VentureBeat

updated
Powerset, the “natural language” search engine company saying it wants to take on Google, has seen a shake up in its management.

The San Francisco company received substantial hype when it emerged last year saying it had found a way to understand the phrases you type into your search queries. However, it has been slow to deliver on the technology, instead dribbling it out in small doses in a “labs” portion of its site.

Steve Newcomb (left), the chief operating officer and one of the founding troika, has left the company. Barney Pell, who was chief executive, is stepping aside and becoming chief technology officer. The company is looking for a chief executive, Pell told VentureBeat in a phone conversation.

The move comes after differences between Newcomb and Pell over the direction of the company, as well as a slip in the company’s delivery date of its product. Initially, it had planned to release its search engine to the public this year. Now, it plans to do that by the second quarter of next year, though it come even later. “This is really hard, what we’re doing,” Pell said, of the company’s ambitious goals. He said one cause of the delay was the time it took for Powerset to license key natural language technology from PARC. Powerset finally got access to PARC’s source code in January this year. The company is now on track, he said.

Pell has published the changes on his blog.

Powerset has also been trying raise money from venture capitalists at a very high valuation, and the shakeup suggests it wasn’t getting very far in its efforts.

This is instructive for founders and entrepreneurs who try to raise money early at stratospheric levels: Powerset last year raised $12.5 million last year from Foundation Capital, the Founders Fund and long list of individuals, giving it a post-money valuation of $42.5 million. Now, in order to raise money again, the company needs to seek a higher valuation, so that Foundation Capital and the other investors feel they are getting their money’s worth. However, that’s hard to do unless you’ve shown you can perform on your plan. We’ve mentioned Powerset’s various product releases, but none show that it is close to prime time — a year after the investment by Foundation et. al.

It isn’t clear to what extent the investor team pushed for the changes. Pell said there was a “fair amount of tension” in the style with which he and Newcomb wanted to run the company. “We were stressing each other out. It was a challenging marriage. That’s very common in start-ups. We were trying to work things out.” That’s when the board recommended to Pell that he bring in some consultants to help on organization, Pell said. Pell indicated he was ready to step aside in order to to position the company for progress. The recommendation was that Newcomb should leave. Newcomb did not respond to a request for comment made several days ago by VentureBeat. Pell said Newcomb has interest in politics, and that a mere “founders” role wasn’t enough to keep him at the company. Pell said other changes were made in the company’s management, to make sure key employees felt increased ownership and responsibility.

It should be noted that this company has continued to show creativity in developing its product. We once called the company a funny farm, but tempered that by observing the company’s resourcefulness. The quixotic nature of its team — embodied by the enthusiastic, relentlessly upbeat Pell — is doing quite a bit in its labs to create interest (see our coverage). It has opened a search box inside its Powerlabs, and offered several use cases such as Powermouse (see coverage). While the shakeup is a blow to the company, it still has substantial technology under its hood, and will not doubt continue to engage the search community. Natural language is a huge challenge, and Powerset is more focused on this problem than any other company.

The company is still not crawling and indexing the entire Web for its search engine. It is still focused mainly on an index Wikipedia, a site that has clear structure and relationships between objects and their definitions. This gives Powerset a nice testing ground for its product. But the wider web is much more complex, and Powerset has yet to tackle it. In part, Powerset has been hampered by limited resources, Pell said. While he said he wants the engine to be ready by second quarter, it might be later than that, he said, possibly even 2009.

Powerset is a quixotic search engine company here in San Francisco that has convinced itself it can take on Google.

While Powerset gears up to release its search engine publicly later this year, it hopes to nurture an army of 50,0000 early testers, or “Powerlabbers,” to bang on different parts of it beforehand — the idea being that the converts will not only improve the product, but will help push it at launch.

To lure those volunteers, Powerset seeks to produce a hail of fun projects for them to work on — more on this below.

As reported, Powerset is significant because its search engine aims to understand phrases — not merely words, as Google does. Powerset’s approach is potentially powerful. However, it requires significant mass education before people change the keyword-like search habits.

Earlier this week, we spent another hour and a half with Powerset to learn about their latest progress. It is still secretive, but it is planning to open considerably over coming weeks.

