Microsoft promises that this is just the beginning of the integration with the recently acquired Powerset, but incorporating better Wikipedia clips into Live Search is a far cry from the original promise of the next generation search startup: true natural language search.
Instead we have more Live Search results with dedicated answers (not sure why Powerset was needed for this), and better Wikipedia results (“Since Wikipedia articles show up in a large percentage of Live Search queries, it’s important that the captions are top notch.”).
The Powerset team is saying they have “plans for deeper integration in the future.” Let’s sure hope so, because better Wikipedia search results is not going to drive me to Live Search as my primary search engine.
Powerset promised to change the way we searched the web. Let’s see some of that.
I hope that is not a story and becomes a reality
Great article!! Keep up the good work.
Powerset is not so powerful yet! Need to spend more time in PhotoShop.
Man what did powerset ever do to you?
They’re actually trying to build something new and useful, unlike the most of the Y-crap or tc50 dripping down the wall.
Agreed. Mike, couldn’t you have been more positive. “Powerset, The Reproductive-Challenge Version.” So negative…
Until Powerset does something significant, they deserve to get slammed. Anyone can try something, actually succeeding is when you can get a positive story. Until then they are worse than Cuil if that’s possible.
Really interesting article =)
Powerset was the best newcomer so far, not like Cuil. I really like the site and it has its target. And yes, some features would be necessary for Live.
Let’s hope that this is just the beginning for Live. I really like the xRank features.
Typical stupid Microsoft. They can’t innovate or create worthwhile functionality for their shoddy products, so they try to cram in clunky extras that don’t fit, don’t work, and end up making Microsoft look stupid. What a clever idea—canned Wikipedia results for searches. I’m sure it took a legion of Microsoft engineers to come up with that one. Live Search is not so bad, if it weren’t created by Microsoft. A friend of mine is doing an informal study where he is testing the functionality of search engines. He has test subjects submit search queries using pages absent of identifying logos and language. I took his test and tried a bunch of searches, and I actually thought Live Search was somewhat better than Google. Once I discovered what had happened, I felt like I had been raped. How foul to think that I had touched a Microsoft product! Give it up Microsoft. No matter how good you products become, no matter how ever-so-hard you try to look competent, people don’t want to use your garbage. I would rather have my organs extracted and donated to science than use a Microsoft Live Search.
Wow…you find it better than Google, and then criticize it for being Microsoft, the sole reason not to use it.
Ever heard of change? Microsoft may have been creating shady products, but some of their newer products are better. I just don’t get how you can say it’s better than Google and then not use it.
This one is pretty simple really and kind of sums it up. I personally have had slightly better luck w/ Google search, but everyone at Google, Microsoft, and Yahoo are racing for a technological solution to the search war and what they simply don’t get is, it’s a branding problem.
People use Google because its Google (and in some cases because its NOT Microsoft). It isn’t geek sheek or any other kind of trendy to be Microsoft. The solution to Microsoft’s woes isn’t in better algorithms, its in better marketing and branding. And ol’ Jason here is a perfect case study…
Mike I think you actually improved their dumb ass logo.
Hi.I’m 17 years old handicapped blogger from Pakistan.I have rickets.A disease of bones.I have started blogging to earn money for my treatment as i belong to poor family.I have written an article which i think you would like.Please sir,visit the link and give one comment.Just because of you i can get some traffic and money for my treatment.I will thankful to you.Please sir help me.Please leave one comment.Will always pray for your prosperity.God give you more success.Thanks ! http://www.propakistani.com/20.....is-biased/
huh. seems like you should add the rickets thing to your about page.
Arrington, there is nothing like that… i think this comment was posted by some opponent of mine. I am owner of this blog (http://www.propakistani.com) and there is nothing like what this commenter said….!
Please consider it SPAM!
Thanks
I’m expecting a 10 million dollar check from the widow of the former president of Nigeria and as soon I get that check I’ll help pay for your treatment.
i get a 10% finders fee for putting you two in touch.
Drink a glass of milk already. The website can wait.
natural language search is a boring. we have natural language search, its “keyword algorithim based.” getting into paragraph structure is needless.
“natural languge domain name networks” is the best way to integrate the net and get rid of the spam algor results we see today.
Looks like this guy http://www.freshmoo.com will be better soon if MS don’t do something about PowerSet.
Hopw that MS really is gonna make it better. It is a nice service.
This sort of wikipedia fascination should soon give rise to a site facilitating “Wikipedia - The Game”, wherein your goal is to change the information on a page just for a brief but important period of time so that someone else using that information gets the wrong answer.
Example- Teacher assigns a research paper. Many of your classmates will go home and use live search (and thus wikipedia) to research it. You go home first thing and change the wikipedia entry so all of the data is wrong. All of your classmates turn in reports based on the wrong data. In the mean time, Wikipedia editors put the correct answer back. Now teacher flunks all the students foolish enough to rely on Microsoft search for the answer. Further, teacher also suspects mass cheating since how could so many could write the exact same *absurdly wrong* answer? Students claim they got answer from wikipedia but teacher looks and doesn’t see anything to support their claims, thus figures they are lying. So not only do your classmates get flunked, but they get suspended for cheating. You come off looking so smart and, more importantly, get to gloat after. +50 points.
Great fun all around!
Hey retard, all wikipedia edits are viewable in the ‘history’ tab, so your teacher could confirm a previous version of the entry.
Get off the interwebs and go study. You might learn something and prevent further idiotic statements in the future (though I doubt it).
wow. TC readers are pissed off today for some reason.
You have been getting quite snarky yourself over the last few days.
that sounds vaguely threatening.
Well, I really like this TechCrunch post because someone commented with a link about freshmoo and, on the heels of that, I found this nice song by Regina Spektor http://www.freshmoo.com/watch/....._The_Call_(Narnia_Soundtrack).html
I saw a press release (multiple postings) announcing Cognition’s comprehensive semantic map (”world’s largest semantic map”). It seems to me that one of the reasons that Powerset can’t yet provide much value to Microsoft is that it is missing this key component in their architecture. Perhaps they should consider licensing Cognition’s semantic map.
By the way, it’s funny that TechCrunch didn’t pick up this piece of news… everyone else did.
Not happy at what MSFT has been doing with Farecast and now Powerset … sadly the only correlation I can think of is MSFT is trying to put racing stripes on a Ford Escort :/
Build a search engine based on natural language processing is really hard. I work as NLP researcher, on a project similar to PowerSet, but it’s at the first step (Mnemoo.com) and this is a research project (university).
For me NLP and AI is really closest, so we need to do more steps in AI before have NLP, and NLP-based search engine, that really work as intended.
What i dont understand is why there are more $$$ on some NLP project and search engine that use “old-school” NLP (and AI)… What future they have ? Instead, for me, what is looking more interesting are solutions based on biologically ispired artificial intelligence (and, of course, NLP) theory.