Powerset: Is There More Than Buzzwords And Patent Threats?
from the do-we-have-anything-useful? dept
There's been so much hype around search startup Powerset that it seems like it's going to be quite difficult to live up to it. The company kicked off by raising a lot of money at an insanely high valuation for a seed stage company, and then used some of that cash to license some natural language technology from PARC. Of course, natural language search has been tried and failed many times before -- sometimes because the technology sucks, but more often because there just isn't that big a benefit to it compared to traditional keyword search (especially as more people have become comfortable with keyword searching). However, Powerset keeps generating lots of attention and hype, and on Thursday apparently revealed a lot more concerning what it's about... we think. That is, the company revealed a lot, but an awful lot of it comes off as simply repeating every buzzword they can think of and reminding everyone they have patents.It's always a signal to be worried if a company kicks off a description of its product by bragging about its patents rather than the actual benefits of its product, but Powerset kicked off the discussion by talking about how "locked down" its patents are. If the company is really doing something special, then people will beat a path to its door, whether or not it has patents. If the technology is useless, the patents will also be meaningless. We don't care about the patents, we care about what's useful. The rest of the talk apparently was about this incredibly confusing buzzword-fest of a social network/ecosystem that the company is apparently trying to build around its search engine:
"Imagine a mashup between Facebook, Digg and Google Apps, but you get to participate in the building of the products that sit on top of our platform. You log into a social network, like you would Facebook, and you get certified to be a Powerlabber. Once certified you can join different interest groups, such as travel, and participate in idea and mashup competitions. QA is embedded and its all bloggable."What does that mean? I've read it many times and I still can't figure it out. He goes on to mention MySpace, Second Life and Wikipedia, of course. It sounds like the company is trying to build the ultimate web platform -- which is a good strategy, but it needs to get away from buzzwords and patents and actually explain what makes it useful.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Reader Comments
Subscribe: RSS
View by: Time | Thread
Intersting...
However useful this turns out to be I also think that they may become victims of their own hype.
[ link to this | view in chronology ]
Powerset does have some secret sauce
Focus on the search index, not the user query. The NLP rocket science is applied to indexing the billions of web pages. NLP is not that helpful in parsing the typical two or three word query, but that is what everyone focuses on.
However, take a quick look at this query. "Who is the best ballplayer of all time?" Powerset breaks this query down very carefully using linguistic ontologies and all sorts of proprietary rules. For example, they know that "ballplayer" can mean Sports. Sports can be separated into categories that involve a "Ball". Things like baseball, basketball, soccer, and football. Note that soccer does not include the word ball, yet Powerset knows this is a sport that includes a ball. Powerset knows that "ballplayer" can mean an individual player of a sport that includes a ball. They know that "best of all time" means history, not time in the clock sense.
Knowing all of this is cool but the real rocket science is in the index. Powerset uses all these rules and linguistic approaches to analyze millions and billions of web pages, and adds "meta data" hooks into each word on each page. As you can imagine this is a huge scaling problem, that has been impossible to solve economically. With Moore's Law applied to constantly reducing the cost of computing, storage, and bandwidth, it is now possible to solve this problem, and within a few years it will be economically viable.
I wrote a blog about Powerset today that goes into more detail. See http://dondodge.typepad.com/the_next_big_thing/2007/06/powerset---open.html
Don Dodge
[ link to this | view in chronology ]
Re: Powerset does have some secret sauce
I don't doubt that there's some secret sauce in there as well, I just think that the benefit needs to be more clearly laid out... because right now, all they've basically laid out is "this thing is awesome! we've got patents, and buzzword, buzzword, buzzword."
They haven't yet said why I'd want to use it.
However, take a quick look at this query. "Who is the best ballplayer of all time?" Powerset breaks this query down very carefully using linguistic ontologies and all sorts of proprietary rules. For example, they know that "ballplayer" can mean Sports. Sports can be separated into categories that involve a "Ball". Things like baseball, basketball, soccer, and football. Note that soccer does not include the word ball, yet Powerset knows this is a sport that includes a ball. Powerset knows that "ballplayer" can mean an individual player of a sport that includes a ball. They know that "best of all time" means history, not time in the clock sense.
That's cool, technically speaking... but how useful is it? How often do people need to do a search for "the best ballplayer of all time"?
Also, I just did a Google search on that phrase, and it actually turns up pretty good responses (especially for a somewhat subjective question).