For the testing phase, Powerset aims to stick volunteer testers on bite-sized pieces of its search problem. It is homing in on sixteen different topical areas – ranging from entertainment to travel and porn – and in each of these areas wants users to provide feedback on its results.

So for example, in the area of entertainment, if a user asks: “Who won an academy award in 2001?” Powerset finds that easy. It will produce answers like Halle Berry, who won the award for Monster’s Ball (see image below). But if you ask “What is the most recent movie Halle Berry started in?” the engine may break down. Powerset tracks the range of questions posed by users, which creates feedback about what is and isn’t working. That way, Powerset hopes to prepare in key, popular topical areas before launch.

.

Take another example, travel. See below for the topical page. Volunteers suggest ideas for useful search themes, and they vote to push the best ideas to the top.

A sign of Powerset’s readiness to think differently is its approach to Web architecture. Powerset will base its site on Ruby on Rails, a new, edgy framework liked by engineers for its nimbleness. But Ruby is controversial because some say it can’t handle vast amounts of traffic efficiently. Few big-traffic sites have built upon it.

The company which released the framework, 37Signals, has used it for four applications, including its popular Basecamp. CNET’s Chow and Chowhound – and most recently by popular messaging site, Twitter, are also built on it.

Powerset chose the framework after considerable research. Nine of Powerset’s team of 66 are working on it (Kevin Clark, the project leader, posted about the decision here).

This is not a company led by one or two brilliant co-founders. Rather, it is a team of now dozens of engineers — who to the outsider seem to share a single quality, a sort of wide-eyed, ebullient confidence, embodied by the relentlessly upbeat chief executive himself, Barney Pell. His two co-founders, Steve Newcomb and Lorenzo Thione, share the same trait. Or, if they have doubts, they try not to show it. That’s why they may pull something off.

Natural language search, as Powerset’s approach is called, faces an enormous challenge. The sheer number of phrases and semantic senses that can be intended by searchers is overwhelming.

Breaking it off in bits makes sense.

Powerlabs, the name given for the topical test features, launches in September, and is taking sign-ups now.

Powerset will specifically target high-school teachers for training on how to use its search engine. If they are recruited, they’ll impart their knowledge to students.

In return, Powerset hopes to get feedback on its main search engine. See below for example of a query: “Who proved Fermat’s last theorem?” Powerset provides a big blue feedback box. This way, if Powerset provides a poor result, testers can alert Powerset’s engine to the shortcoming.

Powerset is also working with databases to fill its result pages with more information. We’ve been told Powerset has partnered with MetaWeb’s Freebase (first reported by Techcrunch, which misspelled the name), though Powerset wouldn’t comment. In the entertainment example above, it pulls the “meta” information stored in Metaweb about Halle Berry into a widget. The widgets are useful, even if they’re not part of the main search engine technology. Powerset hopes to let bloggers embed the widgets into their blogs when they write about related material.

Another example of meta-data being used is on the result below about Steve Jobs and the iPod — you’ll see it pulls bio information and videos.

.

After raising a decent $12.5 million in cash last year, Powerset, the secretive search engine company that wants to take on Google by using “natural language” technology, is going to raise a larger round this year.

This and other gossip about Powerset spilled out Saturday evening during a lively party the company held at Frisson restaurant in San Francisco. Powerset investor, Charles Moldow, mentioned the funding plans to us, and asserted that this time, investors may be skeptical about pushing Powerset’s already high valuation even higher.

In an unfortunate video taken by PartyCrashers of the party, Pell also confirmed the plans, saying once you raise money, you start raising money over again. Troubling was a part at the end of the video (worth watching; see below, it is short), where one Powerset guy, clearly looking as if he had too much to drink, doesn’t express much confidence in the company’s abilities — or at least, it comes across that way.

Techcrunch’s Michael Arrington said earlier today that I gush about Powerset everytime I write about it. That’s the second time in four days he’s claimed that. Clearly, he’s upset that we continue to break news about Powerset, including its launch and sky-high funding last year. In fact, I’ve tried to be pretty even-handed, each time making sure to solicit input from Powerset’s biggest critics like Danny Sullivan or Google’s Peter Norvig. Our first story suggested the investment valuation was “crazy.” Powerset has a very long way to go, and we should make that absolutely clear. If Powerset hones its natural language technology into something usable — and it just may — it will become an acquisition candidate, if not for Google, then for Yahoo and for MSN. As a stand alone company, it is a still a very long shot.

