Does Microsoft + Powerset Beat Google? | Hacker News Reader

Does Microsoft + Powerset Beat Google?

readwriteweb.com

7 pointsAnon848 comments

If it scales yes, why not?

If it doesnt, well no ldo.

How long would it take for Google to implement something similar?

why not?

I haven't seen anything in Powerset that would augment Microsoft's search enough to make a significant impact on their "battle" with Google, never mind beating them.

Powerset only works on wikipedia, it's a cool application of NLP...but scaling it to a full "semantic" search engine for the entire web is an epic challenge. Is Powerset's search technology really better than Google's? I'm not too sure.

It's a good acquisition for msft, but Google is usually better when it comes to choosing startups to buy that are real winners. I honestly think that Google would have scooped them up sometime over the last 3 years if they really thought they were a threat. Just my hunch.

Several reasons.

First - if Microsoft managed to somehow make a more powerful search with PowerSet then they are still lacking in reputation. Rep alone will get a large portion of Google users to stay with them.

Second - People in general don't like Microsoft. Sure they have some great products (don't knock Windows and Office, the world does run on them even though there are other great platforms/editors) but they have lost face due to poor support options and annoying aspects of the interface

Third - Google is not just a search engine.

That's just some of the reasons.

First and Second are the same. Third makes no sense.

First and second are similar yes, however Google is not just a search engine. They have a dedicated base of Gmail users, they have a great advertising service. They have youtube (although are thinking it less and less likely to turn a profit). It would be a huge blow yes if Microsoft somehow overcame Google in search functionality, however it would not be the end of Google.

Google's strength, thanks to PageRank is in Ranking, or, in other words, determining the best page to answer a given query. Like I've argued here before (and I'm no Microsoft fanboy) Powerset's version of Wikipedia might just be enough of a base for Microsoft to topple this. How, you ask?

Simple... As most people will agree Wikipedia's content is pretty good. By having it semantically linked Powerset effectively turned it in to a giant fact database. Now... Microsoft engineers can with a reasonable amount of work use this database to, in a way, fact check the pages they come across. Pages with good content will be ranked higher, while pages with poor or inaccurate content will be ranked poorly, regardless of the number of links they might receive. This would, of course, spell the end of the "Google bomb"...

... and would make a lot of different in what is usually called the "dark web", where, by it's very nature there aren't many (or sometimes any) inbound links.

Now, you might argue that Wikipedia doesn't contain all the facts in the world. And I would agree with you. You might even argue that not all the facts there are trustworthy. And once again, you would be correct. But it's a damn good starting point, and, in the right hands, it can "spread" by validating other sources of information. Say, if one page that contains a "new" fact has a large number of correct facts, it might be ok to assume that the new one is correct as well. Specially, if it is formulated similarly in different places of the web.

In summary, you can imagine it as the seed of a "factrank" that is then used to rank pages. As you well pointed out, however, inertia is a relatively large deal on the web, but not as much as in the physical world. People didn't used to be in love with Google until a few years ago... I'm sure that give the right arguments they can fall in love with another search engine.

The big question is... does Microsoft have what it takes to pull it off?

Reply on news.ycombinator.com