00:31 AM

Search startup Powerset says it can knock off Google

Matt Marshall at VentureBeat reports on Powerset, a new startup which has been boldly telling VC firms that it has search technology that will beat the pants off Google.

The company's website proclaims:

Our unique innovations in search are rooted in breakthrough technologies that take advantage of the structure and nuances of natural language. Using these advanced techniques, Powerset is building a large-scale search engine that breaks the confines of keyword search. By making search more natural and intuitive, Powerset is fundamentally changing how we search the web, and delivering higher quality results.

Powerset is led by Barney Pell, an AI expert who has worked for NASA and VC firm Mayfield. The company is expecting a super-high valuation for its first round, Marshall says.

We’ve heard Powerset is on the verge of raising $10 million, and has been asking to be valued at $20 million before the investment. In other words, if it gets the $10 million, the company is valued after the deal at $30 million ($20M + $10M). That gives venture capitalists a third of the company ownership in return for their investment.

Powerset has reportedly been hiring like crazy, poaching from Yahoo in particular — on the promise that it is about to raise the round.

That last statement is born out by a post on Pell's blog about a mid-summer party celebrating the landing of angel funding.

We thought about having a contest for people to guess what Powerset is up to, since it's unusual to throw a party while a company is still "semi-stealth", but then we decided it might be getting too easy by now and there was no need to tempt fate. Fortnately we did have NDAs and job applications available, in classic silicon valley party style.

I was shocked by Marshall's statement that "search has largely been solved by Google and others, at least for the average person. Yes, there are many incremental improvements that should be made, but is there anything new that Google’s thousand-odd engineers (or Yahoo’s for that matter) can’t figure out and copy within a few months?"

Yes. Search actually is quite broken in that you can't do anything with the results except click on them. You can't analyze the results to come up with answers to the underlying questions that are behind every search query. You can't understand whether the hit results are blogs or About.com pages or spam sites or newspaper articles. There's no system for authority or authenticity of pages or content. And the whole system of valuing pointed-to pages is loaded with vulnerability to spam sites and SEO actions. If I post an article that Google is about to buy Apple, that's going to get a whole lot of googlejuice, regardless of the fact that it is completely and utterly false.

So, no, I wouldn't say search is solved. On a simpler note, Jonathan Grubb comments on the VentureBeat article:

If you search for “I want some pizza” in Google there are no results that take me closer to eating a pizza. “How much is powerset worth” doesn’t give me any good information either. “Who writes venturebeat” doesn’t bring up your name. “How much does an ipod nano cost” doesn’t work, you have to type “ipod nano 1gb price.”

Having said that, it's a long way from NDA parties to the top of the Google mountain, but the maturity of the present search market - invented by Google way back in the late 90s - suggests that there is a new breakthrough coming in search technology. Google is a huge company and huge companies cannot easily discard the technical underpinnings of their empires. Even though Google has maintained an admirable openness to new ideas both internally and through acquisitions, often new ideas require new companies.

I remember being in a meeting with Brewster Kahle (now of Internet Archive) and John Warnock (then CEO of Adobe) in the early 90s, when Brewster had just started WAIS and Adobe had just rolled out Acrobat. He told Warnock, "I hate starting companies but sometimes it's the right thing to do." I think the statement was disingenous (apparently he loves to start new things) but it goes to the point that there are some things big - even smart, even counterculture - companies can't do. In that meeting, for example, Warnock couldn't see how Internet search would have anything to do with Adobe Acrobat.