There has been a lot of speculation surrounding the value of Powerset’s technology, and the Microsoft acquisition of the company.
Last week’s Techcrunch interview with Barney Pell and Ramez Naam contains a lot of insight about the technolo…

From the beta, it seems Powerset is not doing not much more than identifying verbs and nouns and summarizing them.

Can someone identify even one thing that Powerset is doing that’s hard for anyone to code? I am quite certain that it will not make a noticeable difference to Microsoft’s offering.

Microsoft has acquired (after vehemently denying) Powerset, a San Francisco-based search and natural language company (see the Live Search Blog and Powerset Blog for more details). Between Microsoft’s $1.2 Billion acquisition of FAST earlier this year…

Which do you think will rule from human-powered search or semantic-web search?

Microsoft will have to deliver a LOT more value over what Google currently offers to give people enough reason to switch. To do that, this $100M acquisition is just one tool in the kit that they need to deliver the required incremental value. I am excited to see what else they’re going to add. Right now it seems like they still aren’t armed enough to take on the big G.

It’s deceptive.

If I fund a study to say that my social networking script is X times better than PHPFox, and I make it look like an independent review board did it in order to give the illusion of legitimacy, then that would be grounds for customers to request a refund.

http://www.informationweek.com.....=197002546

“In an e-mail dated Nov. 1, 2002, Kevin Johnson, now the head of Windows, wrote: “I don’t like it to be public on the doc that we sponsored it because I don’t think the outcome is as favorable as we had hoped. I just don’t like competitors using it as ammo against us. It is easier if it doesn’t mention that we sponsored it.”

You think Microsoft spent 100 million dollars on a virgin natural language processor without a single usability study on NLP vs. keyword search???

That’s illogical. There are a few papers on NLP, but none that cite real usability tests over Google and other engines’ keyword searches.

So they blindly spent 100M on R&D they could have done themselves?

Maybe, or maybe they did do some studies and didn’t publish the results. 100M is a lot of money. This isn’t Yahoo or a profitable company. This is 100M for R&D. There is a difference.

(It looks like fear of being outed triumphed, and Microsoft ultimately decided to admit its role in commissioning the IDC TCO and subsequent anti-Linux studies.)

How many times that I explained to you here about property rights? What Microsoft did was within its own rights to do that. It paid IDC with its own money, not money siphoned out off the taxpayer. They didn’t break the law here, so what’s the problem? Perhaps you want the Pope to come and run Microsoft, heh?

contractor.powerset.com
vendor.powerset.com
dental.powerset.com
home.powerset. com
specialist.powerset.com
gardener.powerset.com

welcome to the world of natural language location. you can give your algorithim to a college and let them have fun with it or throw it in the garbage because on long tail lit s not gonna help businesses and consumers find anything in a structured common sense uniform manner.

RealityLocator.com says if the internet had started out with LocationEngines instead of SearchEngines we would not have confused all the consumers and engineers that search was the answer when Location has been and always will be the ultimate key to being found on the internet.

http://www.NaturalLocator.com - search solved

Well , if you’ve noticed what I have said consistently in most of my comments about the subject of search here at TechCrunch that numeric-based algorithms are still better than NLP in terms of recall & precision. Also I have stated that the future perhaps lie in the marriage of the two , NLP-based symbolic search and numeric-based search, the ultimatum of search.

Mind you that Microsoft has an army of R&D dudes who’s jobs are to play with algorithm developments all day everyday. Who knows what they’ve been upto. They might have figured out how to combine symbolic NLP and numeric-based Search Live, and perhaps this explains why they went looking for a solution like Powerset? I don’t know the answer but they must see value in Powerset’s technology or otherwise, they wouldn’t have gone that far as to acquire it.

I agree though that it’s not necessarily just about how good your algorithms are - the thing is, once your algorithm is good enough, it’s good enough. A user is not necessarily going to notice that the algorithm just got 0.01% better.

My point is… if there are 10 ‘good enough’ keyword search engines, as a user, why would I bother changing? If I’m a Google user, and my friend tells me to “try Yahoo it’s good” - I’ll probably give it a try, but find that it’s pretty much the same as Google, so I’m going to stay with Google. But if MS pulls this off and, with Powerset, deliver a completely different search experience that also delivers quality results, when my friend tells me to try it, I’m going to think: oh this is a cool and it’s different, let me bookmark this.

Search is like any other market, to compete, you need to differentiate yourself from the competition, so I think MS/Powerset are on the right tracks - whether they will succeed? I guess that depends on the quality of the service! (not market share!!)