Compared to Cuil, yes. But I still like the layout of Yahoo!'s normal search. Whose idea was it that stacked results didn't work for conveying amounts of information?
Stacked results are what work best.
But Cuil had to make themselves look visually inovative, just to be considered 'diferent' from Google. If not journalists would say that they were equal to Google/Yahoo - no matter how better, or worse, the technology and search results.
Even, as it is the case, the UI doesn't help the user.
this kinda stuff has to embarrass cuil, the results are better (yahoo) but still perhaps an engine like goog, Y! or MSN will by their indexing technology (that is supposedly cheaper)
No, it doesn't. Cuill, like Powerset, is a new search engine still doing relevance and query reformulation tuning. It's a lot of work to get results as good as Google and Yahoo! deliver, and a large part of it is a massive body of training data to work with.
People who think this should embarrass Cuill are profoundly missing the point. It's a technology demo to show, "Yes we can make a functional search engine and our technology is not smoke and rainbows".
Powerset faced the same absurd criticism, and it was frustrating because people were so caught up in the search wars that they didn't see the product as its own entity.
I agree with you, but this is also why Cuill's enormous marketing push this week was such a bad idea, even though it was executed brilliantly. (what's the last startup that you remember that had an article about it in the New York Times on its launch day?)
It's not very difficult to make a functional search engine. With the amount of resources cuil had, all you would have to do is provision a few thousand servers and deploy a nutch cluster on them.
Then you pair random images with random search results. Voila, you have cuil.
The criticism is duly applied as Cuil made the claim that they would be better than Google at the same game. Powerset never made such a claim. The people who criticized it didn't have an understanding about what it was actually trying to do.
> It's not very difficult to make a functional search engine. With the amount of resources cuil had, all you would have to do is provision a few thousand servers and deploy a nutch cluster on them.
This is incredibly wrong. What Cuil has done is neither easy (in terms of effort) or trivial (in terms of overall accomplishment). Most people simply don't understand the incredible amount of infrastructure that goes into a search engine. Certainly I did not until being part of what Powerset did, and I was shocked.
> Then you pair random images with random search results. Voila, you have cuil.
I confess I am confused why they don't just turn the feature off. It's certainly not baked yet.
> The criticism is duly applied as Cuil made the claim that they would be better than Google at the same game.
They can be better than Google at the same game at some point in the future. Right now all they have is a faster index generator, which is obviously what they feel they need to pursue funding or acquisition.
> Powerset never made such a claim. The people who criticized it didn't have an understanding about what it was actually trying to do.
I am watching the same series of events I lived through with Powerset unfold for Cuill, and it is no more tolerable from the outside than the inside.
Cuil is much more worth than powerset is.
PowerSet was searching through about 5 million pages? No need for a huge cluster farm!
No offense meant, but I bet if Powerset wouldn't have been sold, you probably would be workless as they would have run out of money very very soon. You can't make money searching through wikipedia, and the technology was nothing more like phrase detection (with verbs, nouns, etc) and simple dictionary lookup. The interface was nice though.
I bet the technology wouldn't work with foreign languages too... ;).
What I think powerset has is the skills to run large clusters, to deploy software on a large scale, and to scale things up. These are things that are interesting, but I think the Microsoft guys can do this as well without buying powerset.
If Microsoft would have spend the 100 Million they gave out for buying powerset on 100 google engineers and give each of them 1/3 million each year (so they leave google and I bet many would!), at the end of 2 years, they would have had a much better product at a far better price!
I'm sorry, I didn't mean to claim that it was easy to make a truly well-functioning search engine (i.e. Google). More that it's pretty easy to index a bunch of pages and return non-relevant results. While scaling is not a trivial problem, $33mm in capital should be enough to buy the expertise required to solve it. It's not a problem that hasn't been solved before. Of course Cuil might one day be better than Google, just like Company X might one day be better than Google.
The point is that they aren't there today so it is completely bizarre in my mind that they would come out guns-a-blazin' with these huge claims. How do we even know that they have a faster index generator than Google? What does this self-reported 130 billion page number even mean? Just that they have downloaded that many pages? It doesn't seem like they have indexed them all, properly anyway.
Cuil made the mistake of telling the user they found 0 results. Instead they should have used Cuil for main results, and then if nothing was found should have displayed Google or Yahoo results(probably Yahoo because of the whole Boss thing).
That way at least half of your results would be relevant.