So, what's the *benefit* of Powerset... I understand it has cool technology, but where's the user benefit?
[ link to this | view in chronology ]
Re: Powerset does have some secret sauce
Also try "Where did Einstein die?" and see how the meaning of "to die" is completely ignored by Powerset, which just happily returns samples containing the German determiner "die".
"Where did Babe Ruth die?" similarly suggests that "where" is just taken as a sort of keyword wild card (anything of the location type), but that the crucial NLP information (that Ruth did the dying at that location) is just ignored.
[ link to this | view in chronology ]
[ link to this | view in chronology ]
As for that quote from Google on statistics and semantics, I've tried Google's supposedly world-changing statistics-based translator, and translations I got were totally incoherent. Babelfish may not be that great, but Google is nowhere near being any threat to it.
[ link to this | view in chronology ]
However, I will point out that Barney has, in the past, created one of the best job search engines around, FlipDog. FlipDog worked because it knew what a job listing looked like and crawled the net looking for, and indexing, job listings from corporate websites. If Powerset can build something similar for the entire web, then they might be onto something, but I think that beating Google is going to be a very, very difficult task since they are the new Microsoft (eg. entrenched incumbent with huge advantages).
That said, the switching costs for search are pretty low, so perhaps it's possible...
Chris.
[ link to this | view in chronology ]
*yawn*
[ link to this | view in chronology ]
Here's a screenshot, from one of the update emails, of "PowerSet" / "PowerLabs" / PowerWhateverCoolNameWeDecideOn.
Can you say... MySpace? lol.
Like I said earlier: Sounds like yet another DotCon.
[ link to this | view in chronology ]
Clarification
I'm the product manager at Powerset for Powerlabs, so I thought that I'd clear up Steve's statement a bit. Powerlabs is going to be Powerset's platform for testing out our newest product ideas and allowing users to test them out and comment on them. The Facebook aspect is the community, the Digg aspect is the ability to rate and comment ideas, and the Google Labs (not Apps) aspect is that we're going to release products before they're ready for prime-time, e.g. in the "Labs".
If you have any questions, you know where to find us =)
-mnj
[ link to this | view in chronology ]
Re: Clarification
I signed up for Powerlabs to find out more about their work, as I'm very interested in Natural Language Processing. According to Mark Johnson's blog (http://deliberateambiguity.typepad.com/blog/2007/06/powerlabs_scree.html), it looks like it won't be opening up until September, but I did receive an e-mail from Mark with a link to this short video, which sums up what's been discussed so far: http://www.youtube.com/watch?v=8D6czWVYc-o
[ link to this | view in chronology ]
Good PR positioning for Sale?
Then again, this is all great to get google, msft, ask and aol interested in a bidding war.
Fear is a great motivational tool in M&A:
http://www.watchmojo.com/web/blog/?p=1765
[ link to this | view in chronology ]
[ link to this | view in chronology ]
Natural languge search
[ link to this | view in chronology ]
Mass confusion! We're talking about different thin
[Disclaimer: I'm NOT a Powerset employee, nor associated with Powerset in any way.]
I think there is a lot of confusion here, at different levels:
1. The Digg-Facebook-GLabs buzzwords apply to PowerLabs, the feedback community for the product [a la Dell Ideastorm], not to the search site or platform.
2. As Don Dodge points out above, the Semantic Processing aspect applies to *both* - the query and the indexed content. (This had not been clear to me before last night.) This understanding of meaning will allow Powerset to provide query results with a much higher level of relevance than keyword search, IMHO.
3. Finally, although there is incredible potential here, Powerset seems to be following a disciplined approach with the following progression: (i) search site (ii) widgets, mashups, APIs (iii) search platform .
Of course, this is all based on what they told us last night - the usual disclaimers apply!
Ob.plug for my own blog post on this topic: Powerset is Not a Google-killer!
[ link to this | view in chronology ]
Natural Search & All That Crap!
Please understand that most of the users are not very sophisticated when entering their queries. Let alone using a multi-keyword or non keyword phrase.
Google is what it is today because it could monetize those specific keywords into adwords for publsihers. How can Powerset do the same to generate revenues from user search queries applying natural search?
Where does Paris hilton buy her underwear?
I wonder what results Powerset deliver?
[ link to this | view in chronology ]
They did tell you their secret sauce.
[ link to this | view in chronology ]
Secret sauce? More like something that doesn't wor
[ link to this | view in chronology ]