July 28, 2008

Chuckling at Cuil: Not Ready for Prime Time

Greetings. The New York Times today ran something of a fluff piece about a new search engine named "Cuil", which includes among its founding staff some former Google engineers.

Cuil (yeah, pronounced as "cool" -- too cute, eh?) appears to be trying to position itself as the "anti-Google" -- particularly by trumpeting a "we don't keep search logs" privacy policy.

I've spent some time experimenting with Cuil. I'm afraid that my initial impressions may be classified as "cruel" rather than cool by some readers ...

Let's start with that "anonymous searches" privacy policy. First off, let's all keep in mind that what Cuil is doing is throwing out data that they might otherwise collect, and that at any point -- either on their own volition or under orders from on high -- collecting identifying search information would typically be as simple as changing a few lines of code.

Perhaps more to the point, I've never been opposed to the collection of such data for reasonable periods of time in raw form. It's useful not only for personalized search implementations and tuning of search algorithms and services over time, but also can be invaluable for fighting network abuses of various sorts as well. I do become concerned when such data is held in non-anonymized forms for long or indefinite periods, increasing the probability of it being abused by outside parties demanding access to that log data for their own purposes.

So it's really a matter of balance. Frankly, when I see an "absolute" privacy policy like Cuil's, I find myself questioning if a reasonable balance is in place there, or rather if Cuil is currently grandstanding for the sake of publicity, and how this will affect the overall usefulness of their product.

And indeed, at least as it stands right now, Cuil needs some serious work. In my testing to date, Cuil's search results generally -- to use a technical term -- suck. Wacky results galore apparent immediately, including combining unrelated results that should have been separated, associating (over and over again!) completely erroneous photos with the wrong texts, and masses of just plain wrong or highly misleading results -- some of which are so ridiculous that one wonders how they became associated within the Cuil index in the first place.

However, I'll admit that in its current state there is a certain entertainment value to Cuil. I rarely laugh out loud when using search engines, but I got some good chuckles and at least one good "choked on the water I was drinking" guffaw from some of the hilariously incorrect, twisted search results that Cuil proudly presented.

No doubt Cuil will be working hard to improve and we'll see how they develop. But somehow I don't think that the Google folks are sweating buckets about these guys right now -- unless Google plans to start a "Get Some Laughs from Purposely Wrong Results (beta)" search engine, that is.


Posted by Lauren at July 28, 2008 01:22 PM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein