes from Dr. Patterson and may have some truth in it. But as Google grows larger, radical change becomes more difficult no matter how many lava lamps there are in the Mountain View office.
The experience Dr. Costello gained in the Web Fountain work for IBM suggests that text analytics will get more than casual treatment. Analytics may play a far larger role in Cuil than it did in either Recall, Xift, or Google for that matter.
The knowledge DNA of the Cuil founders is important. There’s Stanford University, the University of Washington, and AltaVista.com. I make quick judgments about new search technology by looking for this type of knowledge fingerprint.
Other links you may find useful:
- Cuil bios are here.
- Independent Ireland write up about the company is here.
- A run down of Xooglers who jumped ship is here and here.
- A brief description of Xift is here.
- Recall info is here. Scroll down to this headline “Recall Search through Past”
- WebFountain architecture info is here. You have to download the section in which you have interest.
With $33 million in venture funding, it’s tough to determine if Cuil will compete with Google or sell out. This company is on my watch list. If you have information to share about Cuil, please, post it in the comments section.
Stephen Arnold, July 28, 2008