(Updated with more details on how FF works)

A form venture capital funding in Silicon Valley is getting increased interest from founders of start-ups.

It is called the “FF class” of stock, for founders who want to cash out a small percentage of their stake in a company so they don’t have to wait until the company is sold or goes public.

This practice is not entirely new. Many founders through the decades, including at Intuit years ago and Jonathan Abrams at Friendster more recently, have sold shares in their company to their venture backers and gotten cash to enjoy life a little more. But with the favorable start-up climate now, VCs are doing more to accommodate founders, entrepreneurs are getting more sophisticated, hearing more about these sorts of terms, and increasingly asking for them.

The practice is mentioned in the SF Chronicle story by Jessica Guynn, about the Sean Parker of the Founders Fund (the fund’s partners pictured above) and his decision to devise such a stock for Barney Pell, founder of young search start-up Powerset:

Inspired by his personal frustrations as a startup founder, Parker came up with a novel arrangement that he hopes will benefit other founders as they build their companies: a new type of stock that allows founders to cash out a small percentage of their stake in a funding round so they don’t have to wait until the company is sold or goes public.

Pell, who maxed out credit cards, deferred salary and considered taking out a second mortgage until he could raise serious money and interest from the right investors, says he’s relieved that he and his fellow founders don’t have to feel rushed to sell the company to get some return on their investment of energy, time and money.

This Powerset iteration was put together by attorney Steve Venuto, of Orrick, and VentureBeat has been told that Fenwick & West and Wilson Sonsini have also given thumbs up to the practice. Venuto was also the lawyer of Facebook, another company backed by the Founders Fund’s Peter Thiel. (Someone told us Venuto also the only lawyer in the valley who can code, but we haven’t confirmed that.)

The amount of the cash-out is capped to between 10 and 15 percent, depending on the round. It should also be used with caution, because it means you get less as a founder when the company does get sold or go public. There’s a story we’ve heard about the founder of Viaweb, a Paul Graham company, who cashed out in order to buy his wife a Saturn car. It became known later as the “million-dollar-Saturn,” because of the worth that stock would have been had he kept it.

Thestock also lets the investors cash out too, allowing venture capitalists take some money out of the company during a subsequent higher-value round.

The stock type is also significant because a new law, 409A, forces companies to grant options at their fair market price. This brought more scrutiny on the cash-out process. We ran this by Gordy Davidson, of Fenwick & West, who explained the challenge as follows. If a VC buys shares from a founder for $1 each, to let the founder cash out, this effectively establishes a fair market price for the shares. All option grants going forward must thus be granted at $1. This may not be good, because companies may have been able to grant options to employees at 10 cents had no such price been set (this lower price is desirable, because it gives greater upside). The trick is to turn the founders stock into “preferred” shares, which can get around that problem. That, in turn, has some tax implications, but of course — with good lawyers — there are even ways around that.

So here’s how Parker and Venuto structured the FF class (Parker and Thiel named it “FF’ and the name stuck): The FF has a single preference clause that distinguishes it from all other stock. It is convertible to any future class of stock, when certain conditions are true. For example, the holder of FF can convert it into say, a Series B class of stock and sell it to investors, at which point it takes on all the rights and preferences of Series B stock. But it can only be done during the new issuance of that Series B, and only when that Series B is sold to investors; you can’t convert randomly.

Parker tells VentureBeat he spent nine months thinking about how to align founders’ interests more closely with venture capitalists, and this is what he came up with. The stock was first used at Powerset, and has since been used at a secretive companies Philotic in Berkeley, and David Sack’s Geni (no web sites).

[Pictured above is Ken Howery (left), Peter Thiel (middle) and Sean Parker (right). Photo by Chron's Katy Raddatz]

(Update: See our subsequent funding story here)

We promised to bring you more on Powerset, that secretive company that wants to better Google with a new kind of search.

Powerset is going after the holy grail. It is called “natural language” search, or understanding language as it is actually spoken — and that is something that has defied everyone until now, even the Google guys.

Take, for example, if you type “Books by Children” into Google’s search box. Google essentially drops the word “by” and looks for all the pages that are relevant to “books” and “children.” That’s because the English language is so idiomatic that no engine has been able to understand meaning within sentences. Some companies, most famously Ask Jeeves, have tried. You prompted Ask Jeeves’ engine with a question ending with question mark, but as soon as your question got remotely complex, Ask Jeeves broke down • because its engine could only answer specific questions its engineers had prepped it for.

Google, while acknowledging that natural language is a big goal, hasn’t made very big advances in the area. This makes sense, because people have become trained to use “a grunting pidgin language,” as Powerset’s Barney Pell (pictured here) puts it. Pell calls this “Keyword-ese.” Many search engines recognize some advanced query syntaxes — for example to find Web pages that don’t carry certain words, or that have two words within a certain number of words of each other, and so on.

But people have a hard time remembering these advanced syntaxes, and each search engine has a different syntax. Finally, Google’s core engine has been built around this keyword-ese language, and it is hard to change all of the layers that have been built around it.

Which is where Pell and co-founder Steve Newcomb come in. We talked with them at their offices in Palo Alto. Pell is an enthusiastic guy, and has a rich career developing intelligent systems, but also trying to make them work in the market. Pell has just posted about the start-up on his blog here. Newcomb is the operations guys, previously having worked at voice recognition company Promptu.

Search is so crucial in our lives, Pell says, it is like oxygen. “It is a metabolic function,” Pell says. And yet, search is surprisingly underdeveloped.

Powerset is trying to solve the natural language problem, by making its core engine understand concepts of time, place, sentiment and other intent. But Pell and Newcomb stop short of going into the details of their computational linguistics approach, saying it is sensitive. They are also giving no dates about when it will be released.

They insist, however, that it is a radical improvement. So when Craig Silverstein, first employee of Google said it will take many years to get a computer to a point to understand exactly what people are searching for, Powerset thinks differently: “It is not a long way away,” says Pell. “This is not a change of some technology out on the periphery,” adds Newcomb, “we’re changing the core of the engine.”

He says such a transformation hasn’t happened in eight years, since Google invented “page rank,” a concept that ranks a page higher in relevance depending on how many people are linking to it. So while Google has bought companies like Applied Semantics to help find “themes” on web pages, Newcomb shakes his head, and says Powerset wants to do much more. “We’re switching the core out,” he says, adding that when you do that, you’re also going to fundamentally change the image, video, blog and all other searches that Google is doing. Google may be caught in an “innovators dilemma,” co-founder Steve Newcomb says, because it can’t turn on dime.

So, is this so much hype?

They point to the credibility of their investors, who include PayPal co-founder Peter Thiel, analyst Esther Dyson, Reid Hoffman and the Amidzad fund.

And when you talk with Pell, you get a sense for why he may be as prepared as anyone to do this. While an undergraduate at Stanford, he spent much of his time working at SRI’s “natural language group.” There, back in 1988, he began working on the problem that restricts search engines from relying on natural language: In natural language systems, you have to teach the system every single word. If the system, doesn’t know a word, it crashes. So Barney went to work taking the words the system did know, and bridged the gaps formed by the words it didn’t know, which he said created better results. Later he built a language engine that talked with a office processing system — this way, a company could ask the system say, “What were the top five products ordered over the past week?” and the system could spit the results back in the form of a table. His PhD was in machine learning and games.

Then he worked at NASA, and architected the artificial intelligence system that was embedded in the $200 million Deep Space 1 mission. For a week, his system operated the mission autonomously. In 1999, he left and joined the internet revolution, working with Stockmaster.com, and then at Whizbang Labs, which used machine learning and statistical natural language processing to build advanced search applications. His company built Flipdog, a site that extracted job listing information from millions of sites on the Web • which was sold to Monster.com.

Later, exhausted by the post-bubble era, Pell returned to NASA and managed an 80-person operating division deploying information technology for NASA’s missions. When the market recovered, Pell did a brief stint at Silicon Valley venture firm Mayfield, which he left to start Powerset.

On the one hand, Pell and Newcomb have detractors. “We do have a lot of skeptics,” Pell said. But he and Newcomb also believe they are on to something: “We show the demo to people, and their jaws drop, and they say ‘Holy (expletive),’” says Newcomb. “They say: ‘I’ll never use the old search engines again!’”

Update: See search expert Danny Sullivan’s scathing critique of Powerset ambitions